Control Administrator's Guide - Altair PBS Works · Altair Control 2018.4 Administrator’s Guide iv Altair Evolve™ 2017.3 ©1993-2018 Altair 365™ ©1994-2018 Altair intellectual

PBS Works is a brand of

Altair 4

Altair Control 2018.4 Administrator’s Guide

Intellectual Property Rights Notice: Copyrights, Trademarks, Trade Secrets, Patents and Third Party Software Licenses

Updated: February 5, 2019.

Altair® PBS Works® v.2018.4Accelerating Innovation in the Cloud™ Copyright© 1994-2018 Altair Engineering Inc. All Rights Reserved.

PBS Works - Accelerating Innovation in the Cloud™Altair PBS Professional® ©1994-2018Altair Control © 2008-2018; (formerly PBS Control)Altair Access © 2008- 2018; (formerly PBS Access)Altair Accelerator © 1995- 2018; (formerly NetworkComputer)Altair Accelerator plus © 1995- 2018; (formerly WorkloadXelerator)Altair FlowTracer © 1995- 2018; (formerly FlowTracer)Altair Allocator © 1995- 2018; (formerly LicenseAllocator)Altair Monitor © 1995- 2018; (formerly LicenseMonitor)Altair Hero © 1995- 2018; (formerly HERO)

Note:Compute Manager™ ©2012-2017 is now part of Altair AccessDisplay Manager™ ©2013-2017 is now part of Altair AccessPBS Application Services ©2008-2017 is now part of Altair AccessPBS Analytics ©2008-2017 is now part of Altair ControlPBS Desktop ©2008-2012 is now part of Altair Access, specifically Altair Access desktop, which also has Altair Access web and Altair Access mobilee-Compute™ ©2000-2010 was replaced by “Compute Manager” which is now Altair Access

Altair HyperWorks - A Platform for Innovation®Altair AcuConsole ©2006-2018Altair AcuSolve ©1997-2018Altair ElectroFlo ©1992-2018Altair ESAComp ©1992-2018Altair Feko ©1999-2014 Altair Development S.A. (Pty) Ltd.; ©2014-2018 Altair Engineering Inc.Altair Flux ©1983-2018Altair FluxMotor ©2017-2018Altair HyperCrash ©2001-2018Altair HyperGraph ©1995-2018

Special Notice: Pre-release versions of Altair software are provided ‘as is’, without warranty of any kind. Usage of pre-release versions is strictly limited to non-production purposes.

ii


Altair HyperMesh ©1990-2018Altair HyperStudy ©1999-2018Altair HyperView ©1999-2018Altair Virtual Wind Tunnel ©2012-2018Altair HyperXtrude ©1999-2018Altair MotionSolve ©2002-2018Altair MotionView ©1993-2018Altair Multiscale Designer ©2011-2018Altair OptiStruct ©1996-2018Altair Radioss ©1986-2018Altair SimLab ©2004-2018Altair nanoFluidX © 2013-2018 Fluidyna GmbH, © 2018 Altair Engineering Inc.Altair ultraFluidX © 2010-2018 Fluidyna GmbH, © 2018 Altair Engineering Inc.Altair WinProp ©2000-2018Altair ConnectMe ©2014-2018Various other products including Altair solidThinking Platform software products.

Altair Packaged Solution Offerings (PSOs)Altair Automated Reporting Director ©2008-2018Altair GeoMechanics Director ©2011-2018Altair Impact Simulation Director ©2010-2018Altair Model Mesher Director ©2010-2018Altair NVH Director ©2010-2017Altair Squeak and Rattle Director ©2012-2018Altair Virtual Gauge Director ©2012-2018Altair Weight Analytics ©2013-2017Altair Weld Certification Director ©2014-2018Altair Multi-Disciplinary Optimization Director ©2012-2018

solidThinking - Where Innovation Begins™Altair Inspire™ 2019 ©2009-2018 including Altair Inspire Motion and Altair Inspire Structures Altair Inspire Extrude-Metal 2019 ©1996-2018 (formerly Click2Extrude®-Metal) Altair Inspire Extrude-Polymer 2019 ©1996-2018 (formerly Click2Extrude®-Polymer)Altair Inspire Cast 2019 ©2011-2018 (formerly Click2Cast®) Altair Inspire Form 2019 ©1998-2018 (formerly Click2Form®)Altair Inspire Mold 2019 ©2009-2018Altair Compose™ 2019 ©2007-2018 (formerly solidThinking Compose®) Altair Activate™ 2019 ©1989-2018 (formerly solidThinking Activate®)Altair Embed™ 2019 ©1989-2018 (formerly solidThinking Embed®)• Altair Embed SE 2019 ©1989-2018 (formerly solidThinking Embed® SE)• Altair Embed/Digital Power Designer 2019 ©2012-2018Altair SimLab™ 2019 ©2004-2018

iii


Altair Evolve™ 2017.3 ©1993-2018Altair 365™ ©1994-2018

Altair intellectual property rights are protected under U.S. and international laws and treaties. Additionally, Altair software is protected under patent #6,859,792 and other patents pending. All other marks are the property of their respective owners.

ALTAIR ENGINEERING INC. Proprietary and Confidential. Contains Trade Secret Information.

Not for use or disclosure outside of Altair and its licensed clients. Information contained in Altair software shall not be decompiled, disassembled, “unlocked”, reverse translated, reverse engineered, or publicly displayed or publicly performed in any manner. Usage of the software is only as explicitly permitted in the end user software license agreement. Copyright notice does not imply publication.

Third party software licensesAcuConsole contains material licensed from Intelligent Light (www.ilight.com) and used by permission.

Software Security Measures:

Altair Engineering Inc. and its subsidiaries and affiliates reserve the right to embed software security mechanisms in the Software for the purpose of detecting the installation and/or use of illegal copies of the Software. The Software may collect and transmit non-proprietary data about those illegal copies. Data collected will not include any customer data created by or used in connection with the Software and will not be provided to any third party, except as may be required by law or legal process or to enforce our rights with respect to the use of any illegal copies of the Software. By using the Software, each user consents to such detection and collection of data, as well as its transmission and use if an illegal copy of the Software is detected. No steps may be taken to avoid or detect the purpose of any such security mechanisms.

iv

http://www.altairhyperworks.com/ClientCenter3rdPartyLicense.aspx

http://www.altairhyperworks.com/ClientCenter3rdPartyLicense.aspx

http://www.ilight.com


Technical Support

This document is proprietary information of Altair Engineering, Inc.

Location Telephone e-mail

Australia +1 800 174 396 [email protected] +86 (0)21 6117 1666 [email protected] +33 (0)1 4133 0992 [email protected] +49 (0)7031 6208 22 [email protected] +91 80 66 29 4500

+1 800 425 0234 (Toll Free)[email protected]

Italy +39 800 905595 [email protected] +81 3 6225 5821 [email protected] +82 70 4050 9200 [email protected] +91 80 66 29 4500


North America +1 248 614 2425 [email protected] +49 7031 6208 22 [email protected] +46 (0) 46 460 2828 [email protected] +91 80 66 29 4500


South Africa +27 21 831 1500 [email protected] America +55 11 3884 0414 [email protected] Kingdom +44 (0)1926 468 600 [email protected]

iv

Contents

1 Overview......................................................................................................................... 7

1.1 Document Conventions.................................................................................................81.2 About Control............................................................................................................. 91.3 Supported Platforms, Browsers and Product Configurations............................................. 101.4 Control Components, Services and Ports...................................................................... 121.5 Deployment Options...................................................................................................161.6 Third-Party Dependencies........................................................................................... 201.7 PBS Works Licensing..................................................................................................221.8 Before Installing Control.............................................................................................24

2 Preparing to Install Control..................................................................................... 25

2.1 Installation Prerequisites.............................................................................................262.2 Backing Up Cloud Data for a Migration.........................................................................27

Backing Up Cloud Data to Migrate from 2018.2 to 2018.4............................................ 27Backing Up Cloud Data to Migrate from 2018.3 to 2018.4............................................ 28

2.3 Preserving Analytics Configuration Files........................................................................ 302.3.1 Analytics Data Forwarded by the Data Collector..................................................302.3.2 Saving Analytics Configuration Files for Older Versions of PBSA............................. 302.3.3 Saving Analytics 2018.3 Configuration Files........................................................30

2.4 Uninstalling Previous Versions of Control...................................................................... 322.4.1 Uninstalling Older Versions of PBS Control......................................................... 322.4.2 Uninstalling PBS Control 2018.3........................................................................32

2.5 Cleaning Up Installation Artifacts.................................................................................34

3 Installing Control and Third-Party Software........................................................35

3.1 Installing Docker....................................................................................................... 363.1.1 Installing Docker for RHEL and CentOS............................................................. 363.1.2 Installing Docker for SLES and SUSE.................................................................363.1.3 Recommended Docker Logging Configuration......................................................37

3.2 Installing MongoDB.................................................................................................... 383.2.1 Installing MongoDB for RHEL and CentOS.......................................................... 383.2.2 Installing MongoDB for SLES and SUSE............................................................. 39

3.3 Cloud Prerequisite Script............................................................................................ 403.4 Installing Control....................................................................................................... 413.5 Restoring Cloud Data................................................................................................. 43

3.5.1 Restoring Cloud Data for a Migration from 2018.2 to 2018.4.................................433.5.2 Restoring the Cloud Data for a Migration from 2018.3 to 2018.4........................... 45

3.6 Installing the Analytics Data Collector.......................................................................... 48

1

4 Post-Installation Configuration............................................................................... 50

4.1 Settings the Cloud Administrator's Email...................................................................... 514.2 Configuring a SMTP Server for Cloud Bursting Quota Alerts.............................................524.3 Configuring a SMTP Server for Monitoring Alerts............................................................534.4 Opening Ports........................................................................................................... 544.5 Setting up the Analytics Environment...........................................................................554.6 Configuring Analytics..................................................................................................56

4.6.1 Configuring Holidays........................................................................................564.6.2 Configure Prime and Non-Prime Hours...............................................................564.6.3 Configuring Exit Codes.................................................................................... 574.6.4 Configuring Custom Groups..............................................................................58

4.7 Logging into Control.................................................................................................. 594.8 Granting Access Privileges.......................................................................................... 60

4.8.1 Default Roles and Privileges............................................................................. 604.8.2 Adding a New Role......................................................................................... 614.8.3 Assigning a Role to a User...............................................................................614.8.4 Changing the Privileges of a Role..................................................................... 62

4.9 Adding an HPC Cluster...............................................................................................63

5 Uninstalling Control....................................................................................................66

5.1 Stopping Bursted Cloud Nodes.................................................................................... 675.2 Uninstalling Control....................................................................................................685.3 Uninstalling the Data Collector.................................................................................... 695.4 Uninstalling Cloud after a Manual Installation................................................................70

6 Starting and Stopping Services...............................................................................71

6.1 Service Commands.................................................................................................... 726.2 Control Service Commands......................................................................................... 73

6.2.1 Starting Control.............................................................................................. 736.2.2 Stopping Control............................................................................................. 736.2.3 Restarting Control........................................................................................... 736.2.4 Determining the Status of all Control Services....................................................74

6.3 Cloud Service Commands........................................................................................... 766.3.1 Start and Stop Cloud After a Manual Installation.................................................766.3.2 Service Commands After Installing Using the Installer......................................... 77

6.4 Analytics Service Commands.......................................................................................796.4.1 Starting the Analytics Service...........................................................................796.4.2 Stopping the Analytics Service..........................................................................796.4.3 Restarting the Analytics Service........................................................................796.4.4 Determining the Status of the Analytics Service..................................................80

2

7 Analytics Commands.................................................................................................. 82

7.1 Configure Nodes........................................................................................................ 837.2 Configure Holidays..................................................................................................... 887.3 Configure Custom Groups........................................................................................... 947.4 Configure Exit Codes..................................................................................................987.5 Configuring the Data Collector...................................................................................1027.6 Configuring the Parser..............................................................................................1057.7 Resetting the Analytics Dataset................................................................................. 1087.8 Changing the Analytics Database Password................................................................. 1127.9 Creating a Diagnosis File for Troubleshooting Issues.....................................................115

8 Cloud Bursting............................................................................................................116

8.1 Cloud Bursting.........................................................................................................1178.2 Cloud Bursting Terminology.......................................................................................1198.3 Prerequisites for Cloud Bursting.................................................................................1208.4 Cloud Bursting Features............................................................................................1218.5 PCLM CLI Client.......................................................................................................123

8.5.1 Installing the PCLM CLI Client.........................................................................1238.6 Installing and Configuring a Bind DNS Server..............................................................1248.7 Cloud Bursting Startup Script.................................................................................... 1288.8 Configuring Amazon Web Service Cloud Bursting......................................................... 132

8.8.1 Creating an Amazon Web Service User Account................................................ 1328.8.2 Adding an Amazon Web Service Cloud Account................................................. 1338.8.3 Creating a Virtual Private Cloud Network..........................................................1348.8.4 Creating an Internet Gateway.........................................................................1358.8.5 Updating the VPC Route Table........................................................................ 1358.8.6 Creating a Virtual Machine............................................................................. 1368.8.7 Installing and Configuring the PBS MoM on the VM............................................1388.8.8 Creating an OS Image...................................................................................1398.8.9 Creating a Cloud Bursting Scenario................................................................. 140

8.9 Configuring Microsoft Azure Cloud Bursting................................................................. 1438.9.1 Registering the Control Application with Azure.................................................. 1438.9.2 Adding a Microsoft Azure Cloud Account...........................................................1468.9.3 Creating a Resource Group.............................................................................1478.9.4 Creating a Virtual Network............................................................................. 1478.9.5 Creating a Network Security Group and Add Security Rules................................ 1488.9.6 Creating a Virtual Machine............................................................................. 1498.9.7 Installing and Configuring the PBS MoM on the VM............................................1518.9.8 Creating an OS Image...................................................................................1528.9.9 Creating a Bursting Scenario.......................................................................... 153

8.10 Configuring Google Cloud Platform Cloud Bursting......................................................1578.10.1 Creating a Project....................................................................................... 1578.10.2 Creating a Service Account...........................................................................157

3

8.10.3 Adding a GCP Cloud Account........................................................................ 1588.10.4 Creating a Virtual Private Cloud Network........................................................ 1598.10.5 Creating a Virtual Machine............................................................................1608.10.6 Installing and Configuring the PBS MoM on the VM.......................................... 1618.10.7 Creating an OS Image................................................................................. 1628.10.8 Creating a Cloud Bursting Scenario................................................................163

8.11 Configuring Oracle Cloud Platform Cloud Bursting.......................................................1658.11.1 Creating an Oracle Cloud User Account.......................................................... 1658.11.2 Generating an SSH Public for the Oracle Cloud User.........................................1668.11.3 Obtaining the Root Compartment Identifier.....................................................1678.11.4 Obtaining the Tenancy Identifier....................................................................1688.11.5 Adding an Oracle Cloud Account....................................................................1688.11.6 Creating a Virtual Cloud Network.................................................................. 1698.11.7 Creating a Virtual Machine............................................................................1718.11.8 Installing and Configuring the PBS MoM on the VM.......................................... 1748.11.9 Creating an OS Image................................................................................. 1768.11.10 Creating a Cloud Bursting Scenario..............................................................176

8.12 Configuring Orange Cloud Flexible Engine for Cloud Bursting....................................... 1798.12.1 Creating an Orange Cloud Flexible Engine User Account....................................1798.12.2 Selecting a Region.......................................................................................1818.12.3 Adding an Orange Cloud Flexible Engine Account.............................................1818.12.4 Checking Orange Cloud Flexible Engine Account Service Quota.......................... 1828.12.5 Creating a Virtual Private Cloud.................................................................... 1838.12.6 Creating a Virtual Machine............................................................................1848.12.7 Installing and Configuring the PBS MoM on the VM.......................................... 1858.12.8 Creating an OS Image................................................................................. 1868.12.9 Creating a Cloud Bursting Scenario................................................................187

8.13 Configuring HUAWEI Cloud for Cloud Bursting............................................................1908.13.1 Obtain the HUAWEI Cloud Administrator Credentials........................................ 1908.13.2 Adding an HUAWEI Cloud Account................................................................. 1918.13.3 Checking HUAWEI Cloud Account Service Quotas.............................................1928.13.4 Creating a Virtual Private Cloud.................................................................... 1928.13.5 Creating a Virtual Machine............................................................................1938.13.6 Installing and Configuring the PBS MoM on the VM.......................................... 1958.13.7 Creating an OS Image................................................................................. 1968.13.8 Creating a Cloud Bursting Scenario................................................................197

8.14 Configuring Open Telekom Cloud for Cloud Bursting................................................... 2008.14.1 Obtain the OTC Administrator Credentials....................................................... 2008.14.2 Adding an OTC Account................................................................................2018.14.3 Checking OTC Account Service Quotas........................................................... 2028.14.4 Creating a Virtual Private Cloud.................................................................... 2028.14.5 Creating a Virtual Machine............................................................................2048.14.6 Installing and Configuring the PBS MoM on the VM.......................................... 2058.14.7 Creating an OS Image................................................................................. 2068.14.8 Creating a Cloud Bursting Scenario................................................................207

8.15 Testing Cloud Bursting with the CLI......................................................................... 2108.15.1 Bursting an Execution Node in the Cloud........................................................210

4

8.15.2 Unbursting an Execution Node in the Cloud.................................................... 2118.16 Configuring PBS Professional for Cloud Bursting......................................................... 212

8.16.1 Configuring Your Network............................................................................. 2128.16.2 Configuring Cloud Bursting Custom Resources.................................................2128.16.3 Setting Cloud Limits at Server...................................................................... 2158.16.4 Creating and Configuring the Cloud Queue..................................................... 2158.16.5 Configuring the Cloud Bursting Hook..............................................................2168.16.6 Testing Cloud Bursting by Submitting a Job.................................................... 2218.16.7 Managing Cloud and On-premise Jobs............................................................2228.16.8 Disabling Cloud Bursting...............................................................................224

9 PBS Professional Considerations...........................................................................225

9.1 Upgrading or Reconfiguring PBS Professional............................................................... 226

10 Advanced Configuration........................................................................................ 227

10.1 Advanced Configuration by Component..................................................................... 22810.2 Advanced Configuration for the Web Application.........................................................230

10.2.1 Advanced Configuration for Authentication...................................................... 23010.2.2 Advanced Configuration for Licensing............................................................. 23510.2.3 Configuring the Session Timeout................................................................... 23610.2.4 Changing the Gateway HTTPS Port................................................................ 236

10.3 Advanced Configuration for Monitoring......................................................................23710.3.1 Configuring Concurrent Threads for Retrieving Job Details.................................23710.3.2 Configure the Job and Node Details Refresh Interval........................................ 237

10.4 Advanced Configuration for Analytics........................................................................ 23910.4.1 Changing the Location Where the Data Collector Gathers theAccounting Logs.....................................................................................................23910.4.2 Changing the Analytics Hostname..................................................................23910.4.3 Resetting the Analytics Accounting Log Data................................................... 240

10.5 Advanced Configuration for Cloud.............................................................................24210.5.1 Accessing Cloud Analytics............................................................................. 24210.5.2 Accessing the CloudFlow Interface................................................................. 24210.5.3 Changing the Hostname of the Cloud Component............................................ 242

11 Troubleshooting.......................................................................................................244

11.1 Logging................................................................................................................. 24511.1.1 Location of Log Files....................................................................................24511.1.2 Web Application Logging...............................................................................24611.1.3 Configure and Simulator Logging...................................................................24911.1.4 Monitor Logging...........................................................................................25011.1.5 Analytics Logging.........................................................................................25211.1.6 Cloud Logging............................................................................................. 253

11.2 Tracing an Issue.................................................................................................... 254

5

11.2.1 Diagnosis Script......................................................................................... 25411.2.2 Using the Diagnosis Script............................................................................255

11.3 Troubleshooting Port Conflicts.................................................................................. 25611.4 Troubleshooting Cloud Bursting DNS Errors............................................................... 257

A-1 Installing Control Without Internet Connectivity......................................... 258

A-1.1 Installing Docker...................................................................................................259Recommended Docker Logging Configuration............................................................ 259

A-1.2 Installing MongoDB............................................................................................... 260A-1.3 Cloud Prerequisite Script........................................................................................261A-1.4 Installing Control.................................................................................................. 262A-1.5 Installing Cloud Manually....................................................................................... 264

Installing Docker....................................................................................................264Download and Save the Cloud Images..................................................................... 265Load the Cloud Images.......................................................................................... 266Installing the Cloud Containers................................................................................266Linking Cloud with the PBS Works License Server......................................................267Configuring a SMTP Server for Cloud Bursting Quota Alerts.........................................267Starting and Initializing Cloud................................................................................. 268Restoring Cloud Data............................................................................................. 270

A-1.6 Installing the Analytics Data Collector..................................................................... 274

Index.................................................................................................................................276

6

Overview 1

1 Overview

This chapter covers the following:

• 1.1 Document Conventions (p. 8)

• 1.2 About Control (p. 9)

• 1.3 Supported Platforms, Browsers and Product Configurations (p. 10)

• 1.4 Control Components, Services and Ports (p. 12)

• 1.5 Deployment Options (p. 16)

• 1.6 Third-Party Dependencies (p. 20)

• 1.7 PBS Works Licensing (p. 22)

• 1.8 Before Installing Control (p. 24)

Control Administrator's Guide1 Overview p.8

1.1 Document ConventionsCommon typographical conventions for Control technical publications.

PC_HOMEThe Control home directory which contains configuration and logging files. Default is:

/var/spool/pbsworks/2018.4/pbscontrol/home

This default location can be overridden during the installation of Control.

PC_EXECThe Control execution directory which contains binaries and scripts. Default is:

/opt/altair/pbsworks/2018.4/pbscontrol/exec

This default location can be overridden during the installation of Control.

PC_INSTALLERThe Control installer directory which contains installer configuration and log files. Default is:

/opt/altair/pbsworks/2018.4/pbscontrol/installer

The location of this directory is based on the location of PC_EXEC.

PBSA_HOMEThe PBSA installation directory. Default is:

/opt/altair/pbsa/VERSION/portal

Where VERSION is the version of PBSA.

PBS_HOMEThe location where the PBS Professional daemon/service configuration files, accounting logs, etc.are installed. Default is:

/var/spool/PBS

PBS_EXECThe location where the PBS Professional executable programs are installed. Default is:

/opt/pbs/default

PAS_HOMEThe PBS Application Services home directory which contains configuration and logging files.Default is:

/var/spool/pas

PAS_EXECThe location where the PBS Application Services executable programs are installed. Default is:

/opt/altair/pas/VERSION/pasWhere VERSION is the version of PAS.

Proprietary Information of Altair Engineering


1.2 About ControlHPC Administrators control center for managing, optimizing and forecasting HPC Resources.

Altair’s Control is an easy-to-use web application for monitoring and managing jobs and nodesof an High-Performance Computing (HPC) cluster with advanced analytics to support data-drivenplanning and decision making. Also, administrators can perform what-if analysis for determining themost productive way to scale an HPC system's resources by running simulations and manage cloudappliances.

Features include:

• Single pane of glass: configure, deploy, monitor, burst, manage, troubleshoot, simulate, analyze,tune

• Real-time monitoring: simplify troubleshooting and maintenance

• Reporting: Analytics powered by Envision

• Workload simulator: simulate and optimize infrastructure sizing

• Multi-cloud bursting: burst to any cloud for peak loads

• Modern UX: drag-and-drop simplicity



1.3 Supported Platforms, Browsers and ProductConfigurationsA list of supported OS platforms and browsers for the 2018.4 release.

Supported Platforms for ControlCentOS 7.2, 7.3, and 7.4

RHEL 7.2, 7.3, and 7.4

SLES 12 SP2

OpenSUSE 42.2

Note: Minor versions of the operating systems listed above can be installed. However, theinstaller will issue a warning message indicating that the unsupported operating system maynot perform as expected. SLES 12 SP3 has not yet been tested.

Supported BrowsersThe latest version of the following browsers are supported:

• Chrome

• Firefox

• Safari

Supported Product ConfigurationsThe currently supported product configurations are:

Table 1: Supported Product Configurations

PBS Professional MongoDB Control

14.2 & 18.2.x 3.4 & 3.6 2018.4

Hardware RequirementsControl requires a minimum hardware configuration:

Table 2: Hardware Requirements for Control

Hardware Minimum Requirement Recommended

CPU 8 cores 8 cores

Memory (physical) 16 GB 32 GB



Hardware Minimum Requirement Recommended

Disk space 80 GB 100 GB

Additional CPUs, memory, and disk space may be required depending upon the Control componentsinstalled and the size of your site's HPC cluster. Large environments may need more memory forAnalytics.



1.4 Control Components, Services and PortsOverview of the components and services available after installation.

Control ComponentsControl consists of a collection of components and their associated services, accessed and invokedvia a common interface. Each of these components can conceivably reside on a separate host; for thepurposes of this release it is recommended to install everything on a single host.

Components and their associated responsibility are:

Web InterfaceProvides client access to the simulation, analytics, configuration, cloud bursting, and monitoringcapabilities of Control. It also provides traffic routing through the gateway/proxy server andmessaging routing utilizing RabbitMQ.

AnalyticsProvides analytics reporting and simulation results. Sub-components:

• Analytics Server - parses accounting logs and node information and provides analytics.

• Data Collector - copies the accounting logs and node information from your site’s PBSProfessional cluster to the machine where the Analytics server is installed.

SimulatorAllows administrators to run simulations and is responsible for storage of simulation data.

MonitorAllows administrators to monitor and manage HPC nodes and jobs and create alerts.

ConfigureAllows administrators to configure default WLM server, scheduler and security settings, andmanage queues, resources, resource defaults and limits.

Cloud

Provides cloud bursting capabilities so that jobs can be run in the cloud based on demand.

Services Installed by ComponentA number of services are installed with Control. Below is a diagram of what services are installed bycomponent.



Note: When the Web Application, Configure and/or Simulator are installed on the samemachine, a single instance of the Framework services are installed and are shared betweenthe components.

Figure 1: Services by Component

Below is a description of each service:

Component Services/Port Description

pcgui.service (9501) Controls the web application.

pbsworks-nginx.service (9500) Gateway service that controlscommunication betweencomponents.

pbswauth.service (9100) Authentication and licensingservice that providesuser authentication andcommunication with the licenseserver.

Web Interface

pbsworks-notification.service

• Web Socket (9105)

Controls UI notifications.




• HTTP ReST(9106)

pbsworks-notification-mongod.service (9575) Notifications database service.

pbsworks-messagerouter-rabbitmq.service(9840)

RabbitMQ service.

pbsworks-messagerouter-http.service (9820) Message router service (forAMQP clients).

pbsworks-messagerouter-mongod.service(9835)

Message router databaseservice.

pbswadmin-orchestrator.service port (9850) Framework HTTP Orchestratorservice.

pbswadmin-worker.service (no port) Framework AMQP Executorservice.

pbswadmin-mongod.service (9860) Framework database service.

pbsmonitor.service (9600) Controls job and nodemonitoring and alerts.

Monitor

pbsmonitor-postgres.service (9603) Monitor database service.

pbswadmin-orchestrator.service (9850) Framework HTTP Orchestratorservice.

Simulator

pbswadmin-worker.service Framework AMQP Executorservice.

pbswadmin-orchestrator.service (9850) Framework HTTP Orchestratorservice.

Configure

pbswadmin-worker.service Framework AMQP Executorservice.

Analytics pbsworks-pbsa (9000)

• Tomcat-8 (9080)

• On Demand Analysis HTTP (9143)

• MonetDB (9200)

• Envision HTTPS (9443)

• Envision ActiveMQ (9672)

• MongoDB (9700)

Provides all back-end analytics.




pbsworks-dc (9343) Analytics Data Collector

Cloud pbscloud.service (9980)

• ELK Stack (5601)

• CloudFlow (19980)

Cloud user interface

• Log management

• Workload Visualization

Open PortsThe following ports need to be opened for communication with external locations and should be open onyour firewall to allow for this communication:

• 5601 - ELK Stack

• 9443 - Analytics Data Collector

• 9500 - Control Gateway

• 9980 - Cloud UI

• 19980 - CloudFlow

If you are implementing a distributed deployment where Control components are distributed acrossseveral machines, the following ports must also be opened between Control machines.

• 9820 - Message router service (for AMQP clients)

• 9835 - Message router database service

• 9840 - RabbitMQ service

• 9850 - Framework HTTP Orchestrator service

• 9860 - Framework database service



1.5 Deployment OptionsOverview of the recommended installation option, distributed deployments, and single componentdeployments.

Figure 2: Control Recommended Deployment Option

Installation Sequence1. Install Docker.

2. Install MongoDB.

3. Install Control.

4. Install the Analytics Data Collector for every PBS Professional complex. A Data Collector mustbe installed either on the head node of each PBS Professional complex, or on a machine that hasaccess to the PBS Professional accounting logs for that complex.

Installation TypesThe Control installer offers two installation types for installing its components:



Basic

Installs all components in a single installation process. Minimal information is requested whenutilizing this installation type:

• Service User

• License server

• Installation location for the home and execution directories (PC_HOME and PC_EXEC)

AdvancedUsing this option gives you the ability to pick and choose the components to install. Use thisoption:

• to install the Analytics Data Collector on each machine where the PBS Professional accountinglogs are located (typically the PBS Professional Server).

• to distribute components across multiple machines (a distributed deployment).

• when Cloud must be manually installed because there is no internet connectivity. TheAdvanced option is used to install all components, except for Cloud.

• when you do not want to install all components. For example: your site is only interested inthe Monitoring capabilities of Control. The Advanced option is used to install the Monitor andWeb Interface components.

Note: If you are using the Advanced installation type because you are only interestedin installing a single component, then you must also install the Web Interface alongwith the component.

Both installation types assume that all external dependencies have been satisfied prior to installingControl and that the default ports are free and available for use.

Component DependenciesThere are some dependencies between the Control components. There is an interdependency betweenthe Web Interface and all other components, such that the Web Interface must be installed for any ofthe other components to function properly. Additionally, the Simulator is dependent on Analytics forpresenting simulations results. The following diagram shows an overview of these dependencies:

Figure 3: Component Dependencies



Distributed Deployments

Important: It is recommended to contact your Altair support team to assist with adistributed deployment.

Distributed deployments where components are installed across multiple machines are supported. Inthis situation, Control cannot be installed using the Basic installation type. Instead, each componentor subset of components must be installed using the Advanced installation type. When there is ainterdependency between components, the installer will ask where the dependent components havebeen installed.

Important: When installing components distributively, the same username must bespecified for the Service User during the installation of the components.

Example #1: Install the Web Interface singularly on a machine, while installing the rest of thecomponents on a second machine.

Install steps:

1. Install the Web Interface on machine #1

• The installer asks for the hostname/IP address of the machine(s) where Analytics, Cloud,Monitor, Simulator and Configure are installed.

2. Install all other components on machine #2

• The installer asks for the hostname/IP address of the machine(s) where the Web Interface isinstalled.

Example #2: Install the Web Interface and Monitor component on a machine, while installing the rest ofthe components on a second machine.

Install steps:

1. Install the Web Interface and Monitor component on machine #1

• The installer asks for the hostname/IP address of the machine(s) where Analytics, Cloud,Simulator and Configure are installed.



Example #3: Install the Simulator component singularly on a machine, while installing all othercomponents on a second machine.

1. Install the Simulator on machine #1

• The installer asks for the hostname/IP address of the machine(s) where Analytics and theWeb Interface are installed.





Installing A Single ComponentYour site may be interested in the capabilities of a single component such as Monitor, Cloud or Simulate.Again, Control cannot be installed using the Basic installation type. Instead, the component that you areinterested in installing must be installed using the Advanced installation type.

The Web Interface must be installed for any of the other components to function properly. It can eitherbe installed on the same machine as the component or it can be installed distributively.

If you are interested in only the simulation capabilities of Control, Analytics must also be installed, asthe Simulator is dependent on Analytics for presenting simulations results.

No Internet ConnectivityFor sites where there is no internet connectivity, Cloud must be manually installed after installing allother components of Control.

See AlsoInstalling Control Without Internet Connectivity



1.6 Third-Party DependenciesOverview of the third-party software that is required for installing and running Control.

The following third-party software must be installed manually before installing Control:

DockerDocker must be manually installed prior to installing the Cloud component. The Control installerperforms a check for its presence and the installation of Control is halted and an error message isdisplayed when Docker cannot be found.

MongoDB 3.4 and 3.6MongoDB is used for data storage by all components except Monitor. Due to legal reasons, itsinstallation could not be incorporated into the Control installer for the following components:

• Web Interface

• Analytics

• Configure

• Simulator

MongoDB must be manually installed before installing any of the above components. The Controlinstaller performs a check for its presence and the installation of Control is halted and an error messageis displayed when MongoDB cannot be found.

socatsocat must be manually installed prior to installing the Web Interface, prior to installing Configurestand-alone, or prior to installing Simulator stand-alone. The Control installer performs a check for itspresence and the installation of Control is halted and an error message is displayed when socat cannotbe found.

The following third-party software is installed automatically during the installation of Control. In someinstances, an existing installation of the third-party software may result in a conflict and halt theinstallation of Control.

ApacheApache is installed automatically when any of the Control components are installed and overlays theexisting Apache installation if present on the system. However, a problem exists on SLES/OpenSUSEplatforms, where an existing installation of the apache2-prefork package causes a conflict. In this case,the installation of Control is halted and the installer will display an error message directing the user touninstall all packages related to Apache.

Erlang or Erlang-ERTSA customized version of Erlang is installed when the Web Interface is installed, when Configureis installed stand-alone, or when Simulator is installed stand-alone. The presence of an existing



installation of Erlang or Erlang-ERTS on the system causes the installation of Control to be halted. Theinstaller will display an error message directing the user to uninstall the product.

mod_wsgiA customized version of mod_wsgi is installed automatically when any of the Control components areinstalled. The presence of an existing installation of mod_wsgi on the system causes the installationof Control to be halted. The installer will display an error message directing the user to uninstall theproduct.

OpenSSL 1.0.2xOpenSSL 1.0.2.x is used by all components of Control. When the presence of OpenSSL is undetectableon the system where Control is being installed, OpenSSL 1.0.2x is automatically installed. Whenthe presence of OpenSSL is detected on the system, a check for the correct version is done. If theversion of OpenSSL is not 1.0.2x, the installation of Control is halted and the installer displays an errormessage directing the user to uninstall OpenSSL. Otherwise, when the correct version of OpenSSL isdetected, the installation of OpenSSL is skipped.

PBS Professional and PASThe Simulator component requires PBS Professional and PAS. These products are automatically installedwhen the Simulator component is installed. The presence of an existing installation of PBS Professionalor PAS on the system causes the installation of Control to be halted. The installer will display an errormessage directing the user to uninstall the product.

PostgresPostgres is used by the Monitor component for data storage. It is automatically installed when theMonitor component is installed and by default uses port 9603. As long as an existing installation ofPostgres is using a different port, then there should be no conflicts with the instance installed withControl.

RabbitMQA customized version of RabbitMQ is installed when the Web Interface is installed, when Configureis installed stand-alone, or when Simulator is installed stand-alone. The presence of an existinginstallation of RabbitMQ on the system causes the installation of Control to be halted. The installer willdisplay an error message directing the user to uninstall the product.

Other installation considerations:

SELinuxSELinux is not a prerequisite for installing Control. However, if it is installed on the machine where anyof the Control components are being installed, it must be disabled.



1.7 PBS Works LicensingLicensing model for Control.

Licenses must be installed and available via the Altair LMX license server. Three types of licenses areconsumed by Control:

• PBSWorks User license - consumed when a user logs into Control

• PBSWorks Node license - consumed when an HPC cluster is added via the Monitor tab or whenan active node is parsed by Analytics when node information (pbsnodes -av) is forwarded by theAnalytics Data Collector.

• PBSWorks Bursted Node license - consumed when a node is burst into the cloud.

Licensing for Logging into ControlEach user that logs into Control consumes a PBS Works User license. If a user logs into Control andthen opens a second browser window to access Control, a single PBS Works User license is consumed.

To login to Control at least one PBS Works User license and one PBS Works Node license must beinstalled and available via the Altair LMX license server.

Monitor LicensingWhen a cluster is added to Control, a PBS Works Node license is consumed for each HPC cluster node.A node in Control is considered a PBS Professional vnode regardless of the number of cores or devicesassociated with the vnode.

Once the number of PBS Works Node licenses is exceeded, job and node summary and detailinformation is not available for any newly added HPC clusters. A next to the name of the cluster onthe Job Summary or Node Summary page indicates that Node licenses have been exceeded.

Figure 4: PBS Works Node Licenses Exceeded

Configure LicensingNo licenses are consumed when accessing the capabilities of the Configure component.

Analyze LicensingThe Analytics Data Collector forwards node information (pbsnodes -av) to the Analytics component.Each active node consumes a PBSWorks Node license. The total count of PBSWorks Nodes units neededfor the Analytics component levels against the total count of PBSWorks Nodes units needed for theMonitor component.



Simulator LicensingNo licenses are consumed when accessing the capabilities of the Simulator component.

Cloud LicensingNo licenses are consumed when accessing the capabilities of the Cloud component, except for cloudbursting. PBS Works Bursted Node Hours are consumed when nodes are burst in the cloud. Forexample, if 10 nodes are burst for 10 hours, then 100 Bursted Node Hours are consumed.



1.8 Before Installing ControlThings to know before you install.

Access to Control Post-InstallationAll users who have login access to the machine where Control is installed, can login to Control (validNIS/SSH users). The Service User entered during the installation of Control has managerial privileges,therefore has unrestricted access to all features of Control. After installation, the Service User mustassign access privileges for all other users using the Access Control feature.

If a user logs into Control and access privileges have not been established for that user, a messageis displayed indicating that the user does not have access to any of the features of Control and theyshould contact their administrator.

Previous PBS Professional/PAS InstallationControl cannot be installed on a machine that already has PBS Professional or PAS installed. TheControl installer installs its own PBS Professional/PAS for running simulations. An existing install of PBSProfessional or PAS will result in a conflict. This conflict halts the install. The installer provides an errormessage, directing the user to uninstall PBS Professional and PAS.

Important: Control should NOT be installed on the PBS Professional Server.

See AlsoGranting Access Privileges


Preparing to Install Control 2

2 Preparing to Install Control


• 2.1 Installation Prerequisites (p. 26)

• 2.2 Backing Up Cloud Data for a Migration (p. 27)

• 2.3 Preserving Analytics Configuration Files (p. 30)

• 2.4 Uninstalling Previous Versions of Control (p. 32)

• 2.5 Cleaning Up Installation Artifacts (p. 34)

Control Administrator's Guide2 Preparing to Install Control p.26

2.1 Installation PrerequisitesPrerequisites that must be satisfied before the installation of Control.

Prerequisites for Installing ControlBefore running the Control installer, the following prerequisites must be met:

• Installation must be done as root or as a user with sudo permissions using the sudo command.

• Installation must be done on a machine running on a supported platform.

Please review the third-party dependencies before installing Control. Several third-party packages mustbe installed prior to installing Control. Additionally, the presence of some third-party software packageson the system may cause a conflict as a customized version of the package is installed with Control.

Prerequisites for the Service UserDuring the installation of Control, you are asked to provide a username that will be the Control ServiceUser. All Control services including the database instances will run as this user.

• This user must be an existing user on the machine were Control is being installed.

• The Service User must be able to SSH without a password from and into any machine running theSimulator component (e.g. from machine A to machine A).

• When installing components distributively, the same username must be specified for the ServiceUser during the installation of the components.



2.2 Backing Up Cloud Data for a MigrationBack up the Cloud data so that cloud account and bursting scenarios do not have to be recreated.

The procedure for backing up the Cloud data is different depending upon the version of Controlcurrently installed at your site.

Backing Up Cloud Data to Migrate from 2018.2 to 2018.4Back up the Cloud data in preparation for a migration from 2018.2 to 2018.4 so that cloud account andbursting scenarios do not have to be recreated.

The following steps must be done as root or as a user with sudo permissions using the sudo command.

1. Login to the machine where Cloud is installed.

2. Find your current Cloud version:docker images | grep pocker

The version is the second field.pbscloudio.azurecr.io/pocker 2018.2 a6ce3e73b071 2 months ago 581MB

3. Export the currently installed Cloud version number:export CLOUD_OLD_VERSION='2018.2'

4. Export the Cloud version number that you want to upgrade to:export CLOUD_VERSION='2018.4'

5. Export the Control home directory:

• If you installed Cloud using the Control installer and installed in the default location:export PC_HOME=/var/spool/pbsworks/$CLOUD_OLD_VERSION/pbscontrol/home

• If you installed Cloud using the Control installer and installed in a non-default location:export PC_HOME=PC_HOME_DIR

Where PC_HOME_DIR is the location of the Control home directory.

6. Export the current installation location of the Cloud kard.

• If you installed Cloud using the Control installer:export KARD_PATH=$PC_HOME/cloudio/kard

• If you installed Cloud manually:export KARD_PATH=CLOUD_DIR/kard

Replace CLOUD_DIR with the installation location of Cloud.

7. Launch a pocker container:docker run -ti --rm -v /run/docker.sock:/run/docker.sock \-v $KARD_PATH:/pocker/kard pbscloudio.azurecr.io/pocker:$CLOUD_OLD_VERSION bash

8. Stop Cloud services:pocker clean

9. Exit the pocker container:exit



10. Backup your current Cloud data:cd $KARD_PATH/currenttar czvf ~/cloud_kard.tar.gz *

This will create an archive in the home directory of the current user.

Backing Up Cloud Data to Migrate from 2018.3 to 2018.4Back up the Cloud data in preparation for a migration from 2018.3 to 2018.4 so that cloud account andbursting scenarios do not have to be recreated.



2. Find your current Cloud version:docker images | grep pocker

The version is the second field.pbscloudio.azurecr.io/pocker 2018.3 a6ce3e73b071 2 months ago 581MB

3. Export the currently installed Cloud version number:export CLOUD_OLD_VERSION='2018.3'

4. Export the Cloud version number that you want to upgrade to:export CLOUD_VERSION='2018.4'


• If you installed Cloud using the Control installer and installed in the default location:export PC_HOME=/var/spool/pbsworks/$CLOUD_OLD_VERSION/pbscontrol/home





• If you installed Cloud manually:export KARD_PATH=CLOUD_DIR/kard

Replace CLOUD_DIR with the installation location of Cloud.

7. Launch a pocker container:docker run -ti --rm -v /run/docker.sock:/run/docker.sock \-v $KARD_PATH:/pocker/kard pbscloudio.azurecr.io/pocker:$CLOUD_OLD_VERSION bash

8. Stop Cloud services:pocker clean

9. Exit the pocker container:exit

10. Backup your current Cloud data:cd $KARD_PATH/current



tar czvf ~/cloud_kard.tar.gz *

This will create an archive in the home directory of the current user.



2.3 Preserving Analytics Configuration Files

2.3.1 Analytics Data Forwarded by the Data Collector

Once a Data Collector is installed and started, accounting logs and node information (output frompbsnodes -av) is automatically forwarded to the Analytics server. Afterward, any changes to the nodeinformation and newly generated accounting logs are forwarded daily. The node information is usedto configure Analytics so that administrators do not have to manually configure Analytics using thepbsa-config-nodes command. Analytics assumes this node configuration is in effect for all accountinglogs that are forwarded from the PBS Professional Complex, including historical data (accounting logsgenerated before the Data Collector was installed and started).

If you site has already generated Analytics node details files by running the pbsa-config-nodescommand (for example, because you have already installed a previous version of PBSA or Analytics),you can preserve those files and restore them after installation of Control.

2.3.2 Saving Analytics Configuration Files for OlderVersions of PBSAPreserve Analytics configuration files when the currently installed version of PBSA is 13.1, 2018.0,2018.1 or 2018.2.

Save the Analytics configuration files and use them to configure Analytics after installing Control 2018.4

1. Login to the machine where PBSA 13.1, 2018.0, 2018.1 or 2018.2 is installed.

2. Navigate to PBSA_HOME/services/pbsa/parser/modules/pbsdata/config.

3. Copy the following files to a temporary location such as /tmp:

• all holiday_YYYY.conf

• all custom_group_YYYY.conf

• exit_status.conf

• all nodedetals_YYYY.conf

2.3.3 Saving Analytics 2018.3 Configuration FilesPreserve Analytics configuration files when the currently installed version of Control is 2018.3.

Save the Analytics configuration files and use them to configure Analytics after installing Control 2018.4

1. Login to the machine where Control 2018.3 is installed.

2. Navigate to PC_HOME/config/services/pbsa/parser/modules/pbsdata/config.

3. Copy the following files to a temporary location such as /tmp:

• all holiday_YYYY.conf



• all custom_group_YYYY.conf

• exit_status.conf

• all nodedetals_YYYY.conf



2.4 Uninstalling Previous Versions of Control

2.4.1 Uninstalling Older Versions of PBS ControlUninstall versions of PBS Control older than 2018.3 in preparation of installing Control 2018.4.

The following commands must be done as root or as a user with sudo permissions using the sudocommand.

Important: Use the instructions for uninstalling PBS Control, PBSCloud and PBSA that weredocumented in the Administrator's guide that was provided for the version of the productthat is being uninstalled. Ex. If PBS Control 2018.1 is installed at your site, then use theinstructions for uninstalling the product in the PBS Control 2018.1 Administrator's Guide.

1. Login to the machine where the previous version of PBSA is installed.

2. Uninstall PBSA.

3. Login to the machine where the previous version of the PBSA Data Collector is installed.

4. Uninstall the PBSA Data Collector.

5. Repeat step 3 through step 4 to uninstall all Data Collectors.

6. Login to the machine where the previous version of PBS Control is installed.

7. Uninstall PBS Control.

8. Login to the machine where the previous version of PBSCloud is installed.

9. Uninstall PBSCloud.

2.4.2 Uninstalling PBS Control 2018.3Uninstall PBSControl 2018.3 in preparation of installing Control 2018.4.

The following commands must be done as root or as a user with sudo permissions using the sudocommand.

Important: Use the instructions for uninstalling PBS Control that were documented in thePBS Control 2018.3 Administrator's Guide.

1. Login to the machine where PBS Control 2018.3 is installed.

2. Uninstall PBS Control.

See Uninstalling PBS Control in the PBS Control 2018.3 Administrator's Guide.

3. Login to the machine where the PBS Control 2018.3 Data Collector is installed.

4. Uninstall the Data Collector.

See Uninstalling PBS Control in the PBS Control 2018.3 Administrator's Guide.

5. Repeat step 3 through step 4 to uninstall all Data Collectors.



6. If you manually installed PBSCloud because the machine where it was installed did not haveinternet connectivity:

a) Login to the machine where PBSCloud 2018.3 is installed.

b) Uninstall PBSCloud.

See Uninstalling PBSCloud after a Manual Installation in the PBS Control 2018.3Administrator's Guide.



2.5 Cleaning Up Installation ArtifactsManually clean up some installation artifacts in the case of an unsuccessful uninstallation.

If you suspect that a previous installation of Control was not completely uninstalled (for example, dueto filesystem or permissions issues), make sure the following are removed:

• Control home directory - PC_HOME

• Control execution directory - PC_EXEC

• /usr/sbin/pbsworks-nginx

• /usr/bin/pbs-sim-agent


Installing Control and Third-Party Software 3

3 Installing Control and Third-Party Software

Install required third-party software and all components of Control using the Control installer.


• 3.1 Installing Docker (p. 36)

• 3.2 Installing MongoDB (p. 38)

• 3.3 Cloud Prerequisite Script (p. 40)

• 3.4 Installing Control (p. 41)

• 3.5 Restoring Cloud Data (p. 43)

• 3.6 Installing the Analytics Data Collector (p. 48)

Internet connectivity is required for the installation of Cloud.

Note: If Docker or MongoDB are already installed from a previous installation of Control,then you may skip the steps for their installation.

Control Administrator's Guide3 Installing Control and Third-Party Software p.36

3.1 Installing DockerInstall Docker which is used to host Cloud.

3.1.1 Installing Docker for RHEL and CentOSInstall Docker for RHEL and CentOS platforms.


• For more information about installing Docker CE see Get Docker CE for CentOS.

1. Login to the machine where Control is to be installed.

2. Install required packages.yum install -y yum-utils

3. Setup the docker repository:yum-config-manager --add-repo \https://download.docker.com/linux/centos/docker-ce.repo

4. Install the packets and its dependencies:yum install docker-ce

5. Add the user into the docker group:

Enter this command exactly as is. Do not replace $USER with a username.usermod -aG docker $USER

6. Open a new shell to load the new group permissions:

Enter this command exactly as is. Do not replace $USER with a username.su - $USER

Note: You will need to start a new shell once you have added the user to the dockergroup. Otherwise, you will get an error when trying to use the docker command.

7. Start Docker.systemctl start dockersystemctl enable docker

3.1.2 Installing Docker for SLES and SUSEInstall Docker for SLES and SUSE platforms.


• Docker CE is not supported on SLES, therefore Docker EE must be installed.


2. Install and start the latest version of Docker EE by following these instructions Get Docker EE forSLES.


https://docs.docker.com/install/linux/docker-ce/centos/

https://docs.docker.com/install/linux/docker-ee/suse/



3. Verify that the file /usr/bin/docker-init exists.

4. If the file does not exist:

a) Download the tarball for the latest version of Docker from the Docker repository:

https://download.docker.com/linux/static/stable/x86_64/

b) Extract the archive to /tmp:tar zxvf docker-18.06.1-ce.tgz -C /tmp

c) Execute the command:cp /tmp/docker/docker-init /usr/bin/







3.1.3 Recommended Docker Logging ConfigurationSet the maximum size of the Docker log file and the maximum log files to be kept.

To avoid consuming a lot of disk space and to rotate Docker logs, configure your Docker logging systemfollowing this documentation: https://docs.docker.com/config/containers/logging/configure/.

It is recommended to use the following configuration where the maximum size of the log before it isrolled is 10MB and the maximum number of files that are kept is 10.{ "log-driver": "json-file", "log-opts": { "max-size": "10m", "max-file": "10" }}


https://download.docker.com/linux/static/stable/x86_64/docker-18.06.1-ce.tgz

https://docs.docker.com/config/containers/logging/configure/


3.2 Installing MongoDBInstall MongoDB for use by the Web Interface, Analytics, Configure, and Simulator components for datastorage.

Restriction: Only MongoDB 3.4 and 3.6 are supported.

If the Monitoring component is being installed stand-alone, then MongoDB does not need to beinstalled. Otherwise, if you are installing any of the other components (Web Interface, Analytics,Configure, Simulator), MongoDB must be installed.

3.2.1 Installing MongoDB for RHEL and CentOSInstall MongoDB for RHEL and CentOS platforms.

Installation must be done as root or as a user with sudo permissions using the sudo command.

MongoDB is used by the Web Interface, Analytics, Configure and Simulator components for datastorage. Due to legal reasons, its installation could not be incorporated into the 2018.4 installer,therefore it must be manually installed, wherever the Web Interface, Analytics, Configure, andSimulator components are installed. Once MongoDB is installed, no other configuration is necessary.

1. Login to the machine where Control is going to be installed.

2. Create a file /etc/yum.repos.d/mongodb-org-<ver>.repo.

Where <ver> is the version of MongoDB being installed.

3. Edit the file and choose one of the following options:

• Add the following lines for MongoDB 3.4:[mongodb-org-3.4]name=MongoDB Repositorybaseurl=https://repo.mongodb.org/yum/redhat/$releasever/mongodb-org/3.4/x86_64/gpgcheck=1enabled=1gpgkey=https://www.mongodb.org/static/pgp/server-3.4.asc

• Add the following lines for MongoDB 3.6:[mongodb-org-3.6]name=MongoDB Repositorybaseurl=https://repo.mongodb.org/yum/redhat/$releasever/mongodb-org/3.6/x86_64/gpgcheck=1enabled=1gpgkey=https://www.mongodb.org/static/pgp/server-3.6.asc

4. Save the changes made to the file and exit the editor.

5. Install MongoDB using YUM:yum install -y mongodb-org



3.2.2 Installing MongoDB for SLES and SUSEInstall MongoDB for SLES and SUSE platforms.


MongoDB is used by the the Web Interface, Analytics, Configure and Simulator components for datastorage. Due to legal reasons, its installation could not be incorporated into the 2018.4 installer,therefore it must be manually installed, wherever the Web Interface, Analytics, Configure, andSimulator components are installed. Once MongoDB is installed, no other configuration is necessary.


2. Import the MongoDB public key by choosing one of the following options:

• To install MongoDB 3.4, enter the command:rpm --import https://www.mongodb.org/static/pgp/server-3.4.asc

• To install MongoDB 3.6, enter the command:rpm --import https://www.mongodb.org/static/pgp/server-3.6.asc

3. Add the repository so that you can install MongoDB using zypper.

• For MongoDB 3.4, enter the command:zypper addrepo --gpgcheck \"https://repo.mongodb.org/zypper/suse/12/mongodb-org/3.4/x86_64/" mongodb

• For MongoDB 3.6, enter the command:zypper addrepo --gpgcheck \"https://repo.mongodb.org/zypper/suse/12/mongodb-org/3.6/x86_64/" mongodb

4. Install MongoDB:

• For MongoDB 3.4, enter the command:zypper install mongodb-org-3.4.10 \mongodb-org-server-3.4.10 \mongodb-org-shell-3.4.10 \mongodb-org-mongos-3.4.10 \ mongodb-org-tools-3.4.10

• For MongoDB 3.6, enter the command:zypper install mongodb-org-3.6.0 \mongodb-org-server-3.6.0 \mongodb-org-shell-3.6.0 \mongodb-org-mongos-3.6.0 \mongodb-org-tools-3.6.0


https://www.mongodb.org/static/pgp/server-3.6.asc

https://www.mongodb.org/static/pgp/server-3.6.asc


3.3 Cloud Prerequisite ScriptPrerequisite check for Cloud.

Run the below script to check the following prerequisites before installation:

• Presence of Docker and its version

• Space for the images (/var/lib/docker)

• Connectivity with the Cloud Docker repository

Note: The script will complain that it cannot connect to the Cloud Docker repository if youare installing Cloud manually because the machine where you are installing Control does nothave internet connectivity.

#!/bin/bashset -e

# check dockercommand -v docker >/dev/null 2>&1 || { echo >&2 "[ERROR] docker is required but is not installed or unavailable for this user."; exit 1; }command -v docker-init >/dev/null 2>&1 || { echo >&2 "[ERROR] docker-init is required but is not installed or unavailable for this user."; exit 1; }docker info >/dev/null 2>&1 || { echo >&2 "[ERROR] Cannot connect to the Docker daemon. Check that docker is running and that $USER belongs to docker group."; exit 1; }docker_server_version=$(docker info | grep -i "Server Version" | awk -F ": " '{print $2}' | cut -d '.' -f1)if [ $docker_server_version -lt 17 ]; then echo >&2 "[ERROR] Docker version is $docker_server_version. Only versions >= 17 are supported." exit 1fi

# check disk space on docker's root directorydocker_root=$(docker info | grep -i "Docker Root Dir" | awk -F ": " '{print $2}')if [ $(df -P $docker_root | awk 'NR==2 {print $4}') -lt 31457280 ]; then echo >&2 "[ERROR] Not enough space on $docker_root. Only $(df -Ph $docker_root | awk 'NR==2 {print $4}') available, 30GB is required." exit 1fi

# check registry accesscommand -v curl >/dev/null 2>&1 || { echo >&2 "[ERROR] curl is needed to check registry connectivity."; exit 1; }curl --silent --fail -X GET https://a882d237-5d07-48a7-837e-9bd8652d79ec:kq9ZKzrToLGm78xGBDv5lKx5hg5YOB8JDz3usbp+2eo=@pbscloudio.azurecr.io/v2/_catalog &> /dev/null \|| { echo >&2 "[ERROR] Cannot access to docker registry. Please check connection to https://pbscloudio.azurecr.io"; exit 1; }

# prerequisites metecho "[OK] All prerequisites met. You can proceed to Cloud installation."



3.4 Installing ControlInstall all Control components using the Basic installation type.

Review the installation prerequisites before proceeding with the installation.

Internet connectivity is required to install Control. If there is no internet connectivity, then follow theseinstructions: Installing Control Without Internet Connectivity.

Note: Installer log files are available in PC_INSTALLER/logs.

Several components use MongoDB for data storage, therefore there will be multiple database instancesrunning once this installation is complete. The Monitor component and PBS Professional use Postgres fordata storage, therefore there will be two Postgres database instances running after this installation iscomplete.

Every time Control is installed or uninstalled a file called pbswinstall_op_report is appended withparameter information used during the installation/uninstallation process. The file is placed in the /etc directory. The contents of this file may be valuable for diagnosing issues with the installation/uninstallation of Control.


2. Download the installer tarball AltairControl_Installer_2018.4_486_20181218_065323.tar.gzfrom the download location.

3. Navigate to the directory where the installer tarball was copied by changing your workingdirectory.

4. Extract the contents of the tarball:tar -xvf AltairControl_Installer_2018.4_486_20181218_065323.tar.gz

A directory called PBSControl_Installation is created.

5. Navigate to the PBSControl_Installation directory:cd PBSControl_Installation

6. Invoke the installer by running the installation binary:./run_installer

Tip: If a broken installation is detected then the following message is displayed:

There was a non-fixable broken installation!!Would you like to uninstall? (Y/N)

Enter Y and press ENTER. Once the broken installation is completely removed, try toinstall Control again.

7. Read the introduction, verify the prerequisites for installation, and press ENTER.

8. Page through the license agreement by pressing ENTER until you are asked to accept its terms andconditions.

9. Accept the license agreement by entering 1 and pressing ENTER.

10. Choose the Basic installation type by entering 1 and pressing ENTER.



11. Enter the license server in the format port@hostname and press ENTER.

12. Enter a username that who will be the Control Service User and press ENTER.

The user must exist on the machine where you are installing Control. All Control services includingthe database instances run as this user.

13. Enter where you would like to install the execution binaries (PC_EXEC) or press ENTER to use thedefault location.

14. Enter where you would like to install the configuration files (PC_HOME) or press ENTER to use thedefault location.

15. Review the installation summary and then press ENTER to install the product.Message similar to those below are displayed as the installation progresses.Installing...-------------

This may take a while...Unzipping files into home, exec, and installer folders....

Progress is: 0% Web Interface Beginning PBSWorks Shared InstallationProgress is: 12% Web Interface Completed updating PBS Works Shared config files.........All milestones reached!!

16. Once the installation is complete, set the Service User's PAS password by issuing the command:su - SERVICE_USER -c "PAS_EXEC/pas-commands/bin/pas-passwd"

Where SERVICE_USER is the username of the Service User.



3.5 Restoring Cloud DataRestore the Cloud data so that cloud accounts and bursting scenarios do not have to be recreated.

Note: For the 2018.4 version of Control, Pocker a tool for generating Docker images hasbeen replaced with the Open Source tool pkr.

The procedure for restoring the Cloud data is different depending upon the version of Control previouslyinstalled at your site.

3.5.1 Restoring Cloud Data for a Migration from 2018.2 to2018.4Restore the Cloud data after migrating from Control 2018.2 to 2018.4.

1. Export the Cloud version number:export CLOUD_VERSION='2018.4'


• If you installed Cloud using the Control installer and installed in the default location:export PC_HOME=/var/spool/pbsworks/$CLOUD_VERSION/pbscontrol/home





• If you installed Cloud manually:export KARD_PATH=DIR/kard

Replace DIR with the path where Cloud is installed.

4. Launch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v $KARD_PATH:/pkr/kard \pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash

5. Stop pkr and exit the container:pkr cleanexit

6. Restore the Cloud archive:mv $KARD_PATH/$CLOUD_VERSION $KARD_PATH/${CLOUD_VERSION}.backupmkdir $KARD_PATH/$CLOUD_VERSION && cd $KARD_PATH/$CLOUD_VERSIONtar zxvf ~/cloud_kard.tar.gz

7. Relaunch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \



-v /run/docker.sock:/run/docker.sock \-v $KARD_PATH:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash

8. Update the kard version:cd /pkr/kard/$CLOUD_VERSIONsed -i \-e "s#^project_name:.*#project_name: '$(echo $CLOUD_VERSION | sed 's/\.//')'#" \-e "s#^src_path:.*#src_path: /pkr/kard/$CLOUD_VERSION/./src#" \-e "s#^tag:.*#tag: '$CLOUD_VERSION'#" \-e "s#^driver:.*#driver:\n name: compose#" \-e "\$alicense_server:\n \auth_url: $(grep âuth_url site_config/pacioli.conf | cut -d "=" -f2-)\n \host: $(grep ^host site_config/pacioli.conf | cut -d "=" -f2-)" \meta.yml

9. Load the new kard and rebuild the context:pkr kard load $CLOUD_VERSIONpkr kard updatepkr kard make

10. Start MongoDB:pkr start -s mongodb

11. Update the MongoDB database:docker exec -it mongodb mongo -u admin --authenticationDatabase admin \notification-center -p $(cat /pkr/kard/$CLOUD_VERSION/meta.yml | \grep db_mongo | awk '{print $2}') \--eval "db.createUser({ user: 'notification-center', \pwd: '$(cat /pkr/kard/$CLOUD_VERSION/meta.yml \| grep db_notification-center \| awk '{print $2}')', roles: \[ { role: 'readWrite', db: 'notification-center' } ] });"

12. Start Postgres:pkr start -s postgres

13. Exit pkr container:exit

14. Launch a Mistral container:docker run -ti \-v $KARD_PATH/$CLOUD_VERSION/docker-context/mistral-api/mistral.conf:\/home/mistral/data/mistral.conf:ro \-l postgres --network $(docker inspect postgres \-f "{{.HostConfig.NetworkMode}}") mistral-api:$CLOUD_VERSION bash

15. Fix the version of Mistral.mistral-db-manage --config-file /home/mistral/data/mistral.conf stamp 024

See details of available commands https://docs.openstack.org/mistral/newton/guides/upgrade_guide.html.

16. Perform Mistral update:mistral-db-manage --config-file /home/mistral/data/mistral.conf upgrade head

17. Exit Mistral container:exit

18. Relaunch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v $KARD_PATH:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash


https://docs.openstack.org/mistral/newton/guides/upgrade_guide.html



19. Start pkr:pkr cleanpkr start

20. Update Keystone:pkr ext init base

21. Update the database:docker exec -ti bootstrapper-api bashpython -m pclm_common.dbmigrate bootstrapper listpython -m pclm_common.dbmigrate bootstrapper migrate

22. Exit the container:exitexit

3.5.2 Restoring the Cloud Data for a Migration from2018.3 to 2018.4Restore the Cloud data after migrating from Control 2018.3 to 2018.4.
















8. Update the kard version:cd /pkr/kard/$CLOUD_VERSIONsed -i \-e "s#^project_name:.*#project_name: '$(echo $CLOUD_VERSION | sed 's/\.//')'#" \-e "s#^src_path:.*#src_path: /pkr/kard/$CLOUD_VERSION/./src#" \-e "s#^tag:.*#tag: '$CLOUD_VERSION'#" \-e "s#^driver:.*#driver:\n name: compose#" \-e "\$alicense_server:\n \auth_url: $(grep âuth_url site_config/pacioli.conf | cut -d "=" -f2-)\n \host: $(grep ^host site_config/pacioli.conf | cut -d "=" -f2-)" \meta.yml











18. Update the database:docker exec -ti bootstrapper-api bashpython -m pclm_common.dbmigrate bootstrapper migrate

19. Exit the container:exit





exit



3.6 Installing the Analytics Data CollectorInstall the Analytics Data Collector on each PBS Professional complex.



• The Data Collector must have access to the machine hosting Analytics.

If your site has clusters with the same fully qualified domain name (FQDN), then the Analytics data isaggregated for both clusters under a single FQDN. To allow Analytics data for clusters having the sameFQDN to be tracked separately, a unique name must be provided while installing each Data Collector.



1. Login to the PBS Professional Server.













10. Choose the Advanced installation type by entering 2 and pressing ENTER.



11. Choose the Analyze component by entering 5 and pressing ENTER.

12. Choose to install the Data Collector by entering 2 and pressing ENTER.

13. Enter the hostname of the machine where Control Web Application is installed and press ENTER.


15. Enter the hostname of the machine where Analytics was installed.

16. Enter a unique name (FQDN) for the Data Collector.





20. Review the installation summary and then press ENTER to install the product.Message similar to those below are displayed as the installation progresses.Installing... -------------This may take a while...Unzipping files into home, exec, and installer folders.... Progress is: 0% Analyze Beginning PBSWorks Shared InstallationProgress is: 12% Analyze Completed updating PBS Works Shared config files.........All milestones reached!!

21. Update the system path to include the path to the Analytics Commands:export PATH=$PATH:PC_EXEC/dc/scripts/pbsa-commands/bin

Perform post-installation configuration.


Post-Installation Configuration 4

4 Post-Installation Configuration


• 4.1 Settings the Cloud Administrator's Email (p. 51)

• 4.2 Configuring a SMTP Server for Cloud Bursting Quota Alerts (p. 52)

• 4.3 Configuring a SMTP Server for Monitoring Alerts (p. 53)

• 4.4 Opening Ports (p. 54)

• 4.5 Setting up the Analytics Environment (p. 55)

• 4.6 Configuring Analytics (p. 56)

• 4.7 Logging into Control (p. 59)

• 4.8 Granting Access Privileges (p. 60)

• 4.9 Adding an HPC Cluster (p. 63)

Configuration required to complete the installation of Control.

Control Administrator's Guide4 Post-Installation Configuration p.51

4.1 Settings the Cloud Administrator's EmailSpecify the Cloud Administrator's email address so that cloud bursting alerts can be sent to that emailaddress.

1. Open a supported browser and navigate to http://HOSTNAME:9980/pclm/.

Where HOSTNAME is the hostname or IP address of the machine where Cloud is installed.

2. Login using the default Cloud Administrator credentials.

username = admin@altair and password = Altair@123

3. Navigate to to http://HOSTNAME:9980/pclm/#/user-profile.

Where HOSTNAME is the hostname or IP address of the machine where Cloud is installed.

4. For E-mail, enter the email address where cloud bursting alerts should be sent.

5. Click Save.


http://pbscloud-server:9980/pclm/

http://pbscloud-server:9980/pclm/#/user-profile


4.2 Configuring a SMTP Server for Cloud BurstingQuota AlertsSet up a SMTP server so that when a cloud bursting resource alert limit is reached an email is sent tothe Cloud Administrator.

The following steps must be executed as root or as a user with sudo permissions using the sudocommand.

1. Create an environment variable for the Cloud version:export CLOUD_VERSION=2018.4


• If you installed Cloud using the Control installer and installed in the default location:export KARD_PATH=\/var/spool/pbsworks/$CLOUD_VERSION/pbscontrol/home/cloudio/kard

• If you installed Cloud using the Control installer and installed in a non-default location:export KARD_PATH=DIR/cloudio/kard

Replace DIR with the path to the Control home directory.



3. Navigate to the directory $KARD_PATH/current.

4. Create/edit the file meta.yml and add the following lines to the file:smtp: host: SMTP_HOST_URL port: SMTP_PORT login: SMTP_LOGIN password: SMTP_PASSWORD

The SMTP Server port number defaults to 25.

Note: The value of login and password can be left blank if a username and passwordare not necessary to use the SMTP server.

5. Save the file.

6. Launch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v $KARD_PATH/config:/config \-v $KARD_PATH/kard:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash

7. Update the kard:pkr kard update

8. Update the Docker context:pkr kard make




4.3 Configuring a SMTP Server for Monitoring AlertsConfigure a SMTP server so that Monitoring email alerts are sent to specified users.

• The following steps must be executed as root or as a user with sudo permissions using the sudocommand.

A new feature allows alerts to be sent via email to specified users when a nodes goes offline, CPUutilization drops below or rise above a specified percentage, or when a cluster becomes unavailable.These alerts are configured via the Monitoring tab.

If a Postfix SMTP server is installed on the machine where Control is installed and it is running on port25, then email notifications will be sent for the subscribed alerts.

To use another SMTP server to send the emails, provide the hostname and port of the SMPT server andalso the authentication credentials.

1. Login to the machine hosting Control.

2. Edit the file PC_HOME/config/wsbridge/smtp.json.{ "host": "localhost", "port": 25}

3. Update this file with the appropriate SMTP server information.{ "host": "smpt.gmail.com", "port": 587, "user": "[email protected]", "pass": "password"}

4. Save the file.

5. Restart Control:pbs-control-manager restart



4.4 Opening PortsOpen ports for communication with external locations.

Some ports need to be opened for communication with external locations and should be open on yourfirewall to allow for this communication.


2. Open the following ports:

• 5601 - ELK Stack

• 9980 - Cloud UI

• 19980 - CloudFlow

3. Login to the machine where the Web Application is installed.

4. Open the following port:

• 9500 - Control Gateway

5. Login to the machine where the Data Collector is installed.

6. Open the following port:

• 9343 - Analytics Data Collector

7. If Control has been deployed across several machines (distributed deployment):

a) Login to each machine where Control components have been distributed.

b) Open the following ports:

• 9820 - Message router service (for AMQP clients)

• 9835 - Message router database service

• 9840 - RabbitMQ service

• 9850 - Framework HTTP Orchestrator service

• 9860 - Framework database service



4.5 Setting up the Analytics EnvironmentCopy preserved Analytics configuration files and reset the Analytics database password.

Before copying any Analytics configuration files that were preserved prior to uninstalling the olderversion of PBSA/Analytics, review the configuration files automatically created by Analytics in thefollowing location:

PC_HOME/config/pbsa/services/pbsa/parser/modules/pbsdata/config

1. Login to the machine where Analytics is installed.

2. Stop Analytics:systemctl stop pbsworks-pbsa

3. Copy the Analytics configuration files that were preserved prior to uninstalling the older version ofPBSA/Analytics to the directory:

PC_HOME/config/pbsa/services/pbsa/parser/modules/pbsdata/config

Warning: This will overwrite any files generated automatically by Analytics.

4. Update the system path to include the path to the Analytics Commands:export PATH=\$PATH:PC_EXEC/pbsa/scripts/pbsa-commands/bin:PC_EXEC/dc/scripts/pbsacommands/bin

5. Reset the Analytics database password from the default of “pbsworks@123” to a secure site-specific password.pbsa-data-password OLD_PASSWORD NEW_PASSWORD

6. Start Analytics:systemctl start pbsworks-pbsa



4.6 Configuring AnalyticsConfigure holidays, prime and non-prime time hours, exit codes, and custom groups.

Configuration is no longer necessary prior to starting Analytics. However, configuring allows you as anadministrator to supply information so that richer more meaningful charts can be produced.

4.6.1 Configuring HolidaysConfigure site specific holidays.

By default, Analytics is configured with a single holiday, New Year’s day, for the years 2011 through2018.

Use the pbsa-config-holidays command to add a holiday:

Note: Default entries may be removed using the pbsa-config-holidays --remove option.

Any changes to the holiday information will be incorporated in the next parsing cycle.

1. Login to the machine hosting Analytics.

2. Execute the command:pbsa-config-holidays {HOLIDAY|WEEKDAY} [options]

Execute the following command to add a holiday that takes place over a single day.

pbsa-config-holidays MemorialDay --start-date=2016-05-25 --enddate=2016-05-25

Execute the following command to add a holiday that takes place over multiple days.

pbsa-config-holidays Thanksgiving --start-date=2016-11-26 --end-date=2016-11-27

Execute the following command to add a holiday that takes place in a previous year.

pbsa-config-holidays Thanksgiving --start-date=2014-11-27 --end-date=2014-11-28

4.6.2 Configure Prime and Non-Prime HoursAdd prime and non-prime time hours for a weekday.

By default, Analytics is configured with prime and non-prime time hours for the years 2011 through2018. Prime times hours for Monday through Friday begin at 09:00 am. Non-prime time hours begin at5:30 pm. For Saturday and Sunday, the entire day is scheduled as non-prime time.

Use the pbsa-config-holidays command to add prime and non-prime hours per WEEKDAY:

Note: Default entries may be removed using the pbsa-config-holidays --remove option.

Any changes to the prime/non-prime time information will be incorporated in the next parsing cycle.




2. Execute the command:pbsa-config-holidays {HOLIDAY|WEEKDAY} [options]

Execute the following command to configure prime time to start at 9:00 am and non-prime tostart at 6:00 pm for Monday.

pbsa-config-holidays monday --prime-time=09:00 --non-prime-time=18:00

Execute the following command to configure prime time to start at 8:00 am and non-primetime to start at 5:30 pm for Monday. Prime time start defaults to 8:00 am and non-prime startdefaults to 5:30 pm (17:30).

pbsa-config-holidays monday

Execute the following command to configure prime and non-prime time hours for any weekdays(Monday-Friday) not specified separately. For instance, if prime time and non-prime time havealready been specified for the days Monday and Tuesday, then this command will specify primetime and non-prime time for the days Wednesday through Friday.

pbsa-config-holidays weekdays --prime-time=09:00 --non-prime-time=18:00

Execute the following command to configure prime time to start at 9:00 am and non-prime timeto start at 6:00 pm for weekends (Saturday and Sunday).

pbsa-config-holidays weekends --prime-time=09:00 --non-prime-time=18:00

Execute the following command to configure prime time to start at 9:00 am and non-prime tostart at 6:00 pm for Monday for the year 2014 (a previous year’s configuration).

pbsa-config-holidays monday --prime-time=09:00 --non-prime-time=18:00 --start-date=2014-01-01

4.6.3 Configuring Exit CodesConfigure site specific exit codes.

Analytics maps exit codes to a reason to track the success or failure of a job. Analytics is configured bydefault with two exit reasons “Successful” and “Unsuccessful” that are mapped to specific exit codes.

Use the pbsa-config-exits command to map an exit code to a reason.

Analytics is configured with a default group “Still-Running”. Do not delete this default exit code.

If changes are made to exit codes after the initial parsing cycle is complete, any parsed data must bediscarded and all PBS Professional accounting logs have to be reparsed.


2. Execute the command:pbsa-config-exits {EXIT_CODE[,EXIT_CODE] | EXIT_CODE-EXIT_CODE} [options]



Execute the following command to map an exit code of “1” to the reason “Successful” for theapplication PBSJobApp.

pbsa-config-exits 1 --reason=Successful --application=PBSJobApp

Execute the following command to map a list of exit codes to th default reason “Unspecified” forall applications (default).

pbsa-config-exits 2,143,255

4.6.4 Configuring Custom GroupsAdd a user to a custom group to track utilization at a project level.

Custom groups are used to track utilization at a project level for sites running versions of PBSProfessional prior to 11.2. When a user is added to a custom group, they are essentially being addedto a project. The preferred method for tracking utilization at a project level is the use of the -P optionwhen submitting a job via the PBS Professional qsub command.

Analytics is configured with a default group “OTHERS” for the years 2011 through 2018. Do not deletethis default group.

Use the pbsa-config-groups command to add a user to a custom group.

Once the accounting logs are parsed for the first time, modifications to the current year's informationshould only be done in two situations and do not require any reparsing of accounting logs.

1. A user is assigned to a new project.

Use the pbsa-config-groups command to assign the user to the new project. Set the start date totoday's date or later. The end date can be set to a future date or "now".

2. A user completes an assignment.

Use the pbsa-config-groups command to update the user’s completion date by setting the end date tothe assignment completion date.

If the custom group information is modified for any other reason, then discard any parsed data andreparse all PBS Professional accounting logs.


2. Execute the command:pbsa-config-groups USER CUSTOMGROUP [options]

Several engineers may be assigned to a project. To track the total number of jobs executed forthat project, Analytics must be configured so that it knows which users are assigned to thatproject.

The following command configures a user “tsmith” to be assigned to a custom group (i.e.,project) “CFD” from January 22, 2016 to November 23, 2016.

pbsa-config-groups tsmith CFD --start-date=2016-01-22 --end-date=2016-11-23



4.7 Logging into ControlLogin to Control using the URL https://<hostname>:9500/pc.

1. Open a supported web browser.

2. Enter the URL https://HOSTNAME:9500/pc.

Where HOSTNAME is the hostname or IP address of the machine where the Control Web Interfacecomponent is installed.

The Control login screen is displayed.

3. Enter your credentials in the username and password fields.

4. Click Log In.



4.8 Granting Access PrivilegesEstablish access privileges for users other than the Control Service User.

Grant access privileges to the various features of Control by using the new Access Control feature. Youcan either add a user to one of the default roles or you can create a new role and add users to it.

Note: Users must be valid NIS/SSH users.

4.8.1 Default Roles and PrivilegesOverview of the default access control roles.

Access Control RolesTwo roles are available by default: Manager and Operator. These roles cannot be deleted and theirassigned privileges cannot be changed, however additional users can be added to these roles.

By default, the Service User entered during the installation of Control is assigned to the Manager roleand cannot be removed from this role.

ManagerA Manager has the highest level of access privilege. Managers can create cloud accounts andbursting scenarios, can view and act upon jobs and nodes, can configure the HPC workloadmanager, can view and modify analysis charts, can run simulations, and can grant and restrictaccess to the various features of Control.

OperatorOperators can view nodes and jobs information and can view alerts.

PrivilegesBy default, the following privilege levels are available:

Cloud AdminThis privilege allows full access to the features available via the Cloud tab, allowing the user tocreate cloud accounts and bursting scenarios.

Monitor ViewerThis privilege allows view-only access to the features available via the Monitor tab, allowing theuser to view jobs and node information and alerts.

Monitor AdminThis privilege allows full access to the features available via the Monitor tab, allowing the user tonot only view jobs and nodes information, but also to act upon the job and nodes. For example,placing a job on hold, deleting a job, or taking a node offline. Alerts can also be created.

Configure AdminThis privilege allows full access to the features available via the Configure tab, allowing the userto change HPC workload manager settings and parameters.



Analyze AdminThis privilege allows full access to the features available via the Analyze tab, allowing the user toview, modify, and create analysis charts.

Simulate AdminThis privilege allows full access to the features available via the Simulate tab, allowing the user torun simulations and review simulation results.

When a user is not given privileges to one of the available components of Control (Cloud, Configure,Monitor, Analyze, Simulate), then the associated tab is not displayed. For example, the Operator roleby default is given only Monitor Viewer privileges. A user added to the Operator role will only have viewaccess to the Monitor tab after logging in. The other tabs are not displayed.

Please note that access to certain features available via the Monitor and Configure tab are alsocontrolled by access controls defined at the PBS Professional level. This is dependent upon thecredentials that are used to add the cluster. Once a cluster is added, subsequent administrative actionsare performed as the credentials utilized while adding the cluster. For example, if the user used toconnect to the HPC cluster has PBS Professional User privileges, then that user will not be able toconfigure the HPC workload manager settings via the Configure tab or take a node offline via theMonitor tab.

4.8.2 Adding a New RoleCreate a custom role for your site.

Review the default roles and privileges before adding a new role.

Two roles are available by default: Manager and Operator. Add additional roles per your site's needs.

1. Click .

2. Click Roles from the menu located on the left-hand side of the web page.

3. Click Add Role.A role is created with a default role name and undefined privileges.

4. Click located to the right of the name of the role.

a) For Role Name, enter a name that describes the role.

b) Click OK.

5. Choose the privileges to give to this role by clicking located to the right of the privilege.

Privileges are not mutually exclusive, so you may assign more than one.

6. Click Save.The new role is displayed in the Roles list.

4.8.3 Assigning a Role to a UserAssign a role to the user to establish user privileges.

1. Click .



2. Click Users from the menu located on the left-hand side of the web page.

3. Select a user by enabling the checkbox next to the user's name.

Tip: Select multiple users when you want to assign the same roles to multiple users.

Figure 5: Select a User

4. Click Assign Roles.

5. Enable the checkbox next to the Role Name to assign the role to the user.

Tip: More than one role can be assigned to the user.

6. Click OK.

4.8.4 Changing the Privileges of a RoleAdd or remove access privileges for a role.

Review the default roles and privileges before changing the privileges of a role.

You cannot change the access privileges of the default Manager or Operator roles.

1. Click .

2. Click Roles from the menu located on the left-hand side of the web page.

3. Click the name of the role.

Figure 6: Edit a Role

4. Choose one of the following options to change the privileges given to the role:

• Click located to the right of the privilege to enable the privilege.

• Click located to the right of the privilege to disable the privilege.

Privileges are not mutually exclusive, so you may assign more than one.

5. Click Save.



4.9 Adding an HPC ClusterConnect to an HPC cluster so that you can monitor and administer the cluster.

To monitor jobs and nodes via the Monitor tab and configure Workload Manager settings via theConfigure tab, you must first add an HPC cluster. Cluster management for the Monitor services and theConfigure services are mutually exclusive, therefore you must manage the clusters for those servicesindependently - i.e, if you add a cluster via the Monitor tab, that cluster is not available in the Configuretab and vice versa.

To access and connect to an HPC cluster, the cluster must first be added to Control. When you add acluster, the hostname or IP address of the machine hosting the PBS Server must be provided, and thenyou can connect to the HPC by one of two options:

1. without providing the credentials of a user.

Note: Control assumes that SSH key-based authentication, between the machinewhere the Control services are installed and the machine hosting the PBS Server,has been configured for the user account that is the Control Service User (as definedduring the installation of Control). Control attempts to connect to the machine hostingthe PBS Server using the Service User username and SSH key. If the username or SSHkey cannot be verified, an error message is displayed and the credentials of a validuser on the machine hosting the PBS Server must be entered (Option 2).

2. by providing the credentials of a valid user on the machine hosting the PBS Server.

Once the cluster is added, subsequent administrative actions are performed as the credentials utilizedwhile adding the cluster. It is recommended to add the cluster using the credentials of someone withPBS Manager privileges. Connecting to a cluster using non-managerial credentials will severely limit theadministrative actions that can be performed, however monitoring capabilities will not be affected.

Note: If the machine hosting the PBS Server has been configured to run SSH on a non-default port (default is 22), additional information is necessary to add the cluster.

1. Choose one of the following options:

• Click the Monitor tab to monitor and manage HPC cluster jobs and nodes.

• Click the Configure tab to configure an HPC by changing Workload Manager settings, queuesand resources.

2. Click Manage Clusters.A dialog box appears where information about the cluster is entered.



Figure 7: Manage Clusters

3. Enter the name of the cluster as it will appear in the Control web application.

The cluster name must be 20 characters or less.

Tip: Use the unique name supplied during the installation of the Analytics DataCollector so that analytics data is tracked separately for PBS Complexes having thesame Fully Qualified Domain Name (FQDN).

4. Enter the cluster's host name by entering its:

• IP address - a four-part number string such as 18.72.0.3.

• Host name - a text string such as hpc.altair.com.

5. Click Add.

Control attempts to add the cluster as the Control Service User using an SSH key. If the usernameor SSH key is not valid or the machine hosting the PBS Server has been configured to run SSH ona non-default port, follow these steps:

a) Click Show advanced options.

Figure 8: Additional Information Required

b) Enter the SSH port that the machine hosting the PBS Server listens on.

Change the SSH port if the machine hosting the PBS Server has been configured to run SSHon a non-default port.

c) Enter the username and password of an authenticated user on the machine hosting the PBSServer.



d) Click Add.

The cluster is added.


• Repeat Steps 3 through 5 to add another cluster.

• Click Close to close the Manage Clusters dialog box.


Uninstalling Control 5

5 Uninstalling Control


• 5.1 Stopping Bursted Cloud Nodes (p. 67)

• 5.2 Uninstalling Control (p. 68)

• 5.3 Uninstalling the Data Collector (p. 69)

• 5.4 Uninstalling Cloud after a Manual Installation (p. 70)

Control Administrator's Guide5 Uninstalling Control p.67

5.1 Stopping Bursted Cloud NodesStop burst cloud nodes before Control is uninstalled.

Bursted cloud nodes must be stopped before Control can be uninstalled.

1. Delete all jobs from the cloud bursting queues using the qdel command.

2. Login to Control.

3. Click Cloud.

4. Click the Bursting tab on the left-hand side of the web page.

5. Click on the name of the cloud bursting scenario.

6. Determine if there are any cloud burst nodes by viewing the message displayed under theMachines heading.


• If you see the following message, it indicates that there are no cloud burst nodes. Continueonto the next step.No machines are available

• If you do not see the following message, wait until it appears and then continue to the nextstep.No machines are available

8. Disable the cloud bursting hook.



5.2 Uninstalling ControlUninstall all Control components, services and third-party packages.

Uninstallation must be done as root or as a user with sudo permissions using the sudo command.


Note: All services are automatically stopped during the uninstallation process.

1. Login to the machine where Control is installed.




The following message is displayed:

Would you like to uninstall? (Yes/No):



Enter Y and press ENTER.

5. Enter Yes to uninstall the product and press ENTER.

It may take several minutes to uninstall Control.

All services, files and third-party packages are removed, except for the pbswinstall_op_report file.



5.3 Uninstalling the Data CollectorUninstall the Analytics Data Collector.

Uninstallation must be done as root or as a user with sudo permissions using the sudo command.


Note: All services are automatically stopped during the uninstallation process.





The following message is displayed:

Would you like to uninstall? (Yes/No):


There was a non-fixable broken installation!! Would you like to uninstall? (Y/N)

Enter Y and press ENTER.

5. Enter Yes to uninstall the product and press ENTER.

It may take several minutes to uninstall the Data Collector.

All services, files and third-party packages are removed, except for the pbswinstall_op_report

6. Repeat the above steps for all Data Collectors.



5.4 Uninstalling Cloud after a Manual InstallationRemove an instance of Cloud when it has been manually installed.

• You must stop Cloud before uninstalling it.

• Uninstallation must be done as root or as a user with sudo permissions using the sudo command.

Only use the following procedure when Cloud was installed manually. If Cloud was installed using theControl installer, then use the Control installer to remove Cloud.



3. Connect to the pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v DIR/config:/config \-v DIR/kard:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash


4. Clean Cloud services:pkr clean

5. Purge Cloud images:pkr image purge

6. Remove pkr container:exitdocker image rm pbscloudio.azurecr.io/pkr:$CLOUD_VERSIONrm -rf DIR/configrm -rf DIR/kard



Starting and Stopping Services 6

6 Starting and Stopping Services


• 6.1 Service Commands (p. 72)

• 6.2 Control Service Commands (p. 73)

• 6.3 Cloud Service Commands (p. 76)

• 6.4 Analytics Service Commands (p. 79)

Control Administrator's Guide6 Starting and Stopping Services p.72

6.1 Service CommandsService commands for starting and stopping Control.

All Control services are stopped and started using a single command, except for the services associatedwith the Analytics and Cloud components. Those must be stopped and started separately.



6.2 Control Service Commands

6.2.1 Starting ControlStart all Control services, except those associated with the Cloud and Analytics components.

Starting the Control services must be done as root or as a user with sudo permissions using the sudocommand.

Important: The services for the Cloud and the Analytics components must be startedseparately.


2. Enter the following command to start Control:pbs-control-manager start

6.2.2 Stopping ControlStop all Control services, except those associated with the Cloud and Analytics components.

Stopping the Control services must be done as root or as a user with sudo permissions using the sudocommand.

Important: The services for the Cloud and the Analytics components must be stoppedseparately.


2. Enter the following command to stop Control:pbs-control-manager stop

6.2.3 Restarting ControlRestart all Control services, except those associated with the Cloud and Analytics components.

Retarting the Control services must be done as root or as a user with sudo permissions using the sudocommand.

Important: The services for the Cloud and the Analytics components must be restartedseparately.


2. Enter the following command to restart Control:pbs-control-manager restart



6.2.4 Determining the Status of all Control ServicesDetermine whether a Control service is up or down.

Important: A separate command must be executed to determine the status of the Cloudand the Analytics components.


2. Enter the following command to display the status of each Control service:pbs-control-manager status

Messages similar to the following are displayed:pbsmonitor-postgres.service: PBS Monitor Server systemctl-status:loaded active exitedpbsmonitor.service: PBS Monitor Server systemctl-status:loaded active exitedpbswadmin-mongod.service: PBS Works Administrator MongoDB Daemon systemctl-status:loaded active runningpbswadmin-orchestrator.service: PBS Works Service Framework HTTP API systemctl-status:loaded active runningpbswadmin-worker.service: PBS Works Service Framework worker manager systemctl-status:loaded active runningpbswauth.service: PBS Works Auth Server systemctl-status:loaded active exitedpbsworks-messagerouter-http.service: Message-Router Jobs HTTP API systemctl-status:loaded active runningpbsworks-messagerouter-mongod.service: PBSWorks Message Router MongoDB Daemon systemctl-status:loaded active runningpbsworks-messagerouter-rabbitmq.service: RabbitMQ Server for use by Jobs Service (MessageRouter) systemctl-status:loaded active runningpbsworks-nginx.service: PBSWorks NGINX HTTP and reverse proxy server systemctl-status:loaded active runningpbsworks-notification-mongod.service: PBS Works Notification Server MongoDB Daemon systemctl-status:loaded active runningpbsworks-notification.service: PBS Works WebSocket Notification Server systemctl-status:loaded active runningpcgui.service: PBS Control GUI Server systemctl-status:loaded active runningService pbswadmin-launch-prep is not loaded

Note: It is normal for the pbswadmin-launch-prep service to be down as it runsonce, just prior to other services coming up. It prepares the environment and getseverything ready for the other services to run and then shuts down.



When a service is not running the following message is displayed:Service <service> is not loaded.

where <service> is the name of the service that is down.



6.3 Cloud Service Commands

6.3.1 Start and Stop Cloud After a Manual Installation

Starting CloudDocker and pkr instructions for starting Cloud.

Starting Cloud must be done as root or as a user with sudo permissions using the sudo command.

When your server hosting the Cloud component reboots, containers are restarted automatically. If youneed to manually start Cloud containers, please follow below instructions.



3. Launch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v DIR/config:/config \-v DIR/kard:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash


4. Start Cloud:pkr start


Stopping CloudDocker and pkr instructions for stopping Cloud.

Stopping Cloud must be done as root or as a user with sudo permissions using the sudo command.





4. Stop Cloud:pkr stoppkr clean




6.3.2 Service Commands After Installing Using theInstaller

Starting CloudStart the Cloud service.

Starting the Cloud service must be done as root or as a user with sudo permissions using the sudocommand.

1. Login to the machine where the Cloud component is installed.

2. Enter the following command to start Cloud:systemctl start pbscloud

Stopping CloudStop the Cloud service.

Stopping the Cloud component must be done as root or as a user with sudo permissions using the sudocommand.


2. Enter the following command to stop Cloud:systemctl stop pbscloud

Restarting CloudRestart the Cloud service.

Restarting the Cloud service must be done as root or as a user with sudo permissions using the sudocommand.


2. Enter the following command to restart Cloud:systemctl restart pbscloud

Determining the Status of CloudDetermine whether Cloud is up or down.


2. Enter the following command to display the status of Cloud:



systemctl status pbscloud

Message similar to the following are displayed when the service is up and running:● pbscloud.service - PBSCLOUD Loaded: loaded (/etc/systemd/system/pbscloud.service; disabled; vendor preset: disabled) Active: active (running) since Tue 2018-10-02 12:32:28 EDT; 12s ago Main PID: 15536 (docker) Tasks: 12 Memory: 6.8M CGroup: /system.slice/pbscloud.service └─15536 /usr/bin/docker run --rm -v /run/docker.sock:/run/docker.sock -v /var/spool/pbsworks/2018.4/pbscont... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: [139B blob data] Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting configurator-api ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting mistral-api ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting keeper ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting backend ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting graphql ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting hubble ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: [87B blob data] Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting pacioli ... Oct 02 12:32:41 tr-pbsw183-01 docker[15536]: Starting bootstrapper-api ...

Messages similar to the following are displayed when the service is stopped:● pbscloud.service - PBSCLOUD Loaded: loaded (/etc/systemd/system/pbscloud.service; disabled; vendor preset: disabled) Active: inactive (dead) Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping mongodb ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping salt ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping cadvisor ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping node-exporter ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping rabbitmq ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping elasticsearch ... Oct 02 12:30:48 tr-pbsw183-01 docker[10749]: Stopping postgres ... Oct 02 12:31:08 tr-pbsw183-01 docker[10749]: [257B blob data] Oct 02 12:31:13 tr-pbsw183-01 docker[10749]: [63B blob data] Oct 02 12:31:13 tr-pbsw183-01 systemd[1]: Stopped PBSCLOUD.



6.4 Analytics Service Commands

6.4.1 Starting the Analytics ServiceStart Analytics.

Starting Analytics must be done as root or as a user with sudo permissions using the sudo command.


2. Enter the following command to start Analytics.systemctl start pbsworks-pbsa


4. Enter the following command to start the Data Collector.systemctl start pbsworks-dc

5. Repeat step 3 and 4 for all Data Collectors.

6.4.2 Stopping the Analytics ServiceStop Analytics.

Stopping Analytics must be done as root or as a user with sudo permissions using the sudo command.


2. Enter the following command to stop Analytics.systemctl stop pbsworks-pbsa


4. Enter the following command to stop the Data Collector.systemctl stop pbsworks-dc


6.4.3 Restarting the Analytics ServiceRestart Analytics.

Retarting Analytics must be done as root or as a user with sudo permissions using the sudo command.


2. Enter the following command to restart Analytics.systemctl restart pbsworks-pbsa


4. Enter the following command to restart the Data Collector.systemctl restart pbsworks-dc




6.4.4 Determining the Status of the Analytics ServiceDetermine whether Analytics is up or down.


2. Enter the following command to display the status of Analytics.systemctl status pbsworks-pbsa

Message similar to the following are displayed when the service is up and running: Loaded: loaded (/etc/rc.d/init.d/pbsworks-pbsa; bad; vendor preset: disabled) Active: active (exited) since Mon 2018-10-01 14:54:55 EDT; 21h ago Docs: man:systemd-sysv-generator(8) Process: 1394 ExecStart=/etc/rc.d/init.d/pbsworks-pbsa start (code=exited, status=0/SUCCESS) Tasks: 0 Memory: 0B

Messages similar to the following are displayed when the service is stopped;● pbsworks-pbsa.service - LSB: Altair PBSWorks Login server creates customer login sessions Loaded: loaded (/etc/rc.d/init.d/pbsworks-pbsa; bad; vendor preset: disabled) Active: inactive (dead) since Tue 2018-10-02 11:55:53 EDT; 38min ago Docs: man:systemd-sysv-generator(8) Process: 29955 ExecStop=/etc/rc.d/init.d/pbsworks-pbsa stop (code=exited, status=0/SUCCESS) Process: 4029 ExecStart=/etc/rc.d/init.d/pbsworks-pbsa start (code=exited, status=0/SUCCESS)

Oct 02 11:55:40 tr-pbsw183-01 pbsworks-pbsa[29955]: Successfully stopped ActiveMQ server[ OK ]Oct 02 11:55:40 tr-pbsw183-01 pbsworks-pbsa[29955]: Stopping MonetDB server PID 4201Oct 02 11:55:40 tr-pbsw183-01 su[30766]: (to amitt) root on noneOct 02 11:55:52 tr-pbsw183-01 su[31046]: (to amitt) root on noneOct 02 11:55:52 tr-pbsw183-01 pbsworks-pbsa[29955]: Successfully stopped MonetDB server[ OK ]Oct 02 11:55:52 tr-pbsw183-01 pbsworks-pbsa[29955]: Stopping MongoDB server PID 4125Oct 02 11:55:52 tr-pbsw183-01 su[31064]: (to amitt) root on noneOct 02 11:55:53 tr-pbsw183-01 su[31081]: (to amitt) root on noneOct 02 11:55:53 tr-pbsw183-01 pbsworks-pbsa[29955]: Successfully stopped MongoDB server[ OK ]Oct 02 11:55:53 tr-pbsw183-01 systemd[1]: Stopped LSB: Altair PBSWorks Login server creates customer login sessions.


4. Enter the following command to display the status of the Data Collector.systemctl status pbsworks-dc

Messages similar to the following are displayed when the service is up and running:● pbsworks-dc.service - LSB: Altair PBSWorks Data Collector Loaded: loaded (/etc/rc.d/init.d/pbsworks-dc; bad; vendor preset: disabled) Active: active (running) since Wed 2018-09-26 23:04:34 EDT; 5 days ago Docs: man:systemd-sysv-generator(8) Tasks: 33



Memory: 217.4M CGroup: /system.slice/pbsworks-dc.service └─12610 /opt/altair/pbsworks/2018.4/pbscontrol/exec/dc/thirdparty/java/bin/java -XX:ErrorFile=/opt/altair/p...

Sep 26 23:04:30 tr-pbsw183-01 systemd[1]: Starting LSB: Altair PBSWorks Data Collector...Sep 26 23:04:30 tr-pbsw183-01 pbsworks-dc[12567]: Starting PBSWorks Data CollectorSep 26 23:04:34 tr-pbsw183-01 pbsworks-dc[12567]: PBSWorks Data Collector started, pid: 12610[ OK ]Sep 26 23:04:34 tr-pbsw183-01 systemd[1]: Started LSB: Altair PBSWorks Data Collector.

Messages similar to the following are displayed when the service is stopped;● pbsworks-dc.service - LSB: Altair PBSWorks Data Collector Loaded: loaded (/etc/rc.d/init.d/pbsworks-dc; bad; vendor preset: disabled) Active: inactive (dead) since Tue 2018-10-02 12:37:15 EDT; 5s ago Docs: man:systemd-sysv-generator(8) Process: 26277 ExecStop=/etc/rc.d/init.d/pbsworks-dc stop (code=exited, status=0/SUCCESS)

Sep 26 23:04:30 tr-pbsw183-01 systemd[1]: Starting LSB: Altair PBSWorks Data Collector...Sep 26 23:04:30 tr-pbsw183-01 pbsworks-dc[12567]: Starting PBSWorks Data CollectorSep 26 23:04:34 tr-pbsw183-01 pbsworks-dc[12567]: PBSWorks Data Collector started, pid: 12610[ OK ]Sep 26 23:04:34 tr-pbsw183-01 systemd[1]: Started LSB: Altair PBSWorks Data Collector.Oct 02 12:37:04 tr-pbsw183-01 systemd[1]: Stopping LSB: Altair PBSWorks Data Collector...Oct 02 12:37:15 tr-pbsw183-01 pbsworks-dc[26277]: Stopping PBSWorks Data Collector PID 12610: [ OK ]Oct 02 12:37:15 tr-pbsw183-01 systemd[1]: Stopped LSB: Altair PBSWorks Data Collector.



Analytics Commands 7

7 Analytics Commands


• 7.1 Configure Nodes (p. 83)

• 7.2 Configure Holidays (p. 88)

• 7.3 Configure Custom Groups (p. 94)

• 7.4 Configure Exit Codes (p. 98)

• 7.5 Configuring the Data Collector (p. 102)

• 7.6 Configuring the Parser (p. 105)

• 7.7 Resetting the Analytics Dataset (p. 108)

• 7.8 Changing the Analytics Database Password (p. 112)

• 7.9 Creating a Diagnosis File for Troubleshooting Issues (p. 115)

Control Administrator's Guide7 Analytics Commands p.83

7.1 Configure Nodes

Namepbsa-config-nodes

Syntaxpbsa-config-nodes [ options ] HOSTNAME [ HOSTNAME_FAILOVER ] FQDN FILE

DescriptionConfigure Analytics with node details for the PBS Professional Complex represented by HOSTNAME (asspecified in the PBS Professional accounting logs), the failover server HOSTNAME_FAILOVER, and the FQDN(fully qualified domain name) using the FILE containing the output from the pbsnodes-av command.

Note: If the value of HOSTNAME, HOSTNAME_FAILOVER, or FQDN contains a hyphen (-) or aperiod (.), it must be replaced with an underscore (_).

You can set the dates for when the node configuration is in effect using the options --start-date and--end-date.

This command must be executed as root or as a user with sudo privileges using the sudo command onthe machine where Analytics is installed.

Options--config-dir=DIRECTORY

Specify the DIRECTORY where the output of this command is directed. Defaults to PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config.

--end-date=YYYY-MM-DD|nowDefine the effective end date for this PBS Professional node configuration. Default is now whichdenotes that this configuration is currently in effect.

--furnishing-servers=FUR_SERVER[,FUR_SERVER]Specify a comma separated list of FUR_SERVERs when peering is configured.

--helpDisplay help for the command. This option cannot be used in conjunction with any other options.

--newOverwrite any existing node details configuration.

--showDisplay the current configuration on the screen for review. Use this option in conjunction withthe --start-date option to view a configuration from a previous year. Default is to display thecurrent year’s configuration.

--start-date=YYYY-MM-DDDefine the effective start date for this PBS Professional node configuration. Default is the currentdate.



--versionDisplay the command's version number. This option cannot be used in conjunction with any otheroptions.

ExamplesAll examples assume that the PBS Professional command pbsnodes -av has been executed and theoutput from this command has been written to a file called pbsnodes.out:

Configure Current Node Details Using Defaults

Configure node detail information that is effective today and will continue to be in effect goingforward (start date defaults to today and end date defaults to now ).

pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out

Configure a Previous Year’s Node Details pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --start-date=2014-01-01 --end-date=2014-12-31

Configure a Future Year’s Node Details pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --start-date=2022-01-01 --end-date=2022-12-31

Configure Node Details for a Site Implementing Peer Scheduling

Note: Peer scheduling requires a list of furnishing servers.

pbsa-config-nodes Bangalore bangalore.ms.com pbsnodes.out --furnishing-servers=Troy,MV

Redirect the Output of the Command to an Alternate Location

Redirect the output of the command to an alternate location allowing the output of thecommand to be saved without overwriting the contents of the default configuration directory.

pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --config-dir=/tmp/pbsa-config-dir

Refresh the Node Details

Overwrite the current node details. Use this command when you need to discard the currentconfiguration and start afresh.

pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --new

Show Previous Year’s Node Detailspbsa-config-nodes --start-date=2014-01-01 --show



Error Messages

Table 3: Error Messages

Message Text Description

The PBSWORKS_COMMAND_HOME orPBSWORKS_SERVER_HOME variable was not set, andI was unable to find or open a pbsworks-pbsa.conffile.

The environment variablesPBSWORKS_COMMAND_HOME andPBSWORKS_SERVER_HOME have not beenexported to the environment. Contactyour systems administrator to set up theenvironment properly.

Error occurred while settingPBSWORKS_COMMAND_HOME orPBSWORKS_SERVER_HOME


Python version XX is not supported. Supportedversion is 2.4.x or higher.

Please update the Python version to 2.4 orhigher.

Error occurred while get uid for user XXXX. Unable to get user id for current user. Contactyour system administrator to report this error.

Operation not permitted, Please run command asXXXX user or ROOT.

The user does not have permissions to run thiscommand. Please run the command as theService User specified during installation or asroot user.

memory unit XXXX is not supported An error occurred while processing thepbsnodes -av output during the conversion ofmemory units to GB. The memory unit did notequal one of the following valid options: b, kb,mb, gb, tb, or pb.

The configuration file is empty. This error may occur after executing thiscommand using the --show option. Theconfiguration file is empty, therefore nothinghas been configured yet.

Invalid furnishing server names FUR_SERVERS must be a comma separated listof furnishing servers.

Please check the pbsnodes -av output file, some ofthe nodes are offline. Following nodes are offline:[‘xxxxxxxxxx’]

Please investigate the reason for the downednode and either correct the issue within PBSProfessional, or if this is a planned outage,




then manually edit the pbsnodes -av outputfile and modify the state of the node to anonline status, or delete the node.

Incorrect number of arguments. Try pbsa-config-nodes --help.

Please review the usage information forthis command. You can obtain the usageinformation by using the --help option.

Config directory does not exists. Path XXXX. The configuration directory PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config does not exist. An error may haveoccurred during installation. Please re-installthe product.

Unable to instantiate the PBSA ToolKit. Please checkyour PBSWORKS_SERVER_HOME variable.

The environment variablePBSWORKS_SERVER_HOME may not havebeen exported to the environment. Contactyour systems administrator to set up theenvironment properly.

Unable to read the "pbsnodes -av" output file for userXXXX.

Path: XXXX.

The pbsnodes -av output file does not exist.

Unable to read the "pbsnodes -av" output file for userXXXX.

Please check file permissions.

Incorrect file permissions for opening thepbsnodes -av output file for read access.

Unable to open the “nodedetails_<year>.conf ” file. Please check the existence and permissionsof the nodedetails_YYYY.conf file locatedat PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config.

The startdate XXXX provided is not valid. Value of --start-date must be in the formatYYYY-MM-DD.

The enddate XXXX provided is not valid. Value of --end-date must be in the formatYYYY-MM-DD or now.

The startdate XXXX provided must not be greaterthan end date XXXX.

Invalid date range. Value of --start-datemust be less than the value of --end-date.

The startdate year XXXX and enddate year XXXXmust be the same.

Specified date range must not span overseveral years.




Start date format is incorrect. Expected is YYYY-MM-DD.

Dates should be in the format YYYY-MM-DD.



7.2 Configure Holidays

Namepbsa-config-holidays

Syntaxpbsa-config-holidays [ options ] { HOLIDAY | { monday | tuesday | wednesday | thursday |friday | saturday | sunday | weekdays | weekends } }

DescriptionAdd a HOLIDAY or configure prime and non-prime time hours for a day of the week, or for any weekdaysthat have not already been defined, or for weekends (Saturday and Sunday).

You can set the dates for the holiday using the options --start-date and --end-date.

You can set prime and non-prime time hours using the options --prime-time and --non-prime-time.

By default, a single holiday, New Year’s day, is configured for the years 2011 through the current year.For Monday through Friday, prime times hours begin at 8:00 am, and non-prime time hours begin at5:30 pm. For Saturday and Sunday, the entire day is scheduled as non-prime time.




--end-date=YYYY-MM-DD|nowDefine the end date of the holiday. Default is the current date.


--newOverwrite any existing node details configuration.

--non-prime-time=HH:MM|all|none

Define the non-prime time start time. Default is 17:30.

HH:MM designates the start time of non-prime time hours in a 24 hour format.

all designates that all hours are non-prime time hours.

none designates that all hours are prime time hours.

--prime-time=HH:MM|all|none

Define the prime time start time. Default is 08:00.

HH:MM designates the start time of prime time hours in a 24 hour format.



all designates that all hours are prime time hours.

none designates that all hours are non-prime time hours.

--removeRemove a HOLIDAY or prime and non-prime time hours for a day of the week, or for any weekdaysthat have not already been defined, or for weekends.

--showDisplay the current configuration on the screen for review. Use this option in conjunction with the--start-date option to view a configuration from a previous year. Default is to display the currentyear’s configuration.

--start-date=YYYY-MM-DD

Define the start date of the holiday. This option can also be used to configure prime/non-primetime hours for a previous year’s configuration. Default is the current date.


Examples

Configure a Holiday that Occurs Over a Single Daypbsa-config-holidays MemorialDay --start-date=2016-05-25 --end-date=2016-05-25

Configure a Holiday that Occurs Over Multiple Dayspbsa-config-holidays Thanksgiving --start-date=2016-11-26 --end-date=2016-11-27

Configure Prime Time for 9:00 am and Non-Prime for 6:00 pm on Mondaypbsa-config-holidays monday --prime-time=09:00 --non-prime-time=18:00

Configure Prime and Non-Prime Time Hours for Tuesday Using the Defaultspbsa-config-holidays tuesday

Configure Prime Time for 9 am and Non-Prime Time for 6 pm for the Weekendpbsa-config-holidays weekends --prime-time=09:00 --non-prime-time=18:00

Configure Prime and Non-Prime Hours for Remaining Weekdays

Configure prime and non-prime time hours for any weekdays that have not already beenconfigured separately. For example, if prime and non-prime time hours have already beenconfigured for Monday and Tuesday, then this command will specify prime and non-prime timehours for the weekdays Wednesday through Friday.

pbsa-config-holidays weekdays --prime-time=09:00 --non-prime-time=18:00



Configure Prime and Non-prime Hours for Monday for a Previous Year pbsa-config-holidays monday --start-date=2014-01-01 --prime-time=09:00 --non-prime-time=18:00

Remove a Holidaypbsa-config-holidays MemorialDay --remove

Remove Prime and Non-Prime Time Hours for a Specific Daypbsa-config-holidays monday --remove



pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --config-dir=/tmp/pbsa-config-dir

Refresh the Holidays and Prime and Non-Prime Time Hours

Overwrite the current holidays and prime and non-prime time hours. Use this command whenyou need to discard the current configuration and start afresh.

pbsa-config-nodes hwesuse113-01 hwesuse113-01.ms.com pbsnodes.out --new

Show the Holidays and Prime and Non-Prime Time Hours for a Previous Yearpbsa-config-holidays --start-date=2014-01-01 --show

Error Messages






The environment variablesPBSWORKS_COMMAND_HOME andPBSWORKS_SERVER_HOME have not beenexported to the environment. Contact




your systems administrator to set up theenvironment properly.






Either Holiday or prime and non-prime config can bespecified at a time.

Configuring a holiday and prime/non-primetime hours cannot be done with a singlecommand. First issue the command toconfigure the holidays, and then issue a secondcommand to configure the prime/non-primetime hours.


Invalid holiday name. Holiday name must containonly any alphanumeric character and the underscore.

Please enter a valid holiday name whichcontains only alphanumeric characters and/orthe underscore character.

Invalid option --remove in the command. Cannot specify the --remove option whenusing the --new option.

Please define all weekdays (Mon-Fri) separately as"weekdays" value is removed.

The default entry for weekdays has beenremoved therefore, you must now defineprime and non-prime time hours for Mondaythrough Friday separately. For instance, defineprime and non-prime time hours by issuing thefollowing command for each day of the week(Monday through Friday):

pbsa-config-holidays DAY --prime-time=HH:MM --non-prime-time=HH:MM

Please define both weekends (Sat-Sun) separately as"weekends" value is removed.

The default entry for weekends has beenremoved therefore, you must now specifyprime and non-prime time hours for Saturdayand Sunday separately. For instance, define the




prime and non-prime time hours by issuing thefollowing commands:

pbsa-config-holidays Saturday --prime-time=HH:MM --non-prime-time=HH:MM

pbsa-config-holidays Sunday --prime-time=HH:MM --non-prime-time=HH:MM

Please make sure both weekend days (Sat & Sun) aredefined.

Warning message that requests that prime andnon-prime time hours be configured for bothSaturday and Sunday.

Incorrect number of arguments. Try pbsa-config-holidays --help.

Please review the usage information forthis command. You can obtain the usageinformation by using the --help option.

Config directory does not exists. Path XXXX. The configuration directory PC_EXEC/portal/services/pbsa/parser/modules/pbsdata/config does not exist. An error may haveoccurred during installation. Please re-installthe product.



Unable to open the “holiday_<year>.conf ” file. Please check the existence and permissionsof the holiday_YYYY.conf file located atPC_EXEC/portal/services/pbsa/parser/modules/pbsdata/config

The startdate XXXX provided is not valid. Value of --start-date must be in the formatYYYY-MM-DD.

The enddate XXXX provided is not valid. Value of --end-date must be in the formatYYYY-MM-DD or now.

The startdate XXXX provided must not be greaterthan end date XXXX.

Invalid date range. Value of --start-datemust be less than the value of --end-date.

The startdate year XXXX and enddate year XXXXmust be the same.

Specified date range must not span overseveral years.




Start date format is incorrect. Expected is YYYY-MM-DD.


The primetime XXXX provided is not valid. Value of --prime-time must be in the formatof HH:MM in the 24 hour clock, all, or none.

The nonprimetime XXXX provided is not valid. Value of --non-prime-time must be in theformat of HH:MM in the 24 hour clock,all, ornone.



7.3 Configure Custom Groups

Namepbsa-config-groups

Syntaxpbsa-config-groups [ options ] USER CUSTOMGROUP

DescriptionAdd a USER to a CUSTOMGROUP. Custom groups are used to track utilization at a project level for sitesrunning versions of PBS Professional prior to 11.2. When a user is added to a custom group, they areessentially being added to a project.

The preferred method for tracking utilization at a project level is the use of the -P option whensubmitting a job via the qub command. However, for older versions of PBS Professional, use pbsa-config-groups to define a project (custom group).

You can set the date for when a user begins working on project and when the user stops working on theproject using the options --start-date and --end-date.

By default, a custom group “OTHERS” is configured for the years 2011 through the current year.




--end-date=YYYY-MM-DD|nowDefine the date when the user stopped working on the project. Default is now which denotes thatthe user is still currently working in the group.


--newOverwrite any existing custom group configuration.

--showDisplay the current configuration on the screen for review. Use this option in conjunction withthe --start-date option to view a configuration from a previous year. Default is to display thecurrent year’s configuration.

--start-date=YYYY-MM-DDDefine the date when the user began working on the project. Default is the current date.




Examples

Assign a User to a Custom Group By Designating a Start and End Date

Assign a user (tsmith) to a custom group (CFD) from January 22, 2016 to November 23, 2016.

pbsa-config-groups tsmith CFD --start-date=2016-01-22 --end-date=2016-11-23

Assign a User to a Custom Group Using the Defaults

Assign a user to a custom group with a start date of today’s date and an end date of nowdenoting that the user is currently assigned to that custom group.

pbsa-config-groups tsmith CFD

Assign a User to a Custom Group Using a Default End Date

Assign a user (tsmith) to a custom group (CFD) from January 1, 2016 with an end date of nowdenoting that the user is currently assigned to that custom group.

pbsa-config-groups tsmith CFD --start-date=2016-01-01

Assign a User to Several Custom Groups

Assign a user (tsmith) to a custom group (CFD) from January 1, 2016 to June 1, 2016, and toanother custom group (CAD) from January 1, 2016 to September 1, 2016.

pbsa-config-groups tsmith CFD --start-date=2016-01-01--end-date=2016-06-01pbsa-config-groups tsmith CAD --start-date=2016-01-01 --end-date=2016-09-01



pbsa-config-groups tsmith CFD --start-date=2016-01-01 --end-date=2016-06-01 --config-dir=/tmp/pbsa-config-dir

Refresh the Custom Groups

Overwrite the currently defined custom groups. Use this command when you need to discardthe current configuration and start afresh.

pbsa-config-groups tsmith CFD --start-date=2016-01-22 --end-date=2016-11-23 --new




Error Messages



The PBSWORKS_COMMAND_HOME orPBSWORKS_SERVER_HOME variablewas not set, and I was unable to find oropen a pbsworks-pbsa.conf file.

The environment variables PBSWORKS_COMMAND_HOMEand PBSWORKS_SERVER_HOME have not been exported tothe environment. Contact your systems administrator to setup the environment properly.



Python version XX is not supported.Supported version is 2.4.x or higher.

Please update the Python version to 2.4 or higher.

Error occurred while get uid for userXXXX.

Unable to get user id for current user. Contact your systemadministrator to report this error.

Operation not permitted, Please runcommand as XXXX user or ROOT.

The user does not have permissions to run this command.Please run the command as the Service User specified duringinstallation or as root user.

The configuration file is empty. This error may occur after executing this command usingthe --show option. The configuration file is empty, thereforenothing has been configured yet.

Incorrect number of arguments. Trypbsa-config-groups --help.

Please review the usage information for this command. Youcan obtain the usage information by using the --help option.

Invalid option --remove in thecommand.

Cannot specify the --remove option when using the --newoption.

WARN: User "system" can't beremoved.

The default custom group “OTHERS” cannot be deleted.Analytics requires this entry.

Config directory does not exists. PathXXXX.

The configuration directory PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config does not exist. Anerror may have occurred during installation. Please re-installthe product.

Unable to instantiate the PBSAToolKit. Please check yourPBSWORKS_SERVER_HOME variable.

The environment variable PBSWORKS_SERVER_HOME maynot have been exported to the environment. Contact yoursystems administrator to set up the environment properly.




Unable to open the“custom_group_<year>.conf ” file.

Please check the existence and permissions of thecustom_group_YYYY.conf file located at PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config.

The startdate XXXX provided is notvalid.

Value of --start-date must be in the format YYYY-MM-DD.

The enddate XXXX provided is notvalid.

Value of --end-date must be in the format YYYY-MM-DD ornow.

The startdate XXXX provided must notbe greater than end date XXXX.

Invalid date range. Value of --start-date must be less thanthe value of --end-date.

The startdate year XXXX and enddateyear XXXX must be the same.

Specified date range must not span over several years.

Start date format is incorrect. Expectedis YYYY-MM-DD.




7.4 Configure Exit Codes

Namepbsa-config-exits

Syntaxpbsa-config-exits [ options ] { EXIT_CODE [ , EXIT_CODE ] | EXIT_CODE - EXIT_CODE }

DescriptionMap an EXIT_CODE to a REASON. Use the option --reasonto specify the REASON.

By default, three exit reasons “Successful”, “Unsuccessful” and "Still Running" are mapped to specificexit codes. Use the --show option to view the defaults.


Options--application=APPLICATION

Categorize the exit codes by APPLICATION or solver. APPLICATION directly correlates to theapplication name specified via the PBS Professional “software” attribute.

Default is All applications.

--config-dir=DIRECTORYSpecify the DIRECTORY where the output of this command is directed. Defaults to PC_HOME/portal/services/pbsa/parser/modules/pbsdata/config.


--newOverwrite any existing exit code configuration.

--reason=REASONSpecify the exit reason. Default is Unspecified.

--removeRemove the specified exit reason/application from the current configuration. REASON defaults toUnspecified and APPLICATION defaults to All.

--showDisplay the current configuration on the screen for review.




Examples

Assign a User to a Custom Group By Designating a Start and End Date

Assign a user (tsmith) to a custom group (CFD) from January 22, 2016 to November 23, 2016.pbsa-config-groups tsmith CFD --start-date=2016-01-22 --end-date=2016-11-23

Assign a User to a Custom Group Using the Defaults

Assign a user to a custom group with a start date of today’s date and an end date of nowdenoting that the user is currently assigned to that custom group.

pbsa-config-groups tsmith CFD

Assign a User to a Custom Group Using a Default End Date

Assign a user (tsmith) to a custom group (CFD) from January 1, 2016 with an end date of nowdenoting that the user is currently assigned to that custom group.

pbsa-config-groups tsmith CFD --start-date=2016-01-01

Assign a User to Several Custom Groups

Assign a user (tsmith) to a custom group (CFD) from January 1, 2016 to June 1, 2016, and toanother custom group (CAD) from January 1, 2016 to September 1, 2016.

pbsa-config-groups tsmith CFD --start-date=2016-01-01--end-date=2016-06-01pbsa-config-groups tsmith CAD --start-date=2016-01-01 --end-date=2016-09-01



pbsa-config-groups tsmith CFD --start-date=2016-01-01 --end-date=2016-06-01 --config-dir=/tmp/pbsa-config-dir

Refresh the Custom Groups

Overwrite the currently defined custom groups. Use this command when you need to discardthe current configuration and start afresh.

pbsa-config-groups tsmith CFD --start-date=2016-01-22 --end-date=2016-11-23 --new




Error Messages



The PBSWORKS_COMMAND_HOME orPBSWORKS_SERVER_HOME variablewas not set, and I was unable to find oropen a pbsworks-pbsa.conf file.











WARN: User can't update privatesettings for exitreason: "Still Running"

The default entry “Still Running” cannot be deleted. Analyticsrequires this entry.

Invalid option --remove in thecommand.

Cannot specify the --remove option when using the --newoption.

Incorrect number of arguments. Trypbsa-config-exits --help..


Config directory does not exists. The Analytics configuration directory does not exist. An errormay have occurred during installation of Analytics. Please re-install the product.






Unable to open the “exit_status.conf ”file.

Unable to open the “exit_status.conf ” file. Please check theexistence and permissions of the exit_status.conf file locatedat PC_HOME/config/pbsa/services/pbsa/parser/modules/pbsdata/config.



7.5 Configuring the Data Collector

Namepbsa-config-dc

Syntaxpbsa-config-dc [ options ]

DescriptionConfigure the Analytics Data Collector. If the Data Collector is currently running, then it will need to berestarted after issuing this command.

The Data Collector makes the PBS Professional accounting logs and cluster node information accessibleto Analytics. It copies the accounting logs and node information from your site’s PBS Professionalaccounting log location to an intermediate holding area every 5 minutes.

You can change the time interval, however this is only recommended for advanced users of Analytics.The size of your site’s PBS Professional complex and the number and size of the accounting logs willdictate the length of time required for copying the accounting logs.

This command must be executed as root or as a user with sudo privileges using the sudo command onthe machine where the Data Collector is installed.

Options--collection-interval=MM

Define the collection interval MM in minutes for collecting the PBS Professional accounting logs(copying them to an intermediate area prior to transporting to the machine where Analytics isinstalled).

Default is 5 minutes.

--config-dir=DIRECTORYSpecify the DIRECTORY where the output of this command is directed.






Examples

Configure the Data Collector Transport and Collection Interval

configures the Data Collector to collect the PBS Professional accounting logs and cluster nodeinformation and copy them to an intermediate area every 50 minutes.

pbsa-config-dc --collection-interval=50



pbsa-config-dc --collection-interval=50 --config-dir=/tmp/pbsa-config-dir

Error Messages



The DC_SERVER_HOME orDC_CONF_FILE variable was not set,and I was unable to find or open apbsworks-dc.conf file.

The environment variables DC_SERVER_HOME andDC_CONF_FILE have not been exported to the environment.Contact your systems administrator to set up theenvironment properly.

Error occurred while settingDC_SERVER_HOME.

The environment variable DC_SERVER_HOME has notbeen exported to the environment. Contact your systemsadministrator to set up the environment properly.



Operation not permitted, Please runcommand as XXXX user.


Option XXXX argument XXXX must bein minutes.

Try dc-config --help

Then specified option must be expressed in minutes.






Operation not permitted, Please runcommand as

XXXX user or ROOT.


Please specify --collection-interval=COLLECTION_INTERVALoption. Try dc-config --help.

You must specify a collection interval.




Unable to read the dc config file. Please check the existence and permissions of the config.xmland/or modules.xml files located at PC_HOME/config/dc/config.



7.6 Configuring the Parser

Namepbsa-config-parser

Syntaxpbsa-config-parser [ options ]

DescriptionThe Parser reads the PBS Professional accounting logs, parses the appropriate information from theaccounting logs, and then loads this information into a database used by Analytics. Use the pbsa-config-parser command to configure the Parser. Analytics must be restarted after issuing thiscommand.

You can configure when the daily Parser cycle will run. Currently this cycle should not be run more thanonce in a 24-hour period. By default, the Parser cycle is configured to run once a day at 1:00 am.



Specify the DIRECTORY where the output of this command is directed.


--job-duration=DAYS

Define the maximum number of DAYS a running job will be considered for parsing, beyondwhich the Parser will write the PBS Professional “S” (running job) record to the unsupporteddirectory. For example, if DAYS is set to thirty, then a job that runs for more than thirty dayswill not be parsed and will be written to the unsupported directory. The unsupported directoryis where accounting logs records that fail the validation criteria. When an accounting log recordfails validation, the entry is written to file in a unsupported directory. A unsupported directoryis created for each Data Collector. The unsupported directories are located in: PC_EXEC/pbsa/services/pbsa/parser/modules/pbsdata.

--parser-timer=TIMER

Define when the daily parsing cycle should run. TIMER is a Quartz Job Scheduling Library Cronexpression specified between double quotes. For example, "0 0 1 * * ?" specifies that the Parsingcycle should run at 1:00 am every day. According to the documentation at http://www.quartz-scheduler.org, “Cron-Expressions are strings that are actually made up of seven sub-expressions,that describe individual details of the schedule. These sub-expression are separated with white-space, and represent:

• Seconds


http://www.quartz-scheduler.org

http://www.quartz-scheduler.org


• Minutes

• Hours

• Day-of-Month

• Month

• Day-of-Week

• Year (optional field)”



Examples

Configure the Parser

Configure the Parser to run the daily parsing cycle at 2:00 am.

pbsa-config-parser --parser-timer="0 0 2 * * ?" --job-duration=60



pbsa-config-parser --parser-timer="0 0 2 * * ?" --config-dir=/tmp/pbsa-config-dir

Error Messages















Option XXXX argument XXXX must be number. Trypbsa-config-parser --help

The specified option must be expressed as anumeric.

Please specify at least one option either of--parser-timer=PARSER_TIMER, or --jobduration=JOB_DURATION.

You must specify at least one of the followingoptions:

--parser-timer

--job-duration


Config directory does not exists. The Analytics configuration directory doesnot exist. An error may have occurred duringinstallation of Analytics. Please re-install theproduct.



Unable to read the parser config file. Parser configuration file does not exist. Contactyour system administrator.



7.7 Resetting the Analytics DatasetDiscard all Analytics data, reset the Data Collectors and reset the database password to its defaultvalue.

Namepbsa-data-reset

Syntaxpbsa-data-reset [ options ]

DescriptionUse the pbsa-data-reset command to:

• discard the data from the Analytics database.

• delete the history files indicating which PBS Professional accounting log files have already beentransported to the machine where Analytics is installed.

• reset all Data Collectors.

• reset the Analytics database password back to “pbsworks@123”.

Note: It is recommended to use the pbsa-data-password command to change thisdefault password to a site specific one.

Stop Analytics before issuing this command. If this command is executed when the Analytics serviceand database are running, then this command will exit with an error.

Note: All simulation and model data must be removed when resetting the Analyticsdataset. See Resetting the Analytics Accounting Log Data for information on how to deletesimulation and model data.


Options--help

Display help for the command. This option cannot be used in conjunction with any other options.




Error Messages



ThePBSWORKS_COMMAND_HOMEor PBSWORKS_SERVER_HOMEvariable was not set, and Iwas unable to find or open apbsworks-pbsa.conf file.

The environment variables PBSWORKS_COMMAND_HOME andPBSWORKS_SERVER_HOME have not been exported to theenvironment. Contact your systems administrator to set up theenvironment properly.


The environment variables PBSWORKS_COMMAND_HOME andPBSWORKS_SERVER_HOME have not been exported to theenvironment. Contact your systems administrator to set up theenvironment properly.

Python version XX is notsupported. Supported version is2.4.x or higher.


Error occurred while get uid foruser XXXX.


Operation not permitted, Pleaserun command as XXXX user orROOT.


Error in reading the inputs.

Please check log file: XXXX

An error occurred while retrieving system arguments. Contactyour system administrator to report this error.

Property with key XXXX doesn'texists.

An error occurred while reading the database configurationfile monetdb- pbsa.conf. Contact your system administrator toreport this error.

File XXXX must end with(.backup)

An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

File XXXX doesn't exists. An error occurred while reading the DBConnection.propertiesfile. Contact your system administrator to report this error.

Error in reading monetdb-pbsa.conf file

An error occurred while reading the database configuration filemonetdb-pbsa.conf. Contact your system administrator to reportthis error.





Updation of config filedatasources.xml failed.

An error occurred while updating the configuration filedatasources.xml. Contact your system administrator to reportthis error.

Updation of config file .monetdbfailed.

An error occurred while updating the configuration

file monetdb-pbsa.conf. Contact your

system administrator to report this error.

Updation of config file monetdb-user.sql failed.

An error occurred while updating the file

monetdb-user.sql. Contact your system administrator

to report this error.

Error in splitting command: XXXX An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

Error in executing command:XXXX


Command splitting failed: XXXX An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

Server startup failed. An error occurred upon database server start up. Contact yoursystem administrator to report this error.

Server status setting failed. An error occurred while setting the database server status.Contact your system administrator to report this error.

Server status check failed. An error occurred while setting the database server status.Contact your system administrator to report this error.

Datacollectors load historydeletion failed.

An error occurred while deleting the Data Collector load historyfile history.xml. Contact your system administrator to report thiserror.

Parser availability directorydeletion failed.

An error occurred while deleting the Parser availability directory.Contact your system administrator to report this error.




PBSA parser cleaning failed. An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

Data reset failed.



Config directory does not exists. The Analytics configuration directory does not exist. An errormay have occurred during installation of Analytics. Please re-install the product.

Unable to instantiate the PBSAToolKit. Please check yourPBSWORKS_SERVER_HOMEvariable.

The environment variable PBSWORKS_SERVER_HOME may nothave been exported to the environment. Contact your systemsadministrator to set up the environment properly.

Please make sure PBSA engineand Monetdb Server is notrunning.

Cannot reset the data as the Analytics service and database arerunning. Stop the Analytics service.



7.8 Changing the Analytics Database PasswordDiscard all Analytics data, reset the Data Collectors and reset the database password to its defaultvalue.

Namepbsa-data-password

Syntaxpbsa-data-password [ options ] { OLD_PASSWORD NEW_PASSWORD }

DescriptionUse the pbsa-data-password command to change or reset the Analytics database password fromOLD_PASSWORD to NEW_PASSWORD.

A database superuser is created when Analytics is installed having the following credentials: usernameis "pbsworks" and password is "pbsworks@123".

Stop Analytics before issuing this command. If this command is executed when the Analytics serviceand database are running, then this command will exit with an error.


Options--help



Examples

Reset the Analytics database password

Change the current password “pwdJan2018” to a new password “pwd-Feb2018”.

pbsa-data-password pwdJan2018 pwdFeb2018

Error Messages



The PBSWORKS_COMMAND_HOME orPBSWORKS_SERVER_HOME variable

The environment variables PBSWORKS_COMMAND_HOMEand PBSWORKS_SERVER_HOME have not been exported to




was not set, and I was unable to find oropen a pbsworks-pbsa.conf file.

the environment. Contact your systems administrator to setup the environment properly.



Incorrect number of arguments. Trypbsa-data-password --help.








Error in reading the inputs. Pleasecheck log file: XXXX

An error occurred while retrieving system arguments.Contact your system administrator to report this error.

Property with key XXXX doesn't exists. An error occurred while reading the database configurationfile monetdb- pbsa.conf. Contact your system administratorto report this error.

File XXXX must end with (.backup) An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

File XXXX doesn't exists. An error occurred while reading the DBConnection.propertiesfile. Contact your system administrator to report this error.

Error in reading monetdb-pbsa.conf file


An error occurred while reading the database configurationfile monetdb-pbsa.conf. Contact your system administrator toreport this error.

Error in splitting command: XXXX An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.




Error in executing command: XXXX An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

Line must be (key=value). xxxxxxx There is a problem with the DBConnection.properties file.Contact your system administrator to report this error.

Updation of config file context.xmlfailed.

An error occurred while updating the configuration filecontext.xml. Contact your system administrator to report thiserror.

Updation of config file .monetdb failed. An error occurred while updating the configuration filemonetdb-pbsa.conf. Contact your system administrator toreport this error.

Updation of config file monetdb-user.sqlfailed.

An error occurred while updating the file monetdb-user.sql.Contact your system administrator to report this error.

Command splitting failed: XXXX An error occurred while resetting the dataset. Contact yoursystem administrator to report this error.

Server startup failed. An error occurred upon database server start up. Contactyour system administrator to report this error.

Server status setting failed. An error occurred while setting the database server status.Contact your system administrator to report this error.

Server status check failed. An error occurred while setting the database server status.Contact your system administrator to report this error.

Password change failed. Please checklog file: XXXX

An error occurred while changing the password. Contact yoursystem administrator to report this error.

Please make sure PBSA engine andMonetdb Server is not running.

Cannot reset the data as the Analytics service and databaseare running. Stop the Analytics service.



7.9 Creating a Diagnosis File for TroubleshootingIssuesDiscard all Analytics data, reset the Data Collectors and reset the database password to its defaultvalue.

Namepbsa-setup-diagnosis

Syntaxpbsa-setup-diagnosis [ options ]

DescriptionThis command creates a zip file in the current working directory containing information that can assistin the resolution of system issues. This file can be emailed to Altair Systems Support to help with thediagnosis of an issue.


Options--help



Error Messages



Error occurred while findingconfiguration Path <FILENAME>

The Analytics configuration file denoted by <FILENAME>cannot be found. Contact your systems administrator.

Error while creating temporarydirectory

An error occurred while creating the zip file. Contact yoursystems administrator.

Error while read source <DIRECTORY>.Please make sure directory exist anduser have proper permission

An error occurred while creating the zip file. Contact yoursystems administrator.


Cloud Bursting 8

8 Cloud Bursting


• 8.1 Cloud Bursting (p. 117)

• 8.2 Cloud Bursting Terminology (p. 119)

• 8.3 Prerequisites for Cloud Bursting (p. 120)

• 8.4 Cloud Bursting Features (p. 121)

• 8.5 PCLM CLI Client (p. 123)

• 8.6 Installing and Configuring a Bind DNS Server (p. 124)

• 8.7 Cloud Bursting Startup Script (p. 128)

• 8.8 Configuring Amazon Web Service Cloud Bursting (p. 132)

• 8.9 Configuring Microsoft Azure Cloud Bursting (p. 143)

• 8.10 Configuring Google Cloud Platform Cloud Bursting (p. 157)

• 8.11 Configuring Oracle Cloud Platform Cloud Bursting (p. 165)

• 8.12 Configuring Orange Cloud Flexible Engine for Cloud Bursting (p. 179)

• 8.13 Configuring HUAWEI Cloud for Cloud Bursting (p. 190)

• 8.14 Configuring Open Telekom Cloud for Cloud Bursting (p. 200)

• 8.15 Testing Cloud Bursting with the CLI (p. 210)

• 8.16 Configuring PBS Professional for Cloud Bursting (p. 212)

Control Administrator's Guide8 Cloud Bursting p.117

8.1 Cloud BurstingCloud bursting is a configuration that is set up between an HPC cluster and a public cloud to deal withpeaks in cluster demand. When resource capacity at the HPC reaches a certain point, the demand isdirected to a public cloud so there is no interruption of services.

Cloud bursting requires that you already have an account with one of the supported cloud providers:

• Microsoft Azure

• Amazon Web Services (AWS)

• Google Cloud Platform (GCP)

• Oracle Cloud Platform

• Orange Cloud Flexible Engine

• Open Telekom Cloud (OTC)

• HUAWEI Cloud

Cloud bursting is managed through PBS Professional by submitting jobs to one or more designatedcloud queues. A server periodic hook monitors cloud queues, estimates the demand for cloud nodes,and dynamically adjusts the number of nodes by bursting or unbursting as needed.

Figure 9: Typical Cloud Bursting Setup

Steps for Cloud BurstingThere is some setup required at the cloud provider level before a cloud bursting scenario can be createdusing Control. There is also PBS Professional configuration that must be done. General steps that arerequired for bursting:

1. Install a DNS server on the PBS Server.

2. Create a cloud-init script that runs after a cloud node is burst.



3. Create necessary cloud provider components such as a virtual network and a virtual machine.

4. SSH into the cloud provider virtual machine and configure the VM for your site's environment suchas installing PBS, installing cloud-init, mounting file systems, connecting it to the authenticationservice, installing any applications, etc.

5. Create an image of the virtual machine.

6. Provide the necessary cloud provider details, including the cloud-init script, by creating a burstingscenario with Control.

7. Configure PBS cloud bursting custom resources.

8. Creating PBS cloud queues.

9. Configure the PBS cloud bursting hook.

Note: Steps vary depending upon the cloud provider used for cloud bursting.



8.2 Cloud Bursting TerminologyDefinitions for terminology that is used to describe cloud bursting.

PCLM CLI

Cloud Command Line Interface. Client software that bursts, unbursts, and queries the statusof nodes in the cloud. The CLI can be used to test bursting at the command line and is alsoused as part of the cloud bursting hook configuration file. It is installed automatically with PBSProfessional 18.2.x. It must be installed manually for PBS Professional 14.2.6.

BurstThe action of adding a node in the cloud, and adding it to the PBS complex.

UnburstThe action of removing a node from both the PBS complex and the cloud.

Bursting scenarioScenario created in Control containing information about resources provided by cloud service thatare used for bursting.

ImageA pre-configured OS image in the cloud from which virtual machines can be instantiated.

InstanceA single virtual machine in a cloud computing environment.

Instance typeInstance characteristics: the combination of CPU, memory, storage, and networking capacity.

api_keyScenario token generated by Control at the time of a scenario creation. Identifies cloud serviceprovider and account.



8.3 Prerequisites for Cloud BurstingA secure network connection and DNS must be in place before cloud bursting can be enabled.

• An account with one of the supported cloud providers.

• PBS Works Burst Node licenses.

• An on-premise PBS complex; this includes the PBS Server and execution hosts. All the setuprequired for bursting is done on your on-premise PBS complex (it is not done on the PBS Serverinstalled with Control. The PBS Server installed with Control is only for running simulations).

• A DNS server on the PBS server host. Any node burst in the cloud must register with the DNSserver to identify itself in the complex. Any node that is created needs to be identified by itshostname; when a node is burst, the hook script registers the node with the DNS server.

• A dedicated network line, VPN setup, or SSH tunnel to the cloud provider for seamlesscommunication between the two.



8.4 Cloud Bursting FeaturesOverview of cloud bursting features.

Several features are available with cloud bursting:

Provide OS Image upon Job submissionThe first feature allows the OS image that is to be used when a cloud node is burst to be specified atjob submission via the qsub command. If this information is not provided at job submission, then theOS image defaults to the value of the cloud_default_image parameter in the cloud bursting hookconfiguration file.qsub -l select=1:ncpus=4 -q CLOUD_Q -v CLOUD_IMAGE=IMAGE JOB_SCRIPT

Where:

• CLOUD_Q is the name of the cloud queue.

• JOB_SCRIPT is the name of the script to be executed.

• IMAGE is the name of the OS image to be used when the node is burst.

Provide the Instance Type at Job SubmissionThe second feature allows the instance type that is burst to be specified at job submission via the qsubcommand (ex. Standard_DS3_v2). Instance types are defined by the cloud provider. The name of theinstance type must exactly match the cloud provider's.

If this information is not provided at job submission, then the instance type defaults to instance typedefined for the cloud queue.qsub -l select=1:ncpus=4 -q CLOUD_Q -v CLOUD_INSTANCE=INSTANCE_TYPE JOB_SCRIPT

Where:



• INSTANCE_TYPE is the name of cloud provider instance type. This is the type of machine that willbe burst.

Request InfiniBand NodesWhen the user wants to run HPC workloads like MPI jobs, the user needs InfiniBand supported nodesdeployed on the same InfiniBand network.

• The user can request InfiniBand nodes by requesting InfiniBand supported instance types.

• The user can request these nodes to be on the same InfiniBand network using the optionCLOUD_NETWORK=IB.

Both must be done when requesting InfiniBand nodes on the same InfiniBand network. Among thecloud providers, only Azure supports InfiniBand nodes today.qsub -l select=1:ncpus=4 -q CLOUD_Q \-v CLOUD_INSTANCE=INSTANCE_TYPE,CLOUD_NETWORK=IB JOB_SCRIPT

Where:




• INSTANCE_TYPE is the name of cloud provider instance type and should be an instance type thatsupports Infiniband.


To request an OS image, an instance type and InfiniBand nodes in a single qsub request:qsub -l select=1:ncpus=4 -q CLOUD_Q \-v CLOUD_IMAGE=IMAGE,CLOUD_INSTANCE=INSTANCE_TYPE,CLOUD_NETWORK=IB JOB_SCRIPT



8.5 PCLM CLI ClientA Command Line Interface client for bursting and unbursting cloud nodes without using PBSProfessional.

You can burst, unburst and monitor the status of your cloud bursting scenarios through a command lineinterface using the PCLM CLI client. You can burst one VM, several VMs of the same type, or severalVMs of different types in one single bursting call workflow. You can also troubleshoot the burstingscenario by examining the logs messages that are written to PC_HOME/server_logs.

The current version of the PCLM CLI client is 0.4.4. It is compatible with Cloud 2018.4. It has beentested on CentOS and Fedora and will run on any system supporting Python.

By default, PCLM CLI is installed at PBS_EXEC/bin/pclm for PBS Professional 18.2.x. Otherwise, you willneed to install the PCLM CLI manually.

8.5.1 Installing the PCLM CLI ClientSetup the prerequisites, install and test the PCLM CLI client application.

• Make sure Python 2.7 is installed on your system. Follow the instructions from the Installing Pythonpage to install it.

• Install pip. Follow the instructions from the Pip Installation page.

• Install virtualenv. Follow the instructions from the virtualenv Installation page.

Get the latest version of the PCLM CLI client. Setup Python and then use virtualenv and pip to setup thePCLM CLI client.

1. Login to the PBS Server.

2. Download the latest version of the PCLM CLI client.

3. Untar the pclm_cli archive.tar zxvf pclm_cli-VERSION.tgz

Where VERSION is the version of the PCLM CLI Client.

4. Navigate to the pclm_cli folder.cd pclm_cli_VERSION

5. Install, create and activate virtualenv.yum install python-virtualenvvirtualenv .pclm_cli-VERSIONsource .pclm_cli-VERSION/bin/activate

6. Install requirements.pip install -r requirements.txt

7. Install the PCLM CLI client.pip install

8. Test your installation.pclm --help


https://wiki.python.org/moin/BeginnersGuide/Download

https://pip.pypa.io/en/stable/installing/

https://virtualenv.pypa.io/en/stable/installation/


8.6 Installing and Configuring a Bind DNS ServerInstallation and configuration of bind must be done as root or as a user with sudo permissions using thesudo command.

For the below examples, assume the following:

• Fully qualified domain name (FQDN) of the PBS Server = pbs.altair.com

• NIC address of the PBS Server is 10.0.0.5 on the 10.0.0.0/24 network


2. Install BIND using one of the following options:

• For RHEL and CentOS:yum install bind

• For SLES and SUSE: zypper install bind

3. Edit the BIND configuration file /etc/named.conf.

4. Uncomment the following line by removing the # at the beginning of the line.#listen-on port 53 { 127.0.0.1; };

This line in the BIND configuration file tells BIND on which network interfaces and port to acceptclient queries.

5. Add the NIC address of the PBS Server.Example:listen-on port 53 { 127.0.0.1;10.0.0.5; };

The NIC address of the PBS Server should be accessible by all of the PBS execution hosts(compute nodes).

6. Comment out the below line:listen-on-v6 port 53 { ::1; };

Example:#listen-on-v6 port 53 { ::1; };

7. Uncomment the following line by removing the # at the beginning of the line:#allow-query { localhost; };

This line in the BIND configuration file defines the networks from which clients can post DNSrequests.

8. Add the PBS Server network:Example:allow-query { localhost;10.0.0.0/24; };

9. Add the following line to the bottom of the file:include "/etc/named/named.conf.local";

10. Save the file.

11. Make a directory where your site's forward and reverse zone files will be stored:mkdir /etc/named/zones



12. Navigate to that directory:cd /etc/named/zones

13. Create a forward zone file:vi db.DOMAINNAME

Where DOMAINNAME represents the domain name of the PBS Server.

Example:vi db.pbs.altair.com

14. Add the following lines to the file:

$TTL 604800@ IN SOA ns1.DOMAINNAME. admin.DOMAINNAME. ( 3 ; Serial 604800 ; Refresh 86400 ; Retry 2419200 ; Expire 604800 ) ; Negative Cache TTL;; name servers - NS records IN NS headnode.DOMAINNAME. ; PBS_NETWORK - A recordsheadnode.DOMAINNAME. IN A PBS_SERVER_IP_ADDR

Where DOMAINNAME represents the domain name of the PBS Server, PBS_SERVER_IP_ADDRrepresents the NIC address of the PBS Server, and PBS_NETWORK represents the PBS Servernetwork.

Example:$TTL 604800@ IN SOA ns1.pbs.altair.com. admin.pbs.altair.com. ( 3 ; Serial 604800 ; Refresh 86400 ; Retry 2419200 ; Expire 604800 ) ; Negative Cache TTL;; name servers - NS records IN NS headnode.pbs.altair.com. ; 10.0.0/24 - A recordsheadnode.pbs.altair.com. IN A 10.0.0.5

15. Save the file.

16. Create a reverse zone file:vi db.FIRST_THREE_OCTETS_OF_NETWORK_IP

Where FIRST_THREE_OCTETS_OF_NETWORK_IP represents the first three octets of the PBSServer network IP address:

Example:vi db.10.0.0

17. Add the following lines to the file:$TTL 604800@ IN SOA DOMAINNAME. admin.DOMAINNAME. ( 3 ; Serial



604800 ; Refresh 86400 ; Retry 2419200 ; Expire 604800 ) ; Negative Cache TTL; name servers IN NS headnode.DOMAINNAME. ; PTR Records5 IN PTR headnode.DOMAINNAME. ; PBS_SERVER_IP_ADDR

Where DOMAINNAME represents the domain name of the PBS Server and PBS_SERVER_IP_ADDRrepresents the NIC address of the PBS Server.

Example:$TTL 604800@ IN SOA pbs.altair.com. admin.pbs.altair.com. ( 3 ; Serial 604800 ; Refresh 86400 ; Retry 2419200 ; Expire 604800 ) ; Negative Cache TTL; name servers IN NS headnode.pbs.altair.com. ; PTR Records5 IN PTR headnode.pbs.altair.com. ; 10.0.0.5

18. Save the file.

19. Create a file that is used by BIND to determine where the forward and reverse zones file arestored:vi /etc/named/named.conf.local

20. Add the following to the file indicating the location of the forward zone file:zone "DOMAINNAME" { type master; file "/etc/named/zones/db.DOMAINNAME";};


Example:zone "pbs.altair.com" { type master; file "/etc/named/zones/db.pbs.altair.com";};

21. Add the following lines to the file indicating the location of the reverse zone file:zone "NETWORK_OCTETS_REVERSE_ORDER.in-addr.arpa" { type master; file "/etc/named/zones/db.FIRST_THREE_OCTETS_OF_NETWORK_IP"; };

Where FIRST_THREE_OCTETS_OF_NETWORK_IP represents the first three octets of the PBSServer network IP address and NETWORK_OCTETS_REVERSE_ORDER represents first three octetsof the network but in reverse order and .

Example:zone "0.0.10.in-addr.arpa" { type master; file "/etc/named/zones/db.10.0.0";



};

Network IP address is 10.0.0.0/24 so NETWORK_OCTETS_REVERSE_ORDER is 0.0.10

22. Save the file.

23. Restart BIND:service named restart

24. Edit /etc/resolv.conf.

Note: Configuring /etc/resolv.conf needs to be performed on all execution hosts.For the nodes that are burst in the cloud, configuring /etc/resolv.conf is done in thecloud-init script that is ran after a cloud is burst.

25. Add the following lines to the file:search DOMAINNAMEnameserver PBS_SERVER_IP_ADDR


Example:search pbs.altair.comnameserver 10.0.0.5

26. Save the file.

27. Query the DNS server for the IP address:nslookup DOMAINNAME


Example:nslookup pbs.altair.com

The command should return the IP address of DOMAINNAME.

28. Query the DNS server for the domain name:nslookup PBS_SERVER_IP_ADDR

Where PBS_SERVER_IP_ADDR represents the NIC address of the PBS Server.

Example:nslookup 10.0.0.5

The command should return the domain name of PBS_SERVER_IP_ADDR.



8.7 Cloud Bursting Startup ScriptCreate a script that is ran when the cloud node is burst.

IntroductionYour site will want to do some configuration to their cloud nodes after booting. For example, you maywant to install some packages, add users, or start services. A utility specifically designed for this iscloud-init. The cloud-init program is a bootstrapping utility for pre-provisioned disk images that run invirtualized environments, usually cloud-oriented services. Basically, it sets up the server instance to beusable when it’s finished booting. You must install cloud-init on your cloud provider VM to simplify thetask of configuring your instances on boot. For more information see cloud-init.

Several input types are supported by cloud-init.

• Shell scripts

• Cloud config files

The simplest way to configure an instance on boot is to use a shell script. This shell script is uploadedwhen a cloud bursting scenario is created. The shell script must begin with #! in order for cloud-init torecognize it as a shell script.

Below are examples of configuration that should be done via this script after a node has been burstin the cloud. These examples are not intended to be copied and pasted as is, you must configure thestartup script per your site's needs.

Example of a cloud-init Script#!/bin/sh# Configure the DNS resolverecho "resolv.conf setup"rm -f /etc/resolv.confecho "search DOMAINNAME" > /etc/resolv.confecho "nameserver PBS_SERVER_IP_ADDR" >> /etc/resolv.conf

# Map IP address to hostnames via /etc/hostsecho "/etc/hosts setup"rm -f /etc/hostsecho "127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4" > /etc/hostsecho "PBS_SERVER_IP_ADDR headnode headnode.DOMAINNAME" >> /etc/hosts

IP=ìp addr show eth0 | grep "inet\b" | awk '{print $2}' | cut -d/ -f1`HOST=ùname -nècho "$IP $HOST" >> /etc/hosts

# Disable NetworkManager so that it does not overwrite the /etc/resolv.conf filesystemctl disable NetworkManagersystemctl stop NetworkManagersystemctl enable networksystemctl start network

# Configure PBS via /etc/pbs.confecho "pbs setup"systemctl stop pbsrm -f /etc/pbs.confecho "PBS_EXEC=/opt/pbs/default" > /etc/pbs.conf


https://cloud-init.io/


echo "PBS_HOME=/var/spool/PBS" >> /etc/pbs.confecho "PBS_START_SERVER=0" >> /etc/pbs.confecho "PBS_START_MOM=1" >> /etc/pbs.confecho "PBS_START_SCHED=0" >> /etc/pbs.confecho "PBS_START_COMM=0" >> /etc/pbs.confecho "PBS_SERVER=PBS_SERVER_HOSTNAME" >> /etc/pbs.confecho "PBS_CORE_LIMIT=unlimited" >> /etc/pbs.confecho "PBS_SCP=/bin/scp" >> /etc/pbs.confecho "PBS_LEAF_ROUTERS=HOSTNAME,HOSTNAME" >> /etc/pbs.conf

# Configure the MoMecho "mom config setup". /etc/pbs.confecho "\$clienthost $PBS_SERVER" >> /var/spool/pbs/mom_priv/configecho "\$clienthost ${PBS_SERVER//.*}" >> /var/spool/pbs/mom_priv/configecho "\$restrict_user_maxsysid 999" >> /var/spool/pbs/mom_priv/config

# Restart pbssystemctl start pbs

An explanation for each section of the cloud-init script is given below. For the below examples assumethe following: Fully qualified domain name (FQDN) of the PBS Server = pbs.altair.com NIC address ofthe PBS Server is 10.0.0.5 on the 10.0.0.0/24 network. A DNS server has already been configured.

Configure the DNS ResolverThe DNS resolver allows applications running in the operating system to translate human-friendlydomain name into the numeric IP addresses that are required for access to resources on the local areanetwork or the Internet.

When a cloud node is provisioned, the cloud provider DNS information is automatically put in the /etc/resolv.conf file on the provisioned node. You then override the content of this file by providing thesearch domains and nameservers specific to your site in the cloud-init script.#!/bin/sh# Configure the DNS resolverecho "resolv.conf setup"rm -f /etc/resolv.confecho "search DOMAINNAME" > /etc/resolv.confecho "nameserver PBS_SERVER_IP_ADDR" >> /etc/resolv.conf


Example:

# Configure the DNS resolverecho "resolv.conf setup"rm -f /etc/resolv.confecho "search pbs.altair.com" > /etc/resolv.confecho "nameserver 10.0.0.5" >> /etc/resolv.conf

Add additional search domains, if you want to resolve hosts outside the PBS complex. If this informationis not present, cloud bursting will still work but hosts other than the PBS Server and burst nodes will notbe resolvable.

Configure the Host FileMap the hostnames defined when creating the DNS forward and reverse zone files to the PBS Server IPaddress by updating the /etc/hosts file.



# Map IP address to hostnames via /etc/hostsecho "/etc/hosts setup"rm -f /etc/hostsecho "127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4" > /etc/hostsecho "PBS_SERVER_IP_ADDR headnode headnode.DOMAINNAME" >> /etc/hosts

IP=ìp addr show eth0 | grep "inet\b" | awk '{print $2}' | cut -d/ -f1`HOST=ùname -nècho "$IP $HOST" >> /etc/hosts

Example:# Map IP address to hostnames via /etc/hostsecho "/etc/hosts setup"rm -f /etc/hostsecho "127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4" > /etc/hostsecho "10.0.0.5 headnode headnode.pbs.altair.com" >> /etc/hosts

IP=ìp addr show eth0 | grep "inet\b" | awk '{print $2}' | cut -d/ -f1`HOST=`hostname -sècho "$IP $HOST" >> /etc/hosts

#Tell NetworkManager to not overwrite the resolv.conf filesystemctl disable NetworkManagersystemctl stop NetworkManagersystemctl enable networksystemctl start network

Note: The command below may need to be modified depending on your site's operatingsystem. ip addr show eth0 | grep "inet\b" | awk '{print $2}' | cut -d/ -f1

Disable NetworkManager and Use Network InterfaceStop NetworkManager and use network interface so that the contents of /etc/resolv.conf do not getoverwritten:systemctl disable NetworkManagersystemctl stop NetworkManagersystemctl enable networksystemctl start network

Configure PBSUpdate the PBS configuration file /etc/pbs.conf:# Configure pbs.confecho "pbs setup"systemctl stop pbsrm -f /etc/pbs.confecho "PBS_EXEC=/opt/pbs" > /etc/pbs.confecho "PBS_HOME=/var/spool/pbs" >> /etc/pbs.confecho "PBS_START_SERVER=0" >> /etc/pbs.confecho "PBS_START_MOM=1" >> /etc/pbs.confecho "PBS_START_SCHED=0" >> /etc/pbs.confecho "PBS_START_COMM=0" >> /etc/pbs.confecho "PBS_SERVER=PBS_SERVER_HOSTNAME" >> /etc/pbs.confecho "PBS_CORE_LIMIT=unlimited" >> /etc/pbs.confecho "PBS_SCP=/bin/scp" >> /etc/pbs.confecho "PBS_LEAF_ROUTERS=HOSTNAME,HOSTNAME" >> /etc/pbs.conf



Where PBS_SERVER_HOSTNAME represents the hostname of the machine where the PBS Server isinstalled and HOSTNAME tells each endpoint which communication daemon it should talk to.

Note: If PBS is installed in non-default home and execution directory, then update thevalue of PBS_EXEC and PBS_HOME.

Configure the PBS MoM and Restart PBSUpdate the PBS_HOME/mom_priv_config file to configure the MoM:

# Configure /var/spool/pbs/mom_priv/configecho "mom config setup". /etc/pbs.confecho "\$clienthost $PBS_SERVER" >> /var/spool/pbs/mom_priv/configecho "\$clienthost ${PBS_SERVER//.*}" >> /var/spool/pbs/mom_priv/configecho "\$restrict_user_maxsysid 999" >> /var/spool/pbs/mom_priv/config

systemctl start pbs

Note: If PBS is installed in non-default directory, then change the path to the PBS homedirectory (i.e., /var/spool/pbs).

Optional ConfigurationUse the startup script to configure filesystems (/etc/fstab), configure NIS (/etc/yp.conf), mountnecessary filesystems, and any other configuration that your site requires.



8.8 Configuring Amazon Web Service CloudBurstingCreate the necessary components in AWS that are required for cloud bursting and then create a Controlcloud bursting scenario.

Before you can begin creating the AWS components necessary for cloud bursting, you must create andactivate an AWS account. For more information see How do I create and activate a new Amazon WebServices account.

8.8.1 Creating an Amazon Web Service User AccountCreate an Amazon Web Service (AWS) user account so that you can later create a cloud account inControl.

You must already have signed up for an AWS account.

AWS documentation can be found at Creating an IAM User in Your AWS Account.

Create an AWS user account and give this account administrative permissions. Once this task iscomplete, you should have downloaded a CSV file containing the following:

• Access key ID

• Secret access key

This information is necessary to create a cloud account in Control.

1. Login in the AWS console.

2. Using the search box located under AWS services, enter IAM.

3. Click the IAM search result.The Identity and Access Management dashboard is opened.

4. In the navigation pane on the left-hand side of the web page, click Users.

5. Click Add user.This user will have administrative rights to any machines hosted in the cloud.

6. Enter the following information to add a user:

a) For User name, enter a name for the user.

The name can be anything meaningful to your organization, e.g., pc_clouduser.

b) For Access type, enable Programmatic access.

The user requires this type of access because Control needs to make API calls or use theAWS CLI. An access key ID and a secret access key are generated for the user.

7. Click Next:Permissions.

8. Optional: Click Add user to group.

This button may already be selected.

9. Click Create group.


https://aws.amazon.com/premiumsupport/knowledge-center/create-and-activate-aws-account/

https://aws.amazon.com/premiumsupport/knowledge-center/create-and-activate-aws-account/

https://docs.aws.amazon.com/IAM/latest/UserGuide/id_users_create.html


10. Enter the following information to create a group, add the user to the group, and choose apermission policy for the group.

a) For Group name, enter a group name.

The name can be anything meaningful to your organization, e.g., pc_cloudgroup.

b) For Policy type, enable AdministratorAccess.

This policy provides full access to AWS services and resources.

11. Click Create group.You are returned to the Add user page and the new group is enabled indicating that the user isadded to the new group.

12. Click Next: Review.

13. Click Create user.14. Click Download.csv.

15. Download and save this file in a secure location.

This file contains an access key ID and a secret access key and is used later to add a cloudaccount to Control.

16. Click Close.The new user account is displayed.

8.8.2 Adding an Amazon Web Service Cloud AccountAdd an existing AWS user account to Control so that you can create a cloud bursting scenario or anappliance.

You must already created an AWS user account, given it administrative permissions and downloadeda CSV file containing an access key ID and a secret access key. The information in the CSV file isnecessary to create a cloud account in Control. If you do not have this information see Creating anAmazon Web Service User Account.


2. Click the Cloud tab.

3. Under Infrastructure, click Cloud.

4. Click .

5. Enter the following to add a cloud account:

a) For Account name, enter any name for the cloud account.

The name can be anything meaningful to your organization, e.g., aws_cloudaccount.

b) For Access Key ID, copy and paste the value of Access key ID from the downloaded CSVfile.

c) For Secret Access Key, copy and paste the Secret access key value from the downloadedCSV file.

6. Click Create account.



8.8.3 Creating a Virtual Private Cloud NetworkCreate a Virtual Private Cloud (VPC) in AWS so that you can later create a virtual machine for cloudbursting.

Login to your AWS Management Console and choose a region based on the geographical location ofyour users. All cloud resources that are created are placed in this region. For more information seeRegions and Availability Zones. The menu for selecting a region is located at the upper right-handcorner of the AWS Console menu bar.

AWS documentation can be found at Getting Started with IPv4 for Amazon VPC and Working with VPCsand Subnets.

Figure 10: AWS Region

1. Click located in the upper left-hand corner of the web page.

2. Using the search box located under AWS services, enter VPC.

3. Click the VPC search result.The VPC dashboard is opened.

4. In the menu located on the left-hand side of the web page, click Your VPCs.

Create a virtual private cloud.

5. Click Create VPC.

6. Enter the following to create a VPC:

a) For Name, enter any name for the VPC.

The name can be anything meaningful to your organization, e.g., bursting_vpc.

b) For IPv4 CIDR block, provide an address range in CIDR notation.

c) For IPv6 CIDR block, enable No IPv6 CIDR Block.

d) For Tenancy, choose Default.7. Click Yes, Create.

Create a subnet for the VPC:

8. In the menu located on the left-hand side of the web page, click Subnets.

9. Click Create Subnet.a) For Name tag, enter a name for the subnet.

The name can be anything meaningful to your organization, e.g., bursting_subnet.

b) For VPC, choose the VPC that was previously created (e.g. bursting_vpc).

c) For Availability Zone, choose one of the following options:

• Choose an availability zone from the menu.

• Choose No Preference to let Amazon choose an Availability Zone for you.

d) For IPv4 CIDR block, provide an address range in CIDR notation.

10. Click Create.


https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Concepts.RegionsAndAvailabilityZones.html

https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/getting-started-ipv4.html

https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/working-with-vpcs.html#Create-VPC

https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/working-with-vpcs.html#Create-VPC


11. Click Close.

8.8.4 Creating an Internet GatewayCreate an internet gateway and attach it to the bursting VPC.

AWS documentation can be found at Internet Gateways.

An internet gateway must be created so that you can later SSH into the virtual machine that is used forcloud bursting.


2. In the menu located on the left-hand side of the web page, click Internet Gateways.

3. Click Create internet gateway.

4. Enter the following to create an internet gateway:

a) For Name tag, enter a name for the internet gateway.

The name can be anything meaningful to your organization, e.g., bursting_gateway

5. Click Create.

6. Click Close.

7. Select the internet gateway that was just created by enabling the check box next to the name ofthe gateway.

Tip: You may need to deselect any other internet gateways that are displayed in thelist. Amazon creates default resources for your selected region so a default internetgateway may already exist.

8. Click Actions > Attach to VPC.

9. Select the VPC that was created previously (e.g. bursting_vpc).

10. Click Attach.

8.8.5 Updating the VPC Route TableAdd a rule to the VPC route table that allows all internet access and associate the route table to thebursting subnet.

AWS documentation can be found at Route Tables.


2. In the menu located on the left-hand side of the web page, click Route Tables.A list of route tables is displayed.

3. Select the route table attached to your VPC (e.g. bursting_vpc) by enabling the check box next tothe name of the route table.

The VPC column in the route table list specifies the VPC to which the route table is attached.

4. Click the Routes tab at the bottom of the web page.


https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPC_Internet_Gateway.html

https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPC_Route_Tables.html


5. Click Edit.

Figure 11: Add a Route

6. Click Add another route.

Add a rule that allows all traffic access to the internet gateway:

7. Enter the following to add a rule:

a) For Destination enter 0.0.0.0/0.

b) For Target, select the internet gateway that was created previously (e.g. bursting_gateway).

8. Click Save.

Associated the route table to the bursting subnet:

9. Click the Subnet Associations tab.

10. Click Edit.11. Enable the Associate check box.

12. Click Save.

8.8.6 Creating a Virtual MachineCreate a CentOS 7 virtual machine in AWS Elastic Compute Cloud (EC2).

AWS documentation can be found at Launch a Linux Virtual Machine and Launching a Virtual Machinewith Amazon EC2.



3. Using the search box located under AWS services, enter EC2.

4. Click the EC2 search result.The EC2 dashboard is opened.

5. In the menu located on the left-hand side of the web page, click Instances.

6. Click Launch Instance.

7. In the menu located on the left-hand side of the web page, click AWS Marketplace.

8. Using the Search AWS Marketplace Products search box, enter CentOS 7.

9. Locate the CentOS 7 Amazon Machine Image (AMI) and click Select.10. Click Continue.

11. Select an Instance Type appropriate for your site's workload based on:


https://aws.amazon.com/getting-started/tutorials/launch-a-virtual-machine/

https://www.youtube.com/watch?v=SKM0BB0F02Q

https://www.youtube.com/watch?v=SKM0BB0F02Q


• the number of cores

• the amount of memory

• storage

• network performance

Consider the nature of the applications that you plan to deploy on the instance, the number ofusers that you expect to use the applications, and also how you expect the load to scale in thefuture. Remember to also factor in the CPU and memory resources that are necessary for theoperating system.

12. Click Next: Configure Instance Details.

13. Enter the following to configure instance details:

a) For Number of instances, specify 1.

b) For Network, choose the VPC that you previously created (e.g. bursting_vpc).The bursting subnet is automatically populated.

c) For Auto-assign Public IP, select Enable.

14. Click Next: Add Storage.

15. Specify the storage options as per your site's requirements.

Tip: It is recommended to enable Delete on Termination to delete EBS volumeswhen the virtual machine is terminated.

16. Click Next: Add Tags.

17. Optional: You can add tags in key-value pairs for the VM per your site's requirements.

18. Click Next: Configure Security Group.

19. Enter the following to create a security group for the VM:

a) For Assign a security group, enable Create a new security group.

b) For Security group name, enter a name for the security group.

The name can be anything meaningful to your organization, e.g., bursting_security_group.

a) Add new security rules based on your site's requirements.

Figure 12: Security Rules

20. Click Review and Launch.

21. Review the VM information of the VM and click Launch.

22. Create a new public/private key pair for the VM by entering the following:

a) Select Create a new key pair.b) Provide a name for the key pair.

c) Click Download Key Pair.d) Download and save this file in a secure location.



The information in this .pem file is used later to SSH into the VM.

23. Click Launch Instances.

24. At the bottom of the web page, click View Instances.All virtual machines that have been created are displayed.

Your virtual machine is ready when the Instance State is equal to "running" and Status Checks arecomplete. The virtual machine can be accessed using its IPv4 Public IP.

Figure 13: Bursting Virtual Machine

8.8.7 Installing and Configuring the PBS MoM on the VMInstall the PBS Professional MoM on the virtual machine and configure the VM per your site'senvironment.

The username for logging into the virtual machine is dependent on the Amazon Machine Image (AMI)that you used to create the virtual machine. For example, the user name for a CentOS AMI is "centos".

Note: You will need to set the default_user parameter of the PBS Professional burstinghook configuration file to this username.

For information about user names and instructions for connecting and copying files to the virtualmachine see Connecting to Your Linux Instance Using SSH.

You will need the .pem file downloaded while creating the VM.

A cloud-init script is provided during the creation of the bursting scenario. This script is ran after acloud node is burst and is used to set up the system environment so that nodes burst in the cloud cancommunicate with the PBS Professional Server, therefore installation of cloud-init on the VM is required.For more information about installing cloud-init, refer to the Cloudinit Documentation.

1. Login to your site's PBS Professional Server as the user pbsadmin.

2. SSH into the virtual machine as the user "centos "using the IPv4 Public IP assigned to the VM byentering a command similar to:ssh -i /path/my-key-pair.pem centos@IPV4PublicIP

where /path/my-key-pair.pem is the location of the .pem file downloaded while creating thevirtual machine and IPV4PublicIP is the public IP address of the virtual machine.



https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html

http://cloudinit.readthedocs.io/en/latest/index.html


3. Enter the command:sudo -i

4. Using the PBS Professional Installation and Upgrade Guide, install and configure the PBSProfessional MOM.

5. Once the PBS Professional MOM is installed, edit the /etc/pbs.conf file.

6. Change the value of PBS_SERVER to the hostname of the PBS Professional Server.

7. Save the file.

8. Edit the file PBS_HOME/mom_priv/config.

where PBS_HOME is the home directory of PBS Professional.

9. Change the value of $clienthost to the hostname of the PBS Professional Server.

10. Save the file.

11. Configure the VM for your site's environment such as pointing it the DNS server, mounting filesystems, connecting it to the authentication service, installing any applications, etc.

12. Install cloud-init.

13. Connect the VM to your PBS Professional server and submit a job to ensure it is working asexpected.

8.8.8 Creating an OS ImageCreate an image of the virtual machine configured per your site's requirements.

You must have already create a virtual machine, installed the PBS Professional MoM on the VM andconfigured the VM per your site's environment.

AWS documentation can be found at Create an AMI from an Amazon EC2 Instance.





5. In the menu located on the left-hand side of the web page, click Instances.

6. Select the virtual machine created for cloud bursting by enabling the check box next to its name.

7. At the top, click Actions > Instance State > Stop.

8. Click Yes, Stop.

It may take some time for the virtual machine to be stopped.

Important: Do not proceed until the Instance State is "stopped".

9. Click Actions > Image > Create Image.

10. For Image name, enter a name for the image.

The name can be anything meaningful to your organization, e.g., bursting_image.


https://docs.aws.amazon.com/toolkit-for-visual-studio/latest/user-guide/tkv-create-ami-from-instance.html


11. Click Create Image.

12. Click the View pending image ami-xxxxxxxxxx link.

The image is complete when its Status is "available".

You can now delete the virtual machine so that you are no longer charged for it.

8.8.9 Creating a Cloud Bursting ScenarioCreate an AWS bursting scenario so that PBS Professional execution nodes can be burst in the cloud.

Open a browser window and login to your AWS Management Console. Open a second browser windowand login to Control.

A cloud-init script is provided during the creation of the bursting scenario. This script is ran after acloud node is burst and is used to set up the system environment so that nodes burst in the cloud cancommunicate with the PBS Professional Server. For more information about writing cloud-init scripts,refer to the Cloudinit Documentation and the Cloud Bursting chapter of the ControlAdministrator'sGuide.

1. Navigate to the Control browser window.

2. Click Cloud.

3. Under Infrastructure, click Bursting.

4. Click Add Bursting Scenario.

5. Enter information about the bursting scenario such as name, cloud provider, cloud account, andgeographic location.

a) For Name, enter a name for the cloud bursting scenario.

b) For Description, enter a description.

c) For Cloud Account, choose a registered AWS account from the list.

d) Select a region from the list that matches the region selected in AWS.

Figure 15: AWS Region

6. Click Next.7. For Domain name, provide the domain name of your organization for the cloud bursting network.

altair.com, burst.altair.com

8. For Hostname prefix, enter a prefix for the names of the burst nodes.If "node" is entered , the cloud bursting nodes will be available to PBS Professional as node1,node2, node3, etc.

9. Navigate to the AWS browser window.







13. In the menu located on the left-hand side of the web page, click AMIs.

14. Select the Amazon Machine Image (AMI) created for cloud bursting by enabling the check boxnext to its name.

15. In the Details tab located at the bottom of the web page, hover over the AMI ID.

A copy to clipboard icon is displayed.

16. Click .


18. For the AMI ID, paste the AMI ID copied from AWS.


20. In the menu located on the left-hand side of the web page, click Security Groups underNETWORK & SECURITY.

21. Select the Security Group created for the bursting virtual machine by enabling the check box nextto its Group ID.

22. In the Description tab located at the bottom of the web page, hover over the Group ID.


23. Click .


25. For the Security group ID, paste the Group ID copied from AWS.



28. Using the search box located under AWS services, enter VPC.

29. Click the VPC search result.The VPC dashboard is opened.

30. In the menu located on the left-hand side of the web page, click Subnets.

31. Select the subnet for the bursting VPC by enabling the check box next to its name.

32. In the Description tab located at the bottom of the web page, hover over the Subnet ID.


33. Click .


35. For the Subnet ID, paste the Subnet ID copied from AWS.

36. Enable the Add public IP to VMs if you need to connect to the cloud bursting nodes from theinternet.

It is recommended to leave this disabled.

37. For Cloud-init script, click Browse, browse for the script file and select it.

This script will run after a cloud node is burst.

38. Add public SSH keys for any users that you want to have access to the burst nodes:



a) Login to the PBS Professional Server.

b) Copy the public SSH key for the user.

Public key files are usually stored in /.ssh in the user's home directory.

c) Navigate to the Control browser window.

d) Click Add to add the public key.

e) For Public SSH keys, paste the public SSH key.

f) Repeat steps a through e to add additional public SSH keys.

The public keys are copied to each burst node for secure connectivity.

39. Click Instantiate scenario.



8.9 Configuring Microsoft Azure Cloud BurstingCreate the necessary components in Microsoft Azure that are required for cloud bursting and thencreate a Control cloud bursting scenario.

Before you can begin creating the Azure components necessary for cloud bursting, you must purchasean Azure subscription and have created a tenant. For more information about subscriptions see What isan Azure subscription. For more information about tenants see What is an Azure Active Directory tenantand How to get an Azure Active Directory tenant. As a final step, add users to the Active Directory.

8.9.1 Registering the Control Application with AzureRegister Control with Azure so that you can later create a cloud account in Control.

You must have already purchased an Azure subscription, obtained a tenant and have an Azure useraccount.

Azure documentation can be found at Integrating applications with Azure Active Directory

Register the Control web application with the Azure Active Directory and create an API access key.Manage access to the newly create application by assigning a role. Also, obtain your Azure tenant IDand subscription ID. Once this task is complete, you should have the following information stored in afile:

• Application ID generated when registering Control with the Azure Active Directory.

• API access key generated for the newly registered Control application.

• Subscription ID

• Tenant ID


1. Login in to your Microsoft Azure account.

Register Control to the Azure Active Directory.

2. In the menu located on the left-hand side of the web page, click Azure Active Directory.

3. Under MANAGE, click App registrations.


https://docs.microsoft.com/en-us/azure/architecture/cloud-adoption-guide/adoption-intro/subscription-explainer

https://docs.microsoft.com/en-us/azure/architecture/cloud-adoption-guide/adoption-intro/subscription-explainer

https://docs.microsoft.com/en-us/azure/architecture/cloud-adoption-guide/adoption-intro/tenant-explainer

https://docs.microsoft.com/en-us/azure/active-directory/develop/active-directory-howto-tenant?toc=%2Fen-us%2Fazure%2Farchitecture%2Ftoc.json&bc=%2Fen-us%2Fazure%2Farchitecture%2Fbread%2Ftoc.json

https://docs.microsoft.com/en-us/azure/active-directory/fundamentals/add-users-azure-active-directory?toc=%2Fen-us%2Fazure%2Farchitecture%2Ftoc.json&bc=%2Fen-us%2Fazure%2Farchitecture%2Fbread%2Ftoc.json

https://docs.microsoft.com/en-us/azure/active-directory/develop/active-directory-integrating-applications


Figure 16: Application Registration

4. Click New application registration.

5. Enter the following to register Control to the Azure Active Directory:

a) For Name, enter the name of the application.

The name can be anything meaningful to your organization, e.g., pc_cloudapp.

a) For Application type, select Web app / API.

b) For Sign-on URL, enter the URL https://hostname:9500/pc.

where hostname is the hostname of the machine where the Control Web Interface isinstalled. This is the URL that is used to login to Control.

6. Click Create.Once the application registration is complete, its details are displayed, including an Application ID.

Obtain the application id of the newly registered application.

7. Hover over the Application ID and click when it appears.

8. Store the Application ID to a file.

The Application ID is used later to create a cloud account in Control.

Create an api-key for your newly registered application.

9. Click .

Figure 17: Registered Application Settings



10. Under API Access, click Keys.

11. Enter the following to create an api-key:

a) For DESCRIPTION, enter pc_access_key.

b) For EXPIRES, select Never Expires.

12. Click Save.A key value is generated and is displayed under the heading VALUE.

13. Copy the key value.

14. Store the key value to a file.

The key value is used later to create a cloud account in Control.

Obtain your Azure subscription ID.

15. In the menu located on the left-hand side of the web page, click Cost Management + Billing.

16. Under BILLING, click Subscriptions.

Figure 18: Subscriptions

17. Locate and click your subscription.Details about the subscription are displayed, including a Subscription ID.

18. Hover over the Subscription ID and click when it appears.

19. Store the Subscription ID value to a file.

The Subscription ID is used later to create a cloud account in Control.

Assign an access control role to the newly registered application.

20. Click Access control (IAM).

Figure 19: Add Access Controls

21. Click Add.



22. In the Add permissions panel, enter the following to assign a role to the newly registeredapplication:

a) For Role, select Contributor.b) For Assign access to, select Azure AD user, group, or application.

c) For Select, search for the newly registered application, e.g., pc_cloudapp.

23. Click Save.

Obtain your Azure tenant ID:

24. At the top of the web page, click .

25. Click Show diagnostics.

A dialog box is displayed allowing a file called PortalDiagnostics.json to be saved.

26. Open the file using any text editor.

27. Search for tenantId.

28. Store the value of tenantId to a file.

The Tenant ID is used later to create a cloud account in Control

8.9.2 Adding a Microsoft Azure Cloud AccountAdd an Azure cloud account to Control so that you can create a cloud bursting scenario or an appliance.

You should have the following information stored in a file:

• Application ID generated when registering Control with the Azure Active Directory.

• API access key generated for the newly registered Control application.

• Subscription ID

• Tenant ID

This information is necessary to create a cloud account in Control. If you do not have this informationsee Registering the Control Application with Azure.




4. Click .



The name can be anything meaningful to your organization, e.g., azure_cloudaccount.

b) For Client ID, enter the Application ID generated when Control was registered with the AzureActive Directory.

c) For Secret Key, enter the API access key generated for the newly registered Controlapplication.

d) For AD tenant ID, enter your Azure tenant ID.

e) For Subscription ID, enter your Azure subscription ID.




8.9.3 Creating a Resource GroupCreate a resource group that will later be used to create a cloud bursting scenario.

You must have already purchased an Azure subscription and have an Azure user account.

Azure documentation can be found at What is an Azure resource group and Manage Azure resourcesthrough portal.

A resource group is container that holds related resources for an Azure solution. The resource group caninclude all the resources for the solution, or only those resources that you want to manage as a group.You decide how you want to allocate resources to resource groups based on what makes the most sensefor your organization. Once the resource group is created, resources that are placed into the resourcegroup are a virtual network, a virtual machine, and an image of the virtual machine.

1. Login to your Microsoft Azure account.

2. In the menu located on the left-hand side of the web page, click Resource groups.

3. Click + Add.

4. Create a resource group by entering the following:

a) For Resource group name, enter a name for the resource group.

The name can be anything meaningful to your organization, e.g., bursting_resource_group.

b) For Subscription, choose a subscription .

c) For Resource group location, select a location based on the geographical location of users.

5. Click Create.

It may take a moment to create the resource group. All resources (networks, virtual machines,etc.) that are created are placed within this resource group. The name of the resource group isrequired for creating a bursting scenario in Control.

8.9.4 Creating a Virtual NetworkCreate a virtual network so that you can later create a virtual machine for cloud bursting.

You must have already purchased an Azure subscription, have an Azure user account and have createda resource group.

Azure documentation can be found at Virtual Network Documentation.


2. In the menu located on the left-hand side of the web page, click Virtual networks.

3. Click +Add.

4. Enter the following to create a virtual network:

a) For Name, enter a name for the virtual network.

The name can be anything meaningful to your organization, e.g., bursting_virtual_network

b) For Address space, enter an address range for the network using CIDR notation.


https://docs.microsoft.com/en-us/azure/architecture/cloud-adoption-guide/adoption-intro/resource-group-explainer

https://docs.microsoft.com/en-us/azure/azure-resource-manager/resource-group-portal?toc=%2Fen-us%2Fazure%2Farchitecture%2Ftoc.json&bc=%2Fen-us%2Fazure%2Farchitecture%2Fbread%2Ftoc.json

https://docs.microsoft.com/en-us/azure/azure-resource-manager/resource-group-portal?toc=%2Fen-us%2Fazure%2Farchitecture%2Ftoc.json&bc=%2Fen-us%2Fazure%2Farchitecture%2Fbread%2Ftoc.json

https://docs.microsoft.com/en-us/azure/virtual-network/


c) For Subscription, choose a subscription.

d) For Resource group, enable Use existing and choose the previously created burstingresource group.

e) For Location, select the same geographical location as was selected for the previously createdbursting resource group.

f) For Subnet, enter a name for the virtual machine's subnet.

The name can be anything meaningful to your organization, e.g., bursting_subnet

g) For Address range, enter an address range for the subnet in CIDR notation.

5. Click Create.

It may take a moment to create the virtual network. The name of the virtual network is requiredfor creating a bursting scenario in Control.

8.9.5 Creating a Network Security Group and Add SecurityRulesCreate a network security group so that you can perform cloud bursting.


Azure documentation can be found at Network security.


2. In the menu located on the left-hand side of the web page, click Resource groups.

3. Select the resource group that was created for cloud bursting by clicking on its name.

4. Click +Add.

5. In the search box, enter network security group. and press ENTER.

6. Select the Network security group search result.

7. Click Create.

8. Enter the following to create a network security group:

a) For Name, enter a name for the network security group.

The name can be anything meaningful to your organization, e.g.,bursting_network_security_group.

a) For Subscription, choose a subscription.

b) For Resource group, enable Use existing and choose the previously created burstingresource group.

c) For Location, select the same geographical location as was selected for the previously createdbursting resource group.

9. Click Create.

It may take a moment for the network security group to deploy.

10. Once the network security group is deployed a message is displayed indicating success, click onGo to resource.


https://docs.microsoft.com/en-us/azure/virtual-network/security-overview


11. Click Inbound security rules.

Figure 20: Inbound Security Rule

12. Click Add.

13. Enter the following to add an inbound security rule:

a) For Source, choose Any.

Specifies that incoming traffic from any IP address range is allowed.

b) For Source port ranges, enter *

Specifies that incoming traffic from any ports is allowed.

c) For Destination, choose Any.

Specifies that outgoing traffic from any IP address range is allowed.

d) For Destination port ranges, enter 22.

Specifies that outgoing traffic for port 22 is allowed.

e) For Protocol, click TCP.

f) For Action, enable Allow.

14. Click Add.

It may take a moment for the security rule to be created.

8.9.6 Creating a Virtual MachineCreate a CentOS 7.3 virtual machine in Azure.


You may want to view the following web page to learn about Azure Managed Disks before creating aVM. Additionally, a video is available from Microsoft that shows how to create a virtual machine: Createa Linux Virtual Machine.

To provide a secure connection between your site's PBS Server and the virtual machine you must supplythe SSH public key for the user pbsadmin. You can obtain the SSH public key by logging into the PBS


https://docs.microsoft.com/en-us/azure/virtual-machines/windows/managed-disks-overview

https://azure.microsoft.com/en-us/resources/videos/create-linux-virtual-machine/

https://azure.microsoft.com/en-us/resources/videos/create-linux-virtual-machine/


Server as pbsadmin and viewing the contents of the public key files that are usually stored in the .sshdirectory.


2. In the menu located on the left-hand side of the web page, click Virtual Machines.

3. Click Add.

4. In the search bar, enter CentOS-based 7.3 and press ENTER.

5. Select the CentOS-based 7.3 option from the search results list.

6. Click Create.

7. Enter the following to configure the basic settings for the virtual machine:

a) For Name, enter a name for the virtual machine.

The name can be anything meaningful to your organization, e.g., bursting-vm.

b) For VM disk type, choose one of the following options:

• SSD (Solid State Drive)

• HDD (Hard Disk Drive)

Choose SSD for I/O-intensive applications, where low latency and high throughput arecritical. For testing, consider HDD to keep costs down, as you scale up and down quickly.

c) For Username, enter pbsadmin.

Note: You will need to set the default_user parameter of the PBS Professionalbursting hook configuration file to this username.

d) For Authentication type, enable SSH public key.

e) For SSH public key, copy the SSH public key (i.e., id_rsa.pub) for the user pbsadmin from thePBS Server and paste it.

This provides a secure connection between your PBS Server and the virtual machine.

f) For Subscription, choose a subscription.

g) For Resource Group, enable Use Existing and choose the previously created burstingresource group.

h) For Location, select the same geographical location as was selected for the previously createdbursting resource group.

8. Click OK.

9. Choose a VM size appropriate for your site's workload based on:

• the number of cores

• the amount of RAM

• disk drive size

• price

Consider the nature of the applications that you plan to deploy on the instance, the number ofusers that you expect to use the applications, and also how you expect the load to scale in thefuture. Remember to also factor in the CPU and memory resources that are necessary for theoperating system.



10. Click Select.11. Enter the following to configure the optional features of the VM:

a) For Storage, choose whether you want to use managed disks by selecting:

• Yes to use managed disks.

• No to not use managed disks.

Enable this feature to have Azure automatically manage the availability of disks to providedata redundancy and fault tolerance, without creating and managing storage accounts onyour own. This option is recommended by Azure as it is a lot more scalable.

b) For Network Security Group, enable Advanced and choose the network security grouppreviously created for cloud bursting.

c) Accept defaults for the rest of the optional features.

12. Select OK.A summary of your VM is displayed.

13. Click Create.

It may take a few minutes for the VM to be deployed. You will use this virtual machine to createan OS image.

Once the virtual machine is deployed a message is displayed indicating success, click on Go toresource.

8.9.7 Installing and Configuring the PBS MoM on the VMInstall the PBS MoM on the virtual machine and configure the VM per your site's environment.

The user name for logging into the virtual machine is determined by the SSH key that was providedduring the creation of the virtual machine.

A cloud-init script is provided during the creation of the bursting scenario. This script is ran after acloud node is burst and is used to set up the system environment so that nodes burst in the cloud cancommunicate with the PBS Professional Server,therefore installation of cloud-init on the VM is required.For more information about installing cloud-init, refer to the Cloudinit Documentation.

1. Login to your site's PBS Server as the user pbsadmin.

2. SSH into the virtual machine as pbsadmin using the External IP assigned to the VM.






4. Using the PBS Professional Installation and Upgrade Guide, install and configure the PBS MOM.

5. Once the PBS MOM is installed, edit the /etc/pbs.conf file.

6. Change the value of PBS_SERVER to the hostname of the PBS Server.

7. Save the file.

8. Edit the file PBS_HOME/mom_priv/config where PBS_HOME is the home directory of PBS.

9. Change the value of $clienthost to the hostname of the PBS Server.

10. Save the file.



13. Connect the VM to your PBS server and submit a job to ensure it is working as expected.


You must have already create a virtual machine, installed the PBS MoM on the VM and configured theVM per your site's environment.

To create an OS image requires the Azure CLI. Refer to these instructions for installing the CLI How toinstall the Azure CLI.

Before you can create an OS image of the previously created VM, you must first SSH into the VM anddeprovision it. Next you will use the Azure CLI to deallocate and generalize the VM and then create theimage. Generalizing the virtual machine removes any SSH keys and DNS settings from the VM.

Follow Step 1 and Step 2 as documented at the URL How to create an image of a virtual machine orVHD to create an image of the VM.


https://docs.microsoft.com/en-us/cli/azure/?view=azure-cli-latest

https://docs.microsoft.com/en-us/cli/azure/install-azure-cli?view=azure-cli-latest

https://docs.microsoft.com/en-us/cli/azure/install-azure-cli?view=azure-cli-latest

https://docs.microsoft.com/en-us/azure/virtual-machines/linux/capture-image

https://docs.microsoft.com/en-us/azure/virtual-machines/linux/capture-image


Note: Before you can deallocate the virtual machine you may have to execute the followingcommands to set your subscription to be the active subscription.az account listaz account set --subscription yourSubscriptionID

An OS image is created of the virtual machine.


8.9.9 Creating a Bursting ScenarioCreate an Azure bursting scenario so that PBS execution nodes can be burst in the cloud.

Open a browser window and login to your Microsoft Azure account. Open a second browser window andlogin to Control.

Before creating a cloud bursting scenario, read the following Azure article about scale sets What arevirtual machine scale sets.


Note: To create a bursting scenario you must enter a resource group, however theresources necessary for creating the bursting scenario (network, subnet, network securitygroup and the OS image) can all reside in a different resource group. However, the resourcegroups must be in the same geographic location for this to work.


2. Click Cloud.






c) For Cloud Account, choose a previously created Azure cloud account from the list.

d) For Select a region, select the same geographic location as was selected for the previouslycreated bursting resource group.

6. Click Next.Enter information about the Azure resources necessary to burst in the cloud.

7. For Domain name, provide the domain name of your organization for the cloud bursting network.altair.com, burst.altair.com


https://docs.microsoft.com/en-us/azure/virtual-machine-scale-sets/overview#differences-between-virtual-machines-and-scale-sets

https://docs.microsoft.com/en-us/azure/virtual-machine-scale-sets/overview#differences-between-virtual-machines-and-scale-sets




9. Navigate to the Azure browser window.

10. In the menu located on the left-hand side of the web page, click Resource Groups.

11. Copy the name of the resource group created for cloud bursting.


13. For Resource group name, paste the resource group copied from Azure.


15. In the menu located on the left-hand side of the web page, click Virtual Networks.

16. Copy the name of the virtual network created for cloud bursting.


18. For Network name, paste the name of the virtual network copied from Azure.

If the network is contained in a resource group that is different from the one entered for thebursting scenario, it must be entered as Resource Group Name/Virtual Network Name.


20. Select the virtual network created for cloud bursting.

21. Click Subnets.

Figure 22: Subnet

22. For Subnet name, copy the name of the subnet created for the cloud bursting virtual network.


24. For Subnet name, paste the Subnet copied from Azure.

If the subnet is contained in a resource group that is different from the one entered for thebursting scenario, it must be entered as Resource Group Name/Subnet Name.



27. Select the Resource Group created for cloud bursting.



28. In the list, locate the Network Security group created for cloud bursting.

The Type is Network Security group.

29. Copy the name of the network security group.


31. For Network security group name, paste the network security group copied from Azure.

If the network security group is contained in a resource group that is different from the oneentered for the bursting scenario, it must be entered as Resource Group Name/NetworkSecurity Group Name.

32. Choose one of the following options based on whether you chose to use the Azure managed diskfeature when creating the virtual machine:

• Enable Managed Storage if you enabled the managed disk feature.

• Disable Managed Storage if you disabled the managed disk feature.



35. Select the Resource Group created for cloud bursting.

36. In the list, locate the image that was created for cloud bursting.

The Type is Image.

37. Select the image.


• If you chose to use managed disks when you created the VM, copy the name of the image.

• If you did not choose to use managed disks when you created the VM, copy the Linux SourceBLOB URI.


40. For OS image, paste the information copied from Azure.

If the OS image is contained in a Resource group that is different from the one entered for thebursting scenario, it must be entered as Resource Group Name/OS Image Name or ResourceGroup Name/URI.



42. For Maximum number of VMs inside a ScaleSet, specify the limit for your cloud bursting scenario.

If you use InfiniBand, Azure limits the number of VMs on a scaleset to 100. Without InfiniBand, fora scaleset with managed disk and custom image, you can specify a higher limit.

43. For Cloud-init script, click Browse, browse for the script file and add it.



a) Login to the PBS Server.













8.10 Configuring Google Cloud Platform CloudBurstingCreate the necessary components in GCP that are required for cloud bursting and then create a Controlcloud bursting scenario.

8.10.1 Creating a ProjectCreate a GCP Project so that you can create the necessary cloud components for cloud bursting.

GCP documentation can be found at Creating and Managing Projects.

Google Cloud Platform projects form the basis for creating, enabling, and using all Cloud Platformservices including managing APIs, enabling billing, adding and removing collaborators, and managingpermissions for Cloud Platform resources.

1. Login to the GCP console.


3. Click Home.

4. Click Create.

5. For Project Name, enter a name for the project.

The name can be anything meaningful to your organization, e.g., pc_cloudproject.

6. Click Create.

It may take a few moments to create the project.

7. Using a browser, navigate to the following URL: https://console.developers.google.com/apis/library/compute.googleapis.com?project=PROJECTNAME

where PROJECTNAME is the name of the project.

8. Click Enable.

8.10.2 Creating a Service AccountCreate a Google Cloud Platform service account so that you can later create a cloud account in Control.

You must already have signed up for an GCP account.

GCP documentation can be found at Understanding Service Accounts and Compute Engine IAM Roles .

A service account is a special Google account that belongs to an application or a virtual machine,instead of to an individual end user. Create a service account and give this account administrativepermissions for all Compute Engine resources. Once this task is complete, you should have downloadeda JSON file containing the following:

• Project ID

• Client ID

• Client email


https://cloud.google.com/resource-manager/docs/creating-managing-projects

https://cloud.google.com/iam/docs/understanding-service-accounts

https://cloud.google.com/compute/docs/access/iam#compute_admin_role


• Private key ID

• Private key




3. Click IAM & admin > Service accounts.

4. Click CREATE SERVICE ACCOUNT.

5. Enter the following to create a service account:

a) For the Service account name, enter a name for the service account.

The name can be anything meaningful to your organization, e.g., pc-service-account.

b) Click CREATE.

c) For the Project role, click Select a role > Compute Engine > Compute Admin.

This role gives full control of all Compute Engine resources.

d) Click CONTINUE.

e) Under Create key (optional), click CREATE KEY.

f) For Key type, enable JSON.

g) Click CREATE.

h) Save the JSON file in a secure location.

The information in this file is used later to add a cloud account to Control.

i) Click CLOSE.

j) Click DONE.The service account is created.

8.10.3 Adding a GCP Cloud AccountAdd an existing GCP service account to Control so that you can create a cloud bursting scenario or anappliance.

You must have created a GCP service account, given it Compute Admin permissions and downloadeda JSON file containing information about the service account. The information in the JSON file isnecessary to create a cloud account in Control. If you do not have this information see Creating aService Account.




4. Click .



The name can be anything meaningful to your organization, e.g., gcp_cloudaccount.



b) For Project ID, copy and paste the project_id value from the JSON file.

c) For Client ID, copy and paste the client_id value from the JSON file.

d) For Client Mail, copy and paste the client_email value from the JSON file.

e) For Private Key ID, copy and paste the private_key_id value from the JSON file.

f) For Private Key, copy and paste the private_key value from the JSON file.

Important: Do not copy the \n characters at the end of the private_key value.


8.10.4 Creating a Virtual Private Cloud NetworkCreate a Virtual Private Network (VPC) in GCP so that you can later create a virtual machine for cloudbursting.

GCP documentation can be found at Virtual Private Cloud Documentation and Using VPC Networks .



3. Click VPC network > VPC networks.

4. Click CREATE VPC NETWORK.

5. Enter the following to create a VPC:

a) For the Name, enter a name for the VPC.

The name can be anything meaningful to your organization, e.g., bursting-vpc.

b) In the Subnets section, click the Custom tab under Subnet creation mode.

c) For Name, enter a name for the subnet.

The name can be anything meaningful to your organization, e.g., bursting-subnet.

d) For Region, select a Region based on the geographical location of users.

e) For IP address range, enter an IP address range using CIDR notation

f) For Private Google access, enable Off.g) Click Done.

h) For Dynamic routing mode, enable Regional.6. Click Create.

Creating the VPC network may take some time. Do not proceed until the VPC is created.

7. Select the VPC by clicking on its name.

8. Click the Firewall rules tab.

9. Click CREATE FIREWALL RULE.

10. Enter the following to create a firewall rule:

a) For Name, enter a name for the firewall rule.

The name can be anything meaningful to your organization, e.g., ssh-all.

b) For Direction of Traffic, enable Ingress.


https://cloud.google.com/vpc/docs/

https://cloud.google.com/vpc/docs/using-vpc


c) For Action on match, enable Allow.

d) For Targets, select All instances in the network.

e) For Source filter, select IP ranges.

f) For Source IP ranges, enter 0.0.0.0/0.

g) For Protocols and ports, enable Specified protocols and ports.

h) Enable tcp.

i) Enter 22.

11. Click Create.

8.10.5 Creating a Virtual MachineCreate a CentOS 7.3 virtual Machine in GCP.

You must have already created a virtual network with a subnet and added a firewall rule.

GCP documentation can be found at Virtual Machine Instances and Creating and Starting a VM Instance.



3. Click Compute Engine > VM instances.

4. Click CREATE INSTANCE.

5. Enter the following to create a virtual machine:

a) For the Name, enter a name for the virtual machine.

The name can be anything meaningful to your organization, e.g., bursting-vm.

b) For Zone, select a zone that is in the same Region as the subnet of the previously createdVPC.

c) In the Machine type, click the Customize link.

Figure 23: Customize the Machine Type

d) Specify the CPUs, GPUs and RAM.

Consider the nature of the applications that you plan to deploy on the instance, the numberof users that you expect to use the applications, and also how you expect the load to scale inthe future. Remember to also factor in the CPU and memory resources that are necessary forthe operating system.

e) For Boot disk, click Change.

f) Choose CentOS 7.

g) For Boot disk type, choose one of the following options:


https://cloud.google.com/compute/docs/instances/

https://cloud.google.com/compute/docs/instances/create-start-instance


• Standard persistent disk

• SSD persistent disk

Choose SSD for I/O-intensive applications, where low latency and high throughput arecritical. For testing, consider Standard persistent disk to keep costs down.

h) For Size, specify the size of the boot disk.

i) Click Select.j) Under Identity and API access, for Service Account, select No service account.k) For Firewall, choose Allow HTTP traffic.

l) Click Management, disks, networking, SSH Keys.

m) Click the Networking tab.

n) Click Add network interface.

o) For Network, choose the VPC you previously created for bursting.

p) For Network Service Tier, click Standard.

q) Click Done.

r) Delete any default network interfaces that might have been automatically generated.

s) Click the Security tab.

t) For SSK Keys, copy the SSH public key (i.e., id_rsa.pub) for the user pbsadmin from the PBSServer and paste it.

This provides a secure connection between your PBS Server and the virtual machine.

Important: You will need to set the default_user parameter of the PBSProfessional bursting hook configuration file to the username of whatever SSH keyyou copy. GCP does not allow you to SSH into the VM as root so do not copy theroot SSH key.

6. Click Create.Creating the virtual machine may take some time.


The user name for logging into the virtual machine is determined by the SSH key that was providedduring the creation of the virtual machine.



2. SSH into the virtual machine as pbsadmin using the External IP assigned to the VM.









7. Save the file.


where PBS_HOME is the home directory of PBS.


10. Save the file.






GCP documentation can be found at Creating, Deleting, and Deprecating Custom Images.



3. Click Compute Engine > VM instances.A list of virtual machines is displayed.

4. Click next to the virtual machine that was created for cloud bursting.


5. Click Stop from the menu.

6. Click Stop.


https://cloud.google.com/compute/docs/images/create-delete-deprecate-private-images


It may take some time for the VM to be stopped. Do not proceed until the VM is stopped.

7. In the menu located on the left-hand side of the web page, click Images .

8. Click CREATE IMAGE.

9. Enter the following to create an image:

a) For Name, enter a name for the image.

The name can be anything meaningful to your organization, e.g., bursting-image.

b) For Source select Disk.

c) For Source disk, select the previously created virtual machine.

10. Click Create.

It may take some time to create the image. Do not proceed until the image is created.


8.10.8 Creating a Cloud Bursting ScenarioCreate an GCP bursting scenario so that PBS execution nodes can be burst in the cloud.

Open a browser window and login to your GCP console. Open a second browser window and login toControl.



2. Click Cloud.






c) For Cloud Account, choose a previously created GCP cloud account from the list.

d) For Select a region, select the same geographic location as was selected for the previouslycreated bursting VPC.

6. Click Next.Enter information about the GCP resources necessary to burst in the cloud.


8. For Hostname prefix, enter a prefix for the names of the burst nodes.




If "node" is entered , the cloud bursting nodes will be available to PBS Professional as node1,node2, node3, etc.

9. Navigate to the GCP browser window.


11. Click VPC network > VPC networks.

12. Click on the name of the VPC that was created for cloud bursting.VPC network details are displayed.

13. Copy the name of the VPC network.


15. For Network name, paste the name of the VPC network copied from GCP.


17. Copy the name of the VPC network subnet.


19. For Subnetwork name, paste the name of the VPC network subnet copied from GCP.





23. Click Compute Engine > Images.

24. Select the image created for cloud bursting.

25. Click Equivalent REST26. Copy the value for the entry called selfLink.


28. For OS Image URI, paste the link copied from GCP.















8.11 Configuring Oracle Cloud Platform CloudBurstingCreate the necessary components in Oracle Cloud that are required for cloud bursting and then create aControl cloud bursting scenario.

8.11.1 Creating an Oracle Cloud User AccountCreate an Oracle Cloud user account so that you can later create a cloud account in Control.

You must already have signed up for an Oracle Cloud account and have an associated tenancy.

Oracle documentation can be found at Adding Users and Resource Identifiers.

An Oracle user account must be created to add an Oracle cloud account to Control. Once this task iscomplete, you should have the following information stored in a file:

• The user account's Oracle Cloud ID (OCID).


1. Login to the Oracle Cloud Infrastructure console.


3. Click Identity > Users.

4. Click Create User.5. Enter the following to create the user:

a) For NAME, enter a name for the user.

The name can be anything meaningful to your organization, e.g., pc_clouduser.

b) For DESCRIPTION, enter a description of the user.

6. Click Create.The user account is created and displayed in the users list.

7. Click Show located under the name of the user.The user account's OCID is displayed.

8. Click Copy to copy the OCID.

9. Store the OCID to a file.

The OCID is used later to create a cloud account in Control.

10. Click Group from the menu located on the left-hand side of the web page.

11. Click Create Group.

12. Enter the following:

a) For Name, enter Administrators.

b) For Description, enter a description for the group.

c) Click Submit.The group is created and is displayed in the Groups list.


https://docs.cloud.oracle.com/iaas/Content/GSG/Tasks/addingusers.htm

https://docs.cloud.oracle.com/iaas/Content/General/Concepts/identifiers.htm


13. Click on the name of the group.

14. Click Add User to Group.

a) For User, select the user that was previously created (e.g., pc_clouduser).

b) Click Add.


16. Click Identity > Policies.

17. Click Create Policy.

a) For Name, enter a name for the policy.

b) For Policy Versioning, enable Keep Policy Current.c) For Policy Statements, enter: ALLOW GROUP Administrators to manage all-resources IN

TENANCY

d) Click Create.

8.11.2 Generating an SSH Public for the Oracle Cloud UserUse OpenSSL to create a private and public key in a PEM format for the previously created Oracle Clouduser.

If you're using Windows, you'll need to install Git Bash for Windows and run the commands with thattool.

Use OpenSSL to generate an API signing key pair in a PEM format. Once this task is complete, youshould have the following information stored in a file:

• The RSA public key's fingerprint.

1. Generate a private key by executing the following command:openssl genrsa -out oracle_private_key.pem 2048

Note: It is recommended to change the permissions on this file so that only you haveread/write access.

The private key is used later to create a cloud account in in Control.

2. Generate the public key by executing the following command:openssl rsa -pubout -in oracle_private_key.pem -out oracle_public_key.pem



5. Click Identity > Users.

6. Click the name of the previously created user (e.g., pc_clouduser).

7. Click Add Public Key.

8. Copy and paste the contents of the public RSA key file.

9. Click Add.A fingerprint is generated.



Figure 26: Public Key Fingerprint

10. Copy the fingerprint.

11. Store the fingerprint to a file.

The fingerprint is used later to create a cloud account in Control.

8.11.3 Obtaining the Root Compartment IdentifierObtain the Oracle Cloud identifier of the root compartment so that you can later create a cloud accountin Control.


Oracle documentation can be found at Understand Compartments.

When your tenancy is provisioned, a root compartment is created for you. Compartments can be usedto organize and isolate your resources to make it easier to manage and secure access to them. Yourroot compartment holds all of your cloud resources. You can think of the root compartment like a rootfolder in a file system. The first time you sign in to the Oracle Cloud Console and select a service, youwill see your root compartment. All the resources required for cloud bursting will be contained in thisroot compartment. You will need the root compartment's resource identifier to add an Oracle cloudaccount to Control. Once this task is complete, you should have the following information stored in afile:

• The compartment's Oracle Cloud ID (OCID).



3. Click Identity > Compartments.

4. Click Show located under the name of the root compartment.The compartment's OCID is displayed.


https://docs.cloud.oracle.com/iaas/Content/GSG/Concepts/settinguptenancy.htm#two





8.11.4 Obtaining the Tenancy IdentifierObtain the Oracle Cloud identifier of the tenancy that you can later create a cloud account in Control.


You will need the tenancy's Oracle Cloud identifier to add an Oracle cloud account to Control. Once thistask is complete, you should have the following information stored in a file:

• The tenancy's Oracle Cloud ID (OCID).



3. Click Administration > Tenancy Details.

4. Under Tenancy Information, click Show located to the right of OCID:

Figure 27: Tenancy OCID

The tenancy's OCID is displayed.




8.11.5 Adding an Oracle Cloud AccountAdd an existing Oracle Cloud account to Control so that you can create a cloud bursting scenario or anappliance.

You must have created an Oracle Cloud user and added a public key for the user.

You should have the following information stored in a file:

• User OCID generated when creating the Oracle user.

• Tenancy OCID.



• Root compartment OCID.

• Fingerprint of the public key generated for the Oracle user.

• RSA private key that was generated for the Oracle user.

This information is necessary to create a cloud account in Control. If you do not have this informationsee Creating an Oracle Cloud User Account, Generating an SSH Public for the Oracle Cloud User,Obtaining the Root Compartment Identifier and Obtaining the Tenancy Identifier.




4. Click .



The name can be anything meaningful to your organization, e.g., oracle_cloudaccount.

b) For User OCID, enter the user OCID generated when creating the Oracle cloud user.

c) For Tenant OCID, enter the tenancy OCID .

d) For Compartment OCID, enter the root compartment OCID.

e) For Fingerprint, enter the fingerprint generated when adding the public SSH key for theOracle user.

f) For Private Key, enter the generated RSA private key.


8.11.6 Creating a Virtual Cloud NetworkCreate a Virtual Cloud Network (VCN) so that you can perform cloud bursting.


Oracle documentation can be found at Overview of Networking and Creating a Virtual Cloud Network.

An Oracle VCN must be created to create a bursting scenario in Control.



3. Click Networking > Virtual Cloud Networks.

4. Choose a region based on the geographical location of your users.


https://docs.cloud.oracle.com/iaas/Content/Network/Concepts/overview.htm?tocpath=Services%7CNetworking%7C_____0

https://docs.cloud.oracle.com/iaas/Content/GSG/Tasks/creatingnetwork.htm


Figure 28: Select a Region

5. ClickCreate Virtual Cloud Network.

6. Enter the following to create a VNC:

a) For CREATE IN COMPARTMENT, select the root compartment.

b) For NAME, enter a name for the VNC.

The name can be anything meaningful to your organization, e.g., bursting_vnc.

c) Enable CREATE VIRTUAL CLOUD NETWORK PLUS RELATED RESOURCES.

Choosing this option automatically creates a VCN with a CIDR block 10.0.0.0/16, an internetgateway, a route rule to enable traffic to and from the internet gateway, the default securitylist, the default set of DHCP options, and one public subnet per availability domain.

7. Click Create Virtual Cloud Network.A summary of the VNC, internet gateway, default route table and subnets is displayed.

8. Click Close.A list of VNCs is displayed.

9. Click the name of the VNC.The subnets are displayed. A subnet is created for each availability domain (data center) locatedin the previously selected region.

Figure 29: Virtual Machine Subnets and Associated Availability Domain

Next, create a virtual machine that is hosted in one of the region's availability domains.



8.11.7 Creating a Virtual MachineCreate a CentOS 7 virtual machine (VM) in Oracle Cloud.

You must already have signed up for an Oracle Cloud account, have an associated tenancy and havecreated a VNC in a region that is geographically close to your users. The VNC should have a subnetassociated to each of the region's availability domains. If you have not created a VNC see Creating aVirtual Cloud Network.

Virtual machines are hosted in availability domains (data centers) located in a region and are basedon predefined VM shapes. Before proceeding, determine the VM shape that your site requires for cloudbursting based on the number of CPUs, memory, disk space, network bandwidth, and virtual networkinterface cards. While selecting the shape for a VM, consider the nature of the applications that you planto deploy on the instance, the number of users that you expect to use the applications, and also howyou expect the load to scale in the future. Remember to also factor in the CPU and memory resourcesthat are necessary for the operating system.

Next, verify that your Oracle tenancy has the appropriate service limits for the VM shape and theregion's availability domains Checking Tenancy Service Limits .



3. Click Compute > Instances.

4. Choose the region where the previously created VNC is hosted.


5. For COMPARTMENT, select the root compartment.


https://docs.cloud.oracle.com/iaas/Content/Compute/References/computeshapes.htm?Highlight=shapes#vmshapes


Figure 31: Choose Compartment

6. Click Create Instance.

7. Enter the following to create a virtual machine:

a) For NAME, enter a name for the VM.

The name can be anything meaningful to your organization, e.g., bursting_vm.

b) For AVAILABILITY DOMAIN, choose one of the region's availability domains.

The virtual machine is hosted in the chosen availability domain (data center). Choose theavailability domain that best suits your site's cloud bursting requirements based on themachine type of the virtual machine (VM shape) and service limits.

c) For BOOT VOLUME, enable ORACLE-PROVIDED OS IMAGE.

d) For IMAGE OPERATING SYSTEM, choose CentOS 7.

e) For SHAPE TYPE, enable VIRTUAL MACHINE.

f) For SHAPE, select a VM shape.

Choose the VM shape that best suits your site's cloud bursting requirements based onnumber of CPUs, memory, disk space, and network bandwidth.

g) For IMAGE VERSION, select the latest available one.

h) For BOOT VOLUME CONFIGURATION, enable CUSTOM BOOT VOLUME SIZE and enter aboot volume size in GBs.

i) Enable PASTE SSH KEYS and copy the SSH public key (i.e., id_rsa.pub) for the userpbsadmin from the PBS Server and paste it.

j) For VIRTUAL CLOUD NETWORK, choose the VCN that was created for cloud bursting.

k) For SUBNET, choose the subnet associated with the previously chosen availability domain.

8. Click Create Instance.

Creating the virtual machine may take some time. It is done when the state is "Running".



Figure 32: Running Virtual Machine

Checking Tenancy Service LimitsCheck the service limit for your tenancy before creating the components necessary for cloud bursting.


When you sign up for Oracle Cloud Infrastructure, a set of service limits are configured for yourtenancy. The service limit is the quota or allowance set on a resource. For example, your tenancy isallowed a maximum number of compute instances (virtual machines) per availability domain. Theselimits are generally established with your Oracle sales representative when you purchase Oracle CloudInfrastructure. Oracle documentation can be found at Service Limits and Regions and AvailabilityDomains.

When you reach the service limit for a resource, you receive an error when you try to create a newresource of that type. You cannot create a new resource until you are granted an increase to yourservice limit or you terminate an existing resource.

View your tenancy's limits to ensure that there are sufficient resources available in a region's availabilitydomains.


2. Choose the region where the previously created VNC is hosted.


https://docs.cloud.oracle.com/iaas/Content/General/Concepts/servicelimits.htm

https://docs.cloud.oracle.com/iaas/Content/General/Concepts/regions.htm

https://docs.cloud.oracle.com/iaas/Content/General/Concepts/regions.htm




4. Click Governance > Service Limits.

5. Scroll down to the Service Limits section.

6. Click Compute.

Availability domains (data centers) for the region are displayed. For each resource (VM shape) thenumber of nodes that can be burst in the corresponding availability domains are displayed. In thebelow example, three nodes can be burst in each data center in the us-phoenix-1 region for theVM Standard1.1 shape.

Figure 34: Virtual Machine Type Limits

7. Verify that the appropriate service limits are set for your tenancy based on the VM shape chosenfor the virtual machine and the region's availability domains.

To request an increase a service limits for your tenancy see Requesting a Service Limit Increase.


Log into an Oracle-provided Oracle Linux instance as the default user, "opc". This user has sudoprivileges.


https://docs.cloud.oracle.com/iaas/Content/General/Concepts/servicelimits.htm#three


Note: You will need to set the default_user parameter of the PBS Professional burstinghook configuration file to this username.

For more information see Accessing an Instance from UNIX and UNIX-Like Systems.



2. SSH into the virtual machine as the user "opc" using the External IP assigned to the VM.






7. Save the file.

8. Edit the file PBS_HOME/mom_priv/config where PBS_HOME is the home directory of PBS.


10. Save the file.





https://docs.oracle.com/en/cloud/iaas/compute-iaas-cloud/stcsg/accessing-oracle-linux-instance-using-ssh.html#GUID-01F76367-64BC-48EA-8E80-8248776757EE





Oracle documentation can be found at Managing Custom Images.



3. Click Compute > Instances.A list of virtual machines is displayed.

4. Click the name of the virtual machine created for cloud bursting.

5. Click Create Custom Image.

6. Enter the following to create a custom image:

a) For CREATE IN COMPARTMENT, select the root compartment.

b) For NAME, enter a name for the image.

The name can be anything meaningful to your organization, e.g., bursting_image.

7. Click Create Custom Image.

It may take some time to create the image. Do not proceed until the image is created.


8.11.10 Creating a Cloud Bursting ScenarioCreate an Oracle bursting scenario so that PBS execution nodes can be burst in the cloud.

Open a browser window and login to the Oracle Cloud Infrastructure console. Open a second browserwindow and login to Control.



2. Click Cloud.






c) For Cloud Account, choose a previously created Oracle cloud account from the list.


https://docs.cloud.oracle.com/iaas/Content/Compute/Tasks/managingcustomimages.htm?tocpath=Services%7CCompute%7CManaging%20Custom%20Images%7C_____0



d) For Select a region, select the availability domain where the cloud bursting virtual machine ishosted.

Figure 36: Virtual Machine Availability Domain

6. Click Next.Enter information about the Oracle resources necessary to burst in the cloud.



9. Navigate to the Oracle Cloud Infrastructure browser window.


11. Click Networking > Virtual Cloud Networks.

12. Click the name of the VCN create for cloud bursting.

13. Locate the subnet associated with the availability domain where the cloud bursting virtual machineis hosted.

Figure 37: Subnet and Associated Availability Domain

14. Click Show located under the name of the subnet.



The subnet's OCID is displayed.



17. For Subnet OCID, paste the subnet's OCID copied from Oracle.



19. Navigate to the Oracle Cloud Infrastructure browser window.


21. Click Compute > Custom Images.A list of custom images is displayed.

22. Locate the custom image created from the cloud bursting virtual machine.

23. Click the Show link below the name of the image to view the OCID.

24. Click Copy to copy the OCID of the image.


26. For OS Image URI, paste the link copied from Oracle.















8.12 Configuring Orange Cloud Flexible Engine forCloud BurstingCreate the necessary components in Orange Cloud Flexible Engine that are required for cloud burstingand then create a cloud bursting scenario.

Purchase an Orange Business Services account and use your Orange ID and password to access theFlexible Engine console. For more information visit Orange Cloud.

8.12.1 Creating an Orange Cloud Flexible Engine UserAccountCreate an Orange Cloud Flexible Engine user account so that you can later create a cloud account inControl.

You must already have signed up for an Orange Cloud account. You should be able to login to theOrange Cloud Customer space (https://selfcare.cloud.orange-business.com/) with the credentialsprovided to you with your Orange Cloud account. You will also be provided with a Domain name whenyou sign up for your Orange Cloud account.

An Orange Cloud user account must be created to add an Orange Cloud Flexible Engine account toControl. Once this task is complete, you should have the following information stored in a file:

• Username and password

• API password


1. Go to the Orange Cloud Customer space login page.

2. Enter your Orange Cloud credentials.

3. Click Your services.

Figure 38: Orange Cloud Customer Space Services

4.In the navigation bar on the top click Users

5. Click Add user.6. Enter the following user details:


https://cloud.orange-business.com/en/orange-cloud

https://selfcare.cloud.orange-business.com/


a) For Civility, choose Mrs or Mr

b) For Last name, enter the user's last name.

c) For First name, enter the user's first name.

d) For Login, enter a login name for the user.

e) For Email, enter the user's email address.

f) For Phone number, enter the user's phone number.

g) For Mobile phone, enter the user's mobile phone number.

h) For Preferred language, choose the language in which the application should be displayed.

i) Click next.7. In the Roles section enter these details

a) For Billing, choose Visitor.b) For Contracts, choose Account Manager.c) For Dashboard, choose Visitor.d) For Documents, choose Visitor.e) For Orders, choose Visitor.f) For Services, choose Visitor.g) For Subscriptions, choose Visitor.h) For Support, choose Visitor.i) For Users. choose Manager.j) For Flexible Engine Console. choose admin.

k) Click next.8. In the Summary section review your choices.

Click previous to edit your choices entries.

9. Click finish.The new user account is created and displayed in the list of users. Emails are sent to the emailaddress you specified. The emails will contain:

1. Orange ID (Domain Name).

2. Link to set Orange Password.

3. Link to access Cloud Customer Space.

4. Link to log in to the Flexible Engine Console.

5. Link to define your API password.

10. Click the link in the email to set your Orange Password.

11. Click the link in the email to set your API Password.

12. Store the API Password to a file.

The API Password is used later to create a cloud account in Control.



8.12.2 Selecting a RegionDefine a region in the Orange Cloud Flexible Engine console to set up the infrastructure for cloudbursting.

A region is a geographic area where resources used by your ECSs are located. ECSs in the same regioncan communicate with each other over an intranet, but ECSs in different regions cannot. Before settingup the infrastructure for cloud bursting, it is important to ensure that all the resources are defined inthe same region. An Authorization URL is required for adding the Orange Cloud Flexible Engine cloudaccount in Control. This is based on the region selected.

1. Login to the Orange Cloud Flexible Engine console.

2. In the navigation bar on the top select the region for setting up your infrastructure.


Note: For Auth (IAM) URL (Authorization URL), enter the url in the following formatbased on the region you chose in the Orange Cloud Flexible Engine console.

https://iam.<orange region>.<console link> e.g. https://iam.eu-west-0.prod-cloud-ocb.orange-business.com

3. Store the region and Auth (IAM) URL in a file. You will use this to register the account in Control.

8.12.3 Adding an Orange Cloud Flexible Engine AccountAdd an existing Orange Cloud Flexible Engine account to Control so that you can create a cloud burstingscenario or an appliance.

You must have:

• Created an Orange Cloud Flexible Engine user account.

• Obtained the Domain name for the Orange Cloud Flexible Engine account.

• Obtained the API password for the user.

• Defined the region where you will deploy your cloud infrastructure.

• Obtained the Auth (IAM) URL based on the selected region.

This information is necessary to create a cloud account in Control. If you do not have this informationsee Creating an Orange Cloud Flexible Engine User Account and Selecting a Region.






4. Click .



The name can be anything meaningful to your organization, e.g., orange_cloudaccount.

b) For Auth (IAM) URL, enter the url in the following format based on the region you chose inthe Orange Cloud Flexible Engine console.https://iam.<orange region>.<console link> e.g. https://iam.eu-west-0.prod-cloud-ocb.orange-business.com

c) For Domain name, enter the Domain Name you used to login to your Orange account. It isthe same as your Customer ID.

d) For User name, enter the username of the user account you created in the Orange CloudCustomer space.

e) For Password, enter the API password you generated after creating the user.

6. Click Create Account.

8.12.4 Checking Orange Cloud Flexible Engine AccountService QuotaView the your Orange Cloud Flexible Engine resource usage and limits from the console.

Quotas are used to limit the number of resources available to users. It is important to ensure you arenot exceeding your quota while setting up the resources for cloud bursting. If the existing resourcequota cannot meet your service requirements, you can submit a work order to increase your quota.Once your application is approved, Orange Cloud Flexible Engine will update your resource quotaaccordingly and send you a notification.


2. In the navigation bar on the top right hand side, click .

Figure 40: Check Quotas for Resources

Information about resources usage and availability is displayed.



Figure 41: Resources Quotas Table

8.12.5 Creating a Virtual Private CloudCreate a Virtual Private Cloud (VPC) so that you can perform cloud bursting.

Orange Cloud Flexible Engine documentation for VPC can be found at:Virtual Private CloudDocumentation


2. In the top navigation bar select the region where you wish to deploy your cloud infrastructure.

3. From the Network section click Virtual Private Cloud.

4. Click + Create VPC.

5. In the Basic Information section:

a) For Region, ensure the VPC is the same region as the other resources.

b) For Name, enter a name for the VPC.

c) For CIDR Block, enter CIDR values for the VPC.

6. In the Subnet Settings section:

a) For the Subnet Settings choose the AZ (Availability Zone) as the same as the region.

b) For Subnet Name, enter a name to match the VPC Name.

c) Enter CIDR Block for Subnet.

d) For Advanced Settings, click Default.7. Review the Configuration information.

8. Click Create Now.

9. Once the VPC is created, click the Back to VPC List.10. Click Security Group in the left hand side menu.

11. Click + Create Security Group.

12. For Name, enter a name for the Security Group.

13. For Description, enter a suitable Description.


https://docs.prod-cloud-ocb.orange-business.com/en-us/vpc_dld/index.html

https://docs.prod-cloud-ocb.orange-business.com/en-us/vpc_dld/index.html


14. Click OK

Note: By default, the Outbound and Inbound traffic over IPv4 is open. You can addfirewall rules to this security group if required.

8.12.6 Creating a Virtual MachineCreate a Elastic Cloud Server (ECS) using a public image provided by Orange Cloud Flexible Engine forcloud bursting.

You must have already created a Virtual Private Network with a subnet and assigned a Security Group.If you do not have this information, see Creating a Virtual Private Cloud


2. In the Computing section and click Elastic Cloud Server.3. From the menu on the left hand side click Key Pair.4. Click + Create Key Pair.5. Enter a Name for the Key Pair.

6. Click OK.

7. Save the Key Pair (.pem) file to your local disk.

8. Click OK to confirm that you have downloaded the Key Pair file.

9. From the menu on the left hand side, click Elastic Cloud Server.10. Click + Create ECS.

11. For Region, click the region you selected for setting up the infrastructure.

12. For AZ (Availability Zone), select the AZ related to the region.

13. In the Specifications section:

a) For ECS type, click one of the flavor names.

Note: Orange Cloud Flexible Engine provides a set of predefined ECS typesfor specific requirements. Click a flavor name to get the list of availableconfigurations.

b) Review the specifications you have selected.

14. In the Image section:

a) Click Public image.

b) From the drop down menu select CentOS

c) From the version drop down menu select Select OBS_U_CentOS_7.2(40GB)

15. In the Disk section select the defaults.

16. In the VPC section:

a) For VPC, select the VPC you created from the drop down menu.

b) For NIC, choose the default primary NIC.

c) For Security Group, select the Security Group you created for the VPC.



d) For EIP, click Automatically assigne) For Bandwidth, specify it as 5 Mbit/s.

17. For Login Mode, select the Key Pair you generated earlier from the drop down menu.

18. For Auto Recovery, click Enable.

19. For Advanced Settings, click Do not configure.

20. For ECS Name, enter a name.

21. For Quantity, specify 1.

22. Review the Current Configuration.


24. Review the Specfications.

25. Click Submit.The ECS (Virtual Machine) is created and diplayed in the list of ECS.


When you log in to a newly created Linux ECS for the first time, use an SSH key as user cloud (withoutdefault password) and make sure that the ECS has had an EIP bound. Use the key file used during ECScreation to log in to the ECS using SSH. For more information about logging into the Linux ECS, refer tothe Elastic Cloud Server User Guide.



2. Login to the Flexible Engine console.

3. In the Computing section, click Elastic Cloud Server.4. In the search box above the upper right corner of the ECS list, enter the ECS name, IP address, or

ID, and click .

5. Click the name of the target ECS.

6. The page providing details about the ECS is displayed.

7. Copy the Public IP address (Exteral IP) of the ECS.

8. SSH into the VM as the user "cloud" using the External IP assigned to it and the key file yougenerated for the Key Pair that was specified while creating the VM (ECS).






https://docs.prod-cloud-ocb.orange-business.com/en-us/ecs_dld/index.html



13. Save the file.




16. Save the file.





You must have already created a ECS (VM), installed the PBS MoM on the VM and configured it per yoursite's environment.

Orange Cloud Flexible Engine documentation can be found at Creating a Linux Private Image Using anECS. Before creating the Image from the ECS, you must have.

• A Linux ECS in the Stopped state.

• Configured DHCP for the NICs of the ECS

• Configured Network attributes of the ECS

• Detached Data Disks from the ECS

1. Login to the Flexible Engine console.

2. In the Computing section, click Image Management Service.

3. On the Image Management Service page, click + Create Private Image.

4. In the Image Type and Source section, .

a) For Type, click System disk imageb) For Source, click ECSc) Select the target ECS from the ECS list.

5. Set the required information, such as Name and Description.


7. Confirm the parameters and click Submit.8. Switch back to the Image Management Service page to view the image status.

The time required for creating an image varies depending on the image file size. Generally, ittakes about 20 minutes to create an image. The image is successfully created when its imagestatus changes to Normal.

Note: Do not perform any operation on the selected ECS or its associated resourcesduring image creation.



https://docs.prod-cloud-ocb.orange-business.com/en-us/usermanual/ims/en-us_topic_0030713154.html






8.12.9 Creating a Cloud Bursting ScenarioCreate an Orange Cloud Flexible Engine bursting scenario so that PBS execution nodes can be burst inthe cloud.

Open a browser window and login to the Orange Cloud Flexible Engine console. Open a second browserwindow and login to Control.



2. Click Cloud.






c) For Cloud Account, choose a previously created Orange Cloud Flexible Engine account fromthe list.

d) For Select a region, select the availability domain where the cloud bursting virtual machine ishosted.

Figure 42: Virtual Machine Availability Domain

6. Click Next.7. For Domain name, provide the domain name of your organization for the cloud bursting network.

altair.com, burst.altair.com


9. Navigate to the Orange Cloud Flexible Engine console browser window.

10. Click Service List.




Figure 43: Orange Cloud Flexible Engine Console

11. Under Network, click Virtual Private Cloud.

12. Click Virtual Private Cloud from the menu located on the left-hand side of the web page.

13. Click the name of the VPC you created for cloud bursting.

14. Click the name of the Subnet for the VPC.

15. Copy the Subnet ID.


17. For Subnet ID, paste the subnet's ID copied from the Orange Cloud Flexible Engine console.

18. In the Orange Cloud Flexible Engine console browser window click Security Groupfrom the lefthand side menu.

19. Click the name of the Security Group you created for the VPC.

20. Copy the ID of the Security Group.


22. For Security Group ID, paste the security group's ID you copied from the Orange Cloud FlexibleEngine console.



24. Navigate to the Orange Cloud Flexible Engine console browser window.


Figure 44: Orange Cloud Flexible Engine Console

26. Under Computing, click Image Management Service.

27. Click the Private Images tab.

28. Click the name of the VM image you created for cloud bursting.

29. Copy the ID of the image.


31. For OS Image URI, paste the image ID you copied from the Orange Cloud Flexible Engine console.

















8.13 Configuring HUAWEI Cloud for Cloud BurstingCreate the necessary components in HUAWEI Cloud that are required for cloud bursting and then createa cloud bursting scenario.

Before you can begin creating the components necessary for cloud bursting, you must create andactivate an HUAWEI Cloud account.

8.13.1 Obtain the HUAWEI Cloud AdministratorCredentialsGet the credentials for the HUAWEI Cloud administrative user account so that you can later use thisinformation to create a cloud account in Control.

You must already have signed up for an HUAWEI Cloud account.

When subscribing to HUAWEI Cloud, an administrative user account that has permissions for all systemoperations is automatically created. Use this user account to add an HUAWEI Cloud account to Control.Reset the password for this account when the password is unknown.

Note: Alternately, you can create a new user and give the user administrative privileges bysetting its User Group to "admin".

1. Login to the HUAWEI Cloud Console.


Figure 45: HUAWEI Cloud Console

3. Under Management & Deployment, click Identity and Access Management.4. Click Users from the menu located on the left-hand side of the web page.

A list of users is displayed.

5. Click located next to a username to display the user's details.The user account listed as an "admin" is the account to use to create the cloud account in Control.



Figure 46: User Details

6. If you do not know the password for the administrative user account, click Set Credentials.

Figure 47: Set Password

a) Enable Set manually.

b) For Password, enter a password for the user account.

c) For Confirm Password, enter the password a second time.

d) Click OK.

You may have to confirm the password change either by email or by a SMS text.

8.13.2 Adding an HUAWEI Cloud AccountAdd an existing HUAWEI Cloud account to Control so that you can create a cloud bursting scenario.




4. Click .



The name can be anything meaningful to your organization, e.g., huawei_cloudaccount.

b) For Auth (IAM) URL, enter https://iam.ap-southeast-1.myhwclouds.com.

c) For Domain name, enter the Domain Name provided when your subscription to HUAWEICloud was created.

If you do not know your Domain name, contact HUAWEI Cloud support.

d) For Username, enter a username that is an HUAWEI Cloud administrator.

e) For Password, enter the password for the HUAWEI Cloud administrator.




8.13.3 Checking HUAWEI Cloud Account Service QuotasView your OTC account resource usage and limits.

Quotas are used to limit the number of resources available to users. It is important to ensure you arenot exceeding your quota while setting up the resources for cloud bursting. If the existing resourcequota cannot meet your service requirements, you can submit a work order to increase your quota.Once your application is approved, HUAWEI Cloud will update your resource quota accordingly and sendyou a notification.

1. Login to the HUAWEI Cloud console.

2. Click Resources > My Quota.

Figure 48: View Quotas


Figure 49: Resources Quotas


HUAWEI Cloud documentation for creating a VPC can be found at: Creating a VPC and Regions and AZs.



https://support.huaweicloud.com/en-us/usermanual-vpc/en-us_topic_0013935842.html

https://support.huaweicloud.com/en-us/productdesc-ecs/en-us_topic_0030828255.html







a) For Region, select a region.

Regions are geographic areas isolated from each other. Resources are region-specific andcannot be used across regions through internal network connections. For low network latencyand quick resource access, select the nearest region.


c) For CIDR Block, enter an address range for the network using CIDR notation.


a) For the Subnet Settings choose the AZ (Availability Zone).

An Availability Zone is a physical location where resources use independent power suppliesand networks. AZs are physically isolated and AZs in the same VPC are interconnectedthrough an internal network.

b) For Subnet Name, enter a name for the subnet.

c) For CIDR, enter an address range for the subnet using CIDR notation.

d) For Advanced Settings, click Default.7. Click Create Now.

8. Click Back to VPC List.9. Click Security Group in the left hand side menu.


11. For Name, enter a name for the security group.

12. For Description, enter a suitable description.

13. Click OK

Note: By default, the Outbound and Inbound traffic over IPv4 is open. You can addfirewall rules to this security group if required.

8.13.5 Creating a Virtual MachineCreate a Elastic Cloud Server (ECS) using a public image provided by HUAWEI Cloud for cloud bursting.

You must have already created a Virtual Private Network with a subnet and assigned a Security Group.If you do not have this information, see Creating a Virtual Private Cloud.

HUAWEI Cloud documentation for creating an ECS (virtual machine) can be found at Purchase an ECS.


https://support.huaweicloud.com/en-us/qs-ecs/en-us_topic_0092494193.html#EN-US_TOPIC_0092494193__section761474661113





3. Under Computing, click Elastic Cloud Server.4. Click Key Pair from the menu located on the left-hand side of the web page.

5. Click + Create Key Pair.6. For Name, enter a name for the key pair.

7. Click OK.

8. Save the key pair (.pem) file to your local disk.

9. Click OK to confirm that you have downloaded the key pair file.

10. Click Elastic Cloud Server from the menu located on the left-hand side of the web page.

11. Click Buy ECS.

12. For Billing Mode, click Pay-per-use.

13. For Region, select the same region that was chosen for the previously created VPC.

14. For AZ (Availability Zone), select the same AZ that was chosen for the previously created VPC.

15. In the Type section:

a) Choose a ECS type category:

• General computing

• General computing-plus

• Memory-optimized

• Large-memory

• High-performance computing

• Disk-intensive

b) For ECS type, click one of the flavor based on the needs of your site.



b) For Select an OS, select CentOS.

c) For Select an OS version, select CentOS 7.2 64bit(40GB).

17. In the Disk section, select your system disk requirements.


a) For VPC, select the VPC you created for cloud bursting.The NIC information is automatically populated.

b) For Security Group, select the security group you created for cloud bursting.

c) For EIP, click Automatically assignd) For Bandwidth, specify it as 5 Mbit/s.

19. For Login Mode, select Key Pair.



20. For Key Pair, select the key pair file you generated earlier.

21. For Advanced Settings, click Not required.

22. For ECS Name, enter a name for the virtual machine.


24. Click Next.25. Review the specifications.

26. Enable the I have read and agree to the Huawei Image Disclaimer checkbox.

27. Click Submit Application.

28. Click Back to ECS List.It may take some time to create the virtual machine. Once the ECS is created it is displayed in theECS list.


A cloud-init script is provided during the creation of the bursting scenario. This script is ran after acloud node is burst and is used to set up the system environment so that nodes burst in the cloud cancommunicate with the PBS Professional Server, therefore installation of cloud-init on the VM is required.For more information about installing cloud-init, refer to the Cloudinit Documentation and InstallingCloud-Init.




3. Under Computing, click Elastic Cloud Server.4. Copy the Public IP address (External IP) of the ECS.

Figure 53: VM IP Address


6. SSH into the virtual machine as the user "root "using the Public IP address assigned to the VM byentering a command similar to:ssh -i /path/my-key-pair.pem root@PublicIP



https://support.huaweicloud.com/en-us/usermanual-ims/en-us_topic_0030730603.html



where /path/my-key-pair.pem is the location of the .pem file downloaded while creating thevirtual machine and PublicIP is the public IP address of the virtual machine.




10. Save the file.




13. Save the file.





You must have already created a ECS (virtual machine), installed the PBS MoM on the ECS andconfigured it per your site's environment.

HUAWEI Cloud documentation can be found at Creating a Linux Private Image.





4. Click + Create Image.

5. For Region, select the same region that was chosen for the previously created VPC and ECS.

6. In the Image Type and Source section, .

a) For Type, click System disk image.

b) For Source, click ECS.

c) Select the virtual machine created for cloud bursting.

d) For Name, enter a name for the virtual machine.

7. Click Next.8. Review the specifications.




9. Enable the I have read and agree to the Statement of Commitment to Image Creationand Huawei Image Disclaimer checkbox.

10. Click Submit.11. Click Back to Image List.




8.13.8 Creating a Cloud Bursting ScenarioCreate an HUAWEI Cloud bursting scenario so that PBS execution nodes can be burst in the cloud.

Open a browser window and login to the HUAWEI Cloud console. Open a second browser window andlogin to Control.



2. Click Cloud.






c) For Cloud Account, choose a previously created HUAWEI Cloud account from the list.

d) For Select a region, select the same region that was chosen for the previously created VMand VPC.

6. Click Next.Enter information about the HUAWEI Cloud resources necessary to burst in the cloud.






9. Navigate to the HUAWEI Cloud console browser window.






14. Click the name of the VPC's subnet.



17. For Subnet ID, paste the subnet's ID copied from HUAWEI Cloud console.

18. In the HUAWEI Cloud console browser window click Security Group from the menu located onthe left-hand side of the web page.

19. Click the name of the security group you created for the VPC.

20. Copy the ID of the security group.


22. For Security Group, paste the security group's ID you copied from the HUAWEI Cloud console.



24. Navigate to the HUAWEI Cloud console browser window.






29. Copy the ID of the image.


31. For OS Image URI, paste the image ID you copied from the HUAWEI Cloud console.

















8.14 Configuring Open Telekom Cloud for CloudBurstingCreate the necessary components in Open Telekom Cloud (OTC) that are required for cloud bursting andthen create a cloud bursting scenario.

Before you can begin creating the components necessary for cloud bursting, you must create andactivate an OTC Cloud account.

8.14.1 Obtain the OTC Administrator CredentialsGet the credentials for the OTC administrative user account so that you can later use this information tocreate a cloud account in Control.

You must already have signed up for an OTC account.

When subscribing to OTC, an administrative user account that has permissions for all system operationsis automatically created. Use this user account to add an OTC cloud account to Control. Reset thepassword for this account when the password is unknown.

Note: Alternately, you can create a new user and give the user administrative privileges bysetting its User Group to "admin".

1. Login to the OTC Console.


Figure 57: Service List

3. Under Management & Deployment, click Identity and Access Management.4. Click Users from the menu located on the left-hand side of the web page.

A list of users is displayed.

5. Click located next to a username to display the user's details.The user account listed as an "admin" is the account to use to create the cloud account in Control.



Figure 58: User Details

6. If you do not know the password for the admin user account, click Set Credentials.

Figure 59: Set Password

a) Enable Set manually.

b) For Password, enter a password for the user account.

c) For Confirm Password, enter the password a second time.

d) Click OK.

You may have to confirm the password change either by email or by a SMS text.

8.14.2 Adding an OTC AccountAdd an existing OTC Cloud account to Control so that you can create a cloud bursting scenario.




4. Click .



The name can be anything meaningful to your organization, e.g., orange_cloudaccount.

b) For Auth (IAM) URL, enter: https://iam.eu-de.otc.t-systems.com/v3.

c) For Domain name, enter the OTC Domain Name you used to login to the OTC console.

d) For Username, enter a username that is an OTC administrator.

e) For Password, enter the password for the OTC administrator.




8.14.3 Checking OTC Account Service QuotasView your OTC account resource usage and limits.

Quotas are used to limit the number of resources available to users. It is important to ensure you arenot exceeding your quota while setting up the resources for cloud bursting. If the existing resourcequota cannot meet your service requirements, you can submit a work order to increase your quota.Once your application is approved, OTC will update your resource quota accordingly and send you anotification.


2. Click .

Figure 60: View Quotas


Figure 61: Resource Quotas


OTC documentation for creating a VPC can be found at: Creating a VPC and Regions.



https://docs.otc.t-systems.com/en-us/usermanual/vpc/en-us_topic_0013935842.html

https://docs.otc.t-systems.com/en-us/usermanual/tms/en-us_topic_0071335170.html



Figure 62: OTC Service List




a) For Region, select a region.

Figure 63: Regions

A region is a geographical areas and can comprise one or more availability zones (AZs).A region is completely isolated from other regions. Only AZs in the same region cancommunicate with one another through an internal network.


c) For CIDR Block, enter an address range for the network using CIDR notation.


a) For Subnet Name, enter a name for the subnet.

b) For CIDR, enter an address range for the subnet using CIDR notation.

c) For Advanced Settings, click Default.7. Click Create Now.

8. Click Back to VPC List.9. Click Security Group in the left hand side menu.


11. For Name, enter a name for the security group.

12. For Description, enter a suitable description.

13. Click OK.The security group rules are displayed.

14. Click the Inbound tab.

15. Click Add Rule.

a) For Protocol/Application, select TCP.

b) For Port, enter 22.

c) For Source, select IP Address. and enter 0.0.0.0/0.

d) Click OK.



8.14.5 Creating a Virtual MachineCreate a Elastic Cloud Server (ECS) using a public image provided by OTC for cloud bursting.

You must have already created a Virtual Private Network with a subnet and assigned a Security Group.If you do not have this information, see Creating a Virtual Private Cloud.

OTC documentation for creating an ECS (virtual machine) can be found at Create an ECS.




3. Under Computing, click Elastic Cloud Server.4. Click Key Pair from the menu located on the left-hand side of the web page.

5. Click + Create Key Pair.6. For Name, enter a name for the key pair.

7. Click OK.

8. Save the key pair (.pem) file to your local disk.

9. Click OK to confirm that you have downloaded the key pair file.

10. Click Elastic Cloud Server from the menu located on the left-hand side of the web page.

11. Click Create ECS.

12. For Region, select the same region that was chosen for the previously created VPC.

Figure 65: Regions

13. For AZ (Availability Zone), select an availability zone.

14. In the Specifications section:

a) Choose an ECS type category:

• General-purpose

• Dedicated general-purpose

• Memory-optimized

• Large-memory

• High-performance

• Disk-intensive

• GPU-accelerated

b) For ECS type, choose one of the flavors based on the needs of your site.



https://docs.otc.t-systems.com/en-us/usermanual/ecs/en-us_topic_0092494193.html#EN-US_TOPIC_0092494193__section761474661113



b) Select CentOS as the operating system.

c) Select Standard_CentOS7_latest(4GB) for the OS version..

16. In the Disk section, select your system disk requirements.


a) For VPC, select the VPC you created for cloud bursting.The NIC information is automatically populated.

b) For Security Group, select the security group you created for cloud bursting.

c) For EIP, click Automatically assignd) For Bandwidth, specify it as 5 Mbit/s.

18. For Login Mode, select Key Pair.19. For Key Pair, select the key pair file you generated earlier.

20. For Advanced Settings, click Do not configure.

21. For ECS Name, enter a name for the virtual machine.



24. Review the specifications.

25. Click Submit.26. Click Back to ECS List.

It may take some time to create the virtual machine. Once the ECS is created it is displayed in theECS list.


A cloud-init script is provided during the creation of the bursting scenario. This script is ran after acloud node is burst and is used to set up the system environment so that nodes burst in the cloud cancommunicate with the PBS Professional Server, therefore installation of cloud-init on the VM is required.For more information about installing cloud-init, refer to the Cloudinit Documentation and InstallingCloud-Init.




3. Under Computing, click Elastic Cloud Server.4. Copy the Public IP address (External IP) of the ECS.






Figure 67: VM IP Address


6. SSH into the virtual machine as the user "root "using the Public IP address assigned to the VM byentering a command similar to:ssh -i /path/my-key-pair.pem root@PublicIP

where /path/my-key-pair.pem is the location of the .pem file downloaded while creating thevirtual machine and PublicIP is the public IP address of the virtual machine.




10. Save the file.




13. Save the file.





You must have already created a ECS (virtual machine), installed the PBS MoM on the ECS andconfigured it per your site's environment.

OTC documentation can be found at Creating a Linux Private Image.





4. Click + Create System Disk Image.


https://docs.otc.t-systems.com/en-us/usermanual/ims/en-us_topic_0030713180.html


a) For Region, select the same region that was chosen for the previously created VPC and ECS.

Figure 69: Regions

b) For Source, click Server.c) For Server Type, click ECS.

d) For ECS, select the virtual machine created for cloud bursting.

Tip: If the virtual machine is not stopped, then stop it.

e) Click OK when prompted to verify that certain operations have been performed on the ECS.

You do not need to configure or optimize the ECS.

a) For Name, enter a name for the virtual machine.


6. Review the specifications.

7. Click Submit.8. Click Back to Image List.




8.14.8 Creating a Cloud Bursting ScenarioCreate an OTC bursting scenario so that PBS execution nodes can be burst in the cloud.

Open a browser window and login to the OTC Console. Open a second browser window and login toControl.



2. Click Cloud.









c) For Cloud Account, choose a previously created OTC cloud account from the list.

d) For Select a region, select the same region that was chosen for the previously created VMand VPC.

6. Click Next.Enter information about the OTC resources necessary to burst in the cloud.



9. Navigate to the OTC console browser window.






14. Click the name of the VPC's subnet.



17. For Subnet ID, paste the subnet's ID copied from the OTC console.

18. In the OTC console browser window, click Security Group from the menu located on the left-hand side of the web page.

19. Click the name of the security group you created for the VPC.

20. Copy the ID of the security group.


22. For Security Group, paste the security group's ID you copied from the OTC console.



24. Navigate to the OTC console browser window.








29. Copy the Image ID of the image.


31. For OS Image URI, paste the image ID you copied from the OTC console.















8.15 Testing Cloud Bursting with the CLIUse the PCLM Command Line Interface to test cloud bursting.

1. Ensure that the cloud infrastructure (VPN connectivity, networks, firewalls etc.) is configured.

2. The cloud account and bursting scenario must be configured in Control.

3. A Bursting Token (API Key) for the scenario must be generated.

4. By default, PCLM CLI is installed at PBS_EXEC/bin/pclm for PBS Professional 18.2.x. Otherwise,you will need to install the PCLM CLI manually.

8.15.1 Bursting an Execution Node in the CloudUse the PCLM CLI client for cloud bursting.

For more information about cloud provider instance/machine types seeCloud Provider Instance Types.

Cloud bursting can be accomplished from the command line by issuing a single command. You can thenget the status of the bursting activity and also capture the output in raw JSON format.

1. Execute the below command to burst a node in the cloud:pclm --api-endpoint=http://HOSTNAME:9980/pclm/ --api-key API_KEY bootstrapper --wait scenario\ burst '{"mom":[{"deployable_id": "MACHINE_TYPE", "count":COUNT}]}'

Where:

• HOSTNAME is the IP address or hostname of the machine where Cloud is installed.

• API_KEY is API Key generated by Control while creating the cloud bursting scenario.

• boostrapper is the microservices module for bursting.

• --wait causes the application to wait for the node to be burst.

• MACHINE_TYPE is the name of cloud provider instance type. This is the type of machine thatwill be burst. It must match one of the instance types defined by the cloud provider.

• COUNT indicates the number of nodes to burst.

2. Execute this command to view the status of the cloud bursting activity.pclm --api-endpoint=http://HOSTNAME:9980/pclm/ --api-key API_KEY bootstrapper scenario status

The status is displayed as a table. Use --raw to display the output of the command in JSONformat.

The status command will display the machine id, machine private ip and hostname.

The states are:

• RUNNING indicates deployment ongoing.

• SUCCESS indicates deployment finished, machine is ready

• CREATED indicates machines are created. This status is short-lived and rarely visible.

• ERROR indicates that something has gone wrong and machine deployment has failed.



• UNDEPLOYMENT RUNNING indicates that machines are being removed.

• UNDEPLOYMENT ERROR indicates that something has gone wrong while machine was beingremoved.

When an undeployment of a machine is finished, it disappears from the status table.

8.15.2 Unbursting an Execution Node in the CloudUse the PCLM CLI client for unbursting nodes from the cloud.

Use the PCLM CLI to view the status of the nodes while cloud bursting is in progress. Executeunbursting command to remove nodes.

1. Execute this command to get the status of the cloud bursting activity.pclm --api-endpoint=http://HOSTNAME:9980/pclm/ --api-key API_KEY bootstrapper scenario status

Where:

• HOSTNAME is the IP address or hostname of the machine where Cloud is installed.

• API_KEY is API Key generated by Control while creating the cloud bursting scenario.

The status command will display the machine id, machine private ip and hostname for the nodes.

2. Execute this command for unbursting.pclm --api-endpoint=http://HOSTNAME:9980/pclm/ --api-key API_KEY bootstrapper --wait scenario\ unburst MACHINE_ID|MACHINE_PRIVATE_IP|HOSTNAME

Where:

• MACHINE_ID is the ID of the machine to be removed.

• MACHINE_PRIVATE_IP is the IP of the machine to be removed.

• HOSTNAME is the hostname of the machine to be removed.

You can use ID, Private IP or hostname of a machine to unburst it.



8.16 Configuring PBS Professional for CloudBursting

8.16.1 Configuring Your Network

VPNA VPN is required from the on-premise network to the cloud network for seamless communicationbetween the two.

Test the VPN setup: manually bring up the instance on the cloud and check connectivity between thelocal and cloud instances.

Set Up DNSAny node burst in the cloud must register with the DNS server to identify itself in the complex. Usea DNS server installed on the PBS Server host. Any node that is created needs to be identified by itshostname. When a node is burst, the hook script takes care of registering any cloud nodes with theDNS server on the PBS server host. Instructions for installing and configuring a BIND DNS server areprovided in the following section Installing and Configuring a Bind DNS Server.

Important: The DNS server can become a bottleneck if you are running more than 256cloud nodes. Consider adding a replica DNS server on a utility node in the cloud running apbs_comm. Contact Altair for solutions.

8.16.2 Configuring Cloud Bursting Custom Resources

PBS Professional Custom Resources for Cloud BurstingCustom resources that must be added to PBS Professional to implement cloud bursting.

Table 12: Cloud Bursting Custom Resources

Custom Resource Name Type Description

cloud_queue Queue-level Boolean When True, the queue is acloud queue.

cloud_instance_type Queue-level string Instance type for this queue.

cloud_node_instance_type Host-level string. Instance type of this cloudnode.



Custom Resource Name Type Description

cloud_min_instances Server-level integer. Minimum number of instancesto be present in the cloud atany time.

cloud_max_instances Server-level integer. Maximum number of instancesthat can be made available inthe cloud.

cloud_min_uptime Server-level duration. Isexpressed as an integer whoseunits are in seconds.

Minimum time for which acloud burst node can be up.

cloud_provisioned_time Host-level integer. Time at which the cloud node iscreated.

lic_signature Host-level string. Contains node hour informationfor licensing

cloud_node_image Host-level string. OS image that is to be usedwhen a cloud node is burst.

cloud_network Host-level string Used to group a set of nodeswith the same network name.Used to request InfiniBandenabled nodes.

node_location Host-level string Used to differentiate localnodes from cloud nodes. It isuseful when the site wants thescheduler to try to place jobson the on-premise nodes firstbefore trying to run them oncloud nodes, while restrictingjobs from running on both.Used with placement sets.

cloud_max_jobs_check_per_queue Queue-level integer. The number of queued jobsin the cloud queue that arechecked to determine thenumber of nodes that must beburst based on the instancetype requested. Default is 64.



Further Explanation for cloud_max_jobs_check_per_queueThe following example further explains the use of cloud_max_jobs_check_per_queue.

1. The cloud hook first checks for queued jobs in the cloud queue.

2. It then checks the number of resources it requires to run those jobs.

3. It adds up all the resources required for the queued jobs to run.

4. It calculates the number of nodes required to run the job.

Example:

Suppose there is a cloud queue and the instance type specified for the queue is Standard_DS2_V2. Theconfiguration for Standard_DS1_v2 is{“ncpus”:2, “mem”:”8gb”}.

Jobs that are in a queued state in the cloud queue are J1, J2, and J3.

• J1 needs 3 CPUs to run.



The number of CPUs required to run these jobs is 3+2+5 = 10 CPUs.

The number of nodes needed to run all 3 jobs = (10 CPUs/2[Standard_DS2_V2 number of CPUs]) = 5nodes are needed to burst nodes of type Standard_DS2_V2.

cloud_max_jobs_check_per_queue is set to the default value of 64. If 100 jobs are queued in the cloudqueue (J1.….J100), then only the first 64 jobs (J1.…..J64) are used to calculate how many nodes needto be burst. You may change the value of cloud_max_jobs_check_per_queue to tune cloud bursting.

Creating Cloud Bursting Custom ResourcesAdd custom resources required to implement cloud bursting.

Creating custom resources must be done as root or as a user with sudo permissions using the sudocommand.


2. Copy and execute the following commands to create the custom resources required for cloudbursting:qmgr -c "create resource cloud_queue type=boolean" qmgr -c "create resource cloud_instance_type type=string"qmgr -c "create resource cloud_node_instance_type type=string,flag=h"qmgr -c "create resource cloud_min_instances type=long"qmgr -c "create resource cloud_max_instances type=long"qmgr -c "create resource cloud_min_uptime type=long"qmgr -c "create resource cloud_provisioned_time type=long,flag=h"qmgr -c "create resource lic_signature type=string,flag=h"qmgr -c "create resource cloud_node_image type=string,flag=h"qmgr -c "create resource cloud_network type=string,flag=h"qmgr -c "create resource cloud_max_jobs_check_per_queue type=long"

3. Navigate to PBS_HOME/sched_priv.

4. Edit the sched_config file.



5. Add cloud_node_image, cloud_node_instance_type, and cloud_network to the resourcesline:resources: "ncpus, mem, arch, host, vnode, netwins, aoe, cloud_node_image, cloud_node_instance_type, cloud_network"

8.16.3 Setting Cloud Limits at ServerSet server limits for the minimum number of instances to be present in the cloud at any time,maximum number of instances that can be made available in the cloud, and minimum amount of timefor which a cloud burst node can be up.

Setting server limits must be done as root or as a user with sudo permissions using the sudo command.


2. Copy and execute the following commands to set server limits:qmgr -c "set server resources_available.cloud_min_instances = 0"qmgr -c "set server resources_available.cloud_max_instances = MAX_INSTANCES"qmgr -c "set server resources_available.cloud_min_uptime = MIN_UPTIME"

Where MAX_INSTANCES is the maximum number of instances that can be made availablein the cloud and MIN_UPTIME is the minimum time for which a cloud burst node can be up.Recommendation is to set the minimum uptime to at least 15 minutes, test cloud bursting andadjust as necessary.

8.16.4 Creating and Configuring the Cloud Queue

Cloud Provider Instance TypesInformation about cloud provider instance types.

The instance type (also called shapes or machine types) determines the hardware of the host computerused for your cloud bursting nodes. Each instance type offers different compute, memory, and storagecapabilities. You will need to know what type of instance type you wanted burst in the cloud beforeyou can configure the cloud queue. For information about instance types per cloud provider see thefollowing table:

Table 13: Cloud Provider Instance Types

Provider URL Example Instance Types

Azure Sizes for virtual machines inAzureand High performancecompute VM sizes

Standard_DS1_v2, Standard_D2s_v3, Standard_NC6

Standard_H16r, Standard_H16mr (InfiniBand)

AWS Amazon EC2 Instance Types t2.medium, r4.large, p3.2xlarge


https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes

https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes

https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-hpc

https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-hpc

https://aws.amazon.com/ec2/instance-types/


Provider URL Example Instance Types

GCP Machine Types n1-standard-8, n1-highmem-2, n1-highcpu-64

Oracle VM Shapes and GPU Instances VM.Standard1.1, VM.DenseIO1.16, VM.GPU3.1

Configuring the Queue for Cloud BurstingAdd and configure the queue necessary for cloud bursting.

Creating and configuring a queue must be done as root or as a user with sudo permissions using thesudo command.


2. Copy and execute the following commands to create and configure the cloud queue:qmgr -c "create queue cloudq queue_type=execution,enabled=true,started=true"qmgr -c "set queue cloudq resources_available.cloud_queue = True"qmgr -c "set queue cloudq resources_available.cloud_instance_type = INSTANCE_TYPE"

Where INSTANCE_TYPE is the name of cloud provider instance type. This is the type of machinethat will be burst. It must match one of the instance types defined in the cloud bursting hookconfiguration file.

Instance type string has to match exactly with instance type in hook config file

3. Optionally, set the maximum number of jobs (MAX_JOBS) that can be placed in the cloud queue:qmgr -c "set queue cloudq cloud_max_jobs_check_per_queue = MAX_JOBS"

Where MAX_JOBS is the number of queued jobs in the cloud queue that are checked to determinethe number of nodes that must be burst based on the instance type requested. Default is 64. Canbe set to zero so that there is no limit.

8.16.5 Configuring the Cloud Bursting Hook

Cloud Bursting Periodic HookA python hook that performs all the cloud bursting work.

The cloud bursting periodic hook is automatically installed with PBS Professional 14.2.6 and 18.2.Export the cloud bursting hook configuration file and make any changes required for your site. Aftermaking the necessary changes, re-import the hook with the updated configuration values. The cloudbursting configuration file contains the following parameters:

default_userArtifact from a previous release of cloud bursting. No longer used, but is required to be defined inthe configuration file. Recommendation is to leave it set to "pbsuser".


https://cloud.google.com/compute/docs/machine-types

https://docs.cloud.oracle.com/iaas/Content/Compute/References/computeshapes.htm?Highlight=shapes#vmshapes

https://cloud.oracle.com/compute/gpu/features


resourcesResources to be considered for calculating number of nodes to burst. Resource name added inresources line should also be added to instance_types. Resources added should be of numerictype. Currently, resources that are supported are: "ncpus", "mem", "ngpus".

max_nodes_per_burst

Maximum number of nodes allowed to burst in a single hook cycle.

Maximum number of cloud node licenses to renew per hook cycle.

cloud_node_wait_timeoutMaximum time to wait for freshly burst node to become usable. Default is 180 seconds. You canset this to a higher value, but a lower value defaults to 180 seconds.

pclm_cmdPCLM command to be used by the hook script for bursting, unbursting, and status.

api_keyScenario token generated by Control at the time of a scenario creation.

pclm_cmd_licPCLM command to be used by the hook script in order to get the license signature.

Register_to_DNSWhen True, switch to use the DNS server on the PBS server host. When False, use customer'sDNS server. Remaining values are used when this value is set to True.

domainDNS domain name of complex.

zonefileForward zone filename, along with file path

reverse_zonefileReverse zone filename, along with file path

reverse_zone_octetsOctet value for reverse mapping

reload_cmdCommand that reloads the DNS service configuration and zone files. Default example is for aBIND DNS server.

cloud_min_uptime

Minimum time for which a cloud burst node will be up. This value is overridden by the resourcevalue set on the server.

cloud_min_instances

Minimum number of instances to be present in the cloud at any time. Does not apply duringstartup; cloud nodes are not immediately burst on startup. This is the minimum number that aremaintained after they are initially burst on demand. This value is overridden by the same valueset on the server.



cloud_max_instances

Maximum number of instances that can be made available in the cloud. This value is overriddenby the resource value set on the server.

max_vms_for_infiniband_scalesetDefines the number of nodes on an single InfiniBand switch. Supported by Azure only. This limit isset by Azure and the current default is 100. Customers may negotiate with Azure to increase thislimit. In this case, the value of max_vms_for_infiniband_scaleset should be updated.

cloud_default_imageOS image that is to be used by default when a cloud node is burst. This is overridden when theOS image is provided at job submission via the qsub command.

instance_typesInstance type specifications: type of instance on cloud, for example Standard_DS3_v2. Instancetypes are defined by the cloud provider. The name of the instance type and the CPU and memorysignature must exactly match the cloud provider's.

Default Cloud Bursting Hook Configuration File{ "default_user":"pbsuser", "resources":["ncpus", "mem", "ngpus"], "max_nodes_per_burst":50, "cloud_node_wait_timeout":180, "pclm": { "pclm_cmd": "</path/of/pclm> --raw --api-endpoint=<PCLM_SERVER:PORT> --api-key %s bootstrapper %s %s", "api_key": "<API KEY HERE>", "pclm_cmd_lic": "</path/of/pclm> --raw --api-endpoint=<PCLM_SERVER:PORT> --api-key %s pacioli license sign %s %s" }, "dns": { "Register_to_DNS": true, "domain": "<DOMAIN NAME HERE>", "zonefile": "</path/to/forward_zone_file>", "reverse_zonefile": "</path/to/reverse_zone_file>", "reverse_zone_octets": 3, "reload_cmd": "/sbin/rndc reload" }, "cloud":{ "cloud_min_uptime": 180, "cloud_min_instances": 0, "cloud_max_instances": 10, "max_vms_for_infiniband_scaleset" : 100, "cloud_default_image" : "<CloudDefaultImage>" }, "instance_types":{ "Standard_DS2_v2": {"ncpus": 2, "mem": "7gb" }, "Standard_DS3_v2": {"ncpus": 4, "mem": "14gb" }, "Standard_DS1_v2": {"ncpus": 1, "mem": "3.5gb"} }}



Configuring the Cloud Bursting HookConfigure the cloud bursting hook per your site's requirements and then re-import it.

Configuring the hook must be done as root or as a user with sudo permissions using the sudocommand.

Export the cloud bursting hook configuration file, update it's parameters per your site's requirements,and then re-import the hook.

1. Export the cloud bursting hook configuration to a file:qmgr -c "export pbshook PBS_cloud_bursting application/x-config default" > FILENAME

Where FILENAME is the name of the file where the hook configuration is written.

qmgr -c "export pbshook PBS_cloud_bursting application/x-config default" > config.json

2. Edit the file.

3. Set resources to a comma separated list of resources that are to be considered for calculating thenumber of nodes to burst.["ncpus", "mem", "ngpus"] OR ["ncpus", "mem"]

4. Set the value of max_nodes_per_burst to the maximum number of nodes allowed to burst in asingle PCLM call.

5. Set the value of cloud_node_wait_timeout to the maximum time to wait for freshly burst nodeto become usable.

Default is 180 seconds. You can set this to a higher value, but a lower value defaults to 180seconds.

6. For pclm_cmd:

a) Set the value of </path/of/pclm> to the path where the PCLM CLI is installed.

By default, the PCLM CLI is installed at PBS_EXEC/bin/pclm for PBS Professional 18.2.x.

b) Set the value of <PCLM_SERVER:PORT> to http://HOSTNAME:9980/pclm/.

Where HOSTNAME is the IP address or hostname of the machine where Cloud is installed.

7. Set the value of api_key to the token generated by Control during the creation of the burstingscenario.

8. For pclm_cmd_lic:

a) Set the value of </path/of/pclm> to the path where the PCLM CLI is installed.

By default, the PCLM CLI is installed at PBS_EXEC/bin/pclm for PBS Professional 18.2.x.

b) Set the value of <PCLM_SERVER:PORT> to http://HOSTNAME:9980/pclm/.

Where HOSTNAME is the IP address or hostname of the machine where Cloud is installed.

9. For the dns section:

a) Set the value of Register_to_DNS to true.

b) Set the value of domain to the DNS server domain name previously installed.

c) Set the value of zonefile to the path to the DNS zone file.

d) Set the value of reverse_zonefile to the path to the DNS reverse zone file.



e) Set the value of reload_cmd to the path of the command to reload the DNS serverconfiguration and zone files.

Default value is for a BIND DNS server.

10. For the cloud section:

a) Set the value of cloud_min_uptime to the minimum time for which a cloud burst node will beup.

This value is overridden by the resource value set on the server.

b) Set the value of cloud_min_instances to the minimum number of instances to be present inthe cloud at any time.

This value is overridden by the same value set on the server.

c) Set the value of cloud_max_instances to the maximum number of instances that can bemade available in the cloud.

This value is overridden by the resource value set on the server.

d) Set the value of max_vms_for_infiniband_scaleset to define the maximum number ofnodes allowed on a single InfiniBand switch.

This value should match the value of Maximum number of VMs inside a ScaleSet as specifiedin the Azure bursting scenario. If you are not using InfiniBand, do not change the value ofthis parameter.

Note: Supported by Azure only.

e) Set the value of cloud_default_image to OS image that should be used for bursting.

This OS image is used if the image is not requested as part of the job submission request.

11. For the instance_types section, add any instance types per your cloud provider that you wantburst.

Instance types are defined by the cloud provider. The name of the instance type and the CPU andmemory signature must exactly match the cloud provider's.

Cloud Provider Instance Type Example

Azure "Standard_DS2_v2": {"ncpus": 2, "mem": "7gb" },"Standard_DS3_v2": {"ncpus": 4, "mem": "14gb" },"Standard_DS1_v2": {"ncpus": 1, "mem": "3.5gb"},"Standard_NC6": {"ncpus": 6, "mem": "56gb", "ngpus": 1}

AWS "t3.large": {"ncpus": 2, "mem": "8gb" }, "t2.xlarge":{"ncpus": 4, "mem": "16gb" }, "p3.2xlarge": {"ncpus": 8, "mem": "61gb", "ngpus": 1}

GCP "n1-standard-2": {"ncpus": 2, "mem": "7.5gb" },"n1-highmem-8": {"ncpus": 8, "mem": "152gb" },"n1-highcpu-16": {"ncpus": 16, "mem": "14.4gb", "ngpus": 1}

Oracle Cloud "VM.Standard1.1": {"ncpus": 1, "mem": "7gb" },"VM.DenseIO1.8": {"ncpus": 8, "mem": "120gb" },



Cloud Provider Instance Type Example

"VM.GPU3.1": {"ncpus": 6, "mem": "90gb", "ngpus": 1}

12. Re-import the hook with the updated configuration values:qmgr -c "import pbshook PBS_cloud_bursting application/x-config default FILENAME"

Where FILENAME is the name of the file containing the hook configuration.qmgr -c "import pbshook PBS_cloud_bursting application/x-config default config.json"

13. Optionally, set the cloud bursting hook frequency.

The default frequency for the cloud bursting hook to execute is every 2 minutes.

qmgr -c "set pbshook PBS_cloud_bursting freq=FREQUENCY"

Where FREQUENCY is the frequency for the hook to execute in minutes.

14. Set the cloud bursting hook alarm:

You must set the hook alarm; otherwise, the hook will time out after the default value of 600seconds. Tune the alarm value to something greater than the default. Consider the followingfactors:

• Time required to burst nodes on the cloud

• Time required to unburst nodes on the cloud

• Number of cloud queues

Tip: Recommend that this value should be set to at least 20 minutes (1200 seconds).

qmgr -c "set pbshook PBS_cloud_bursting alarm=ALARM_TIME"

15. Enable the cloud bursting hook.

The cloud bursting server periodic hook is disabled by default.qmgr -c "set pbshook PBS_cloud_bursting enabled=True"

8.16.6 Testing Cloud Bursting by Submitting a JobTest cloud bursting by submitting jobs to cloud queues.

Tip: To troubleshoot issues with the cloud bursting hook, specify that all log events shouldbe captured in the PBS Server logs. Make sure to reset this value after testing is complete.qmgr -c "set server log_events = 2047"

1. Submit jobs to the cloud queue.qsub -l select=1:ncpus=4 -q CLOUD_Q JOB_SCRIPTqsub -l select=1:ncpus=4 -q CLOUD_Q JOB_SCRIPT

Where:





2. Check the status of the jobs:qstat -s

3. Tail the PBS Professional server logs.tail -f PBS_HOME/server_logs/CURRENT_SERVER_LOG_FILE

Where CURRENT_SERVER_LOG_FILE is the most current PBS Server log file.

4. Check the current log file to verify that the cloud bursting hook is started.

Search for the name of the cloud bursting hook.Server@master;Hook;HOOK_NAME;started

Where HOOK_NAME is the name of the cloud bursting hook.

5. Check the status of the nodes to verify that PBS Professional burst to cloud nodes:pbsnodes -av

6. Check the status of the jobs.

Once the node is added, jobs should start running.qstat -s

8.16.7 Managing Cloud and On-premise JobsSolutions to handle various job distribution needs.

Your site may want to run certain types of jobs on-premises or in the cloud. PBS Professional providesvarious methods to collect and distribute jobs. For more information see Routing Jobs in the PBSProfessional Administrator’s Guide. Here are some solutions to handle various job distribution needs.

Associate Node with QueuesIn general, start by associating nodes with queues, so that a job in a specific queue is sent to thedesired nodes. Associate each on-premise vnode with the appropriate local queue. Set the queueattribute on the vnode to the name of the local queue:

qmgr -c "active <list of on-premise nodes>"qmgr -c "set node queue=<name of local queue>"

For more information see Creating Queues and Associating Vnodes With One Queue in the PBSProfessional Administrator’s Guide.

Use Hooks or Routing Queues to Send Jobs to the Appropriate QueueTo send jobs to the appropriate queue on submission, use hooks or routing queues. For moreinformation see the PBS Professional Hooks Guide and Routing Queues in the PBS ProfessionalAdministrator’s Guide.

Use Peering to Pull Jobs to Empty NodesTo allow jobs to be run where resources are available, use peering to pull jobs to empty nodes. Formore information see Peer Scheduling in the PBS Professional Administrator’s Guide.

Use Placement SetsCreate a custom resource and use it to indicate that a node is on-premises.



qmgr -c "create resource node_location type=string,flag=h"

To make sure that each job runs either on premises or in the cloud, but not both, you can useplacement sets. Set the resources_available.node_location resource on your local vnodes to onevalue, for example, "local", to indicate that they are on-premise:qmgr -c "active <local vnode list>"qmgr -c "set node resources_available.node_location=local"

Make sure that cloud vnodes have resources_available.node_location unset.

Enable placement sets:qmgr -c "set server node_group_enable=true"qmgr -c "set server node_group_key=node_location"

Allow creation of placement sets from unset resources:qmgr -c "set sched only_explicit_psets=false"

Optionally, prevent any single job from running on both local and cloud vnodes:qmgr -c "set sched do_not_span_psets=true"

At each local queue, set default resources so that jobs go to local vnodes:qmgr -c "set queue <queue name> resources_default.node_location=local"qmgr -c "set queue <queue name> resources_min.node_location=local"qmgr -c "set queue <queue name> resources_max.node_location=local"

Add the node_location resource to the resources line in the sched_config file so that it includes:"... , node_location"

Restart or HUP the scheduler:

kill -HUP <scheduler PID>

For more information see Placement Sets in the PBS Professional Administrator’s Guide.

Job Distribution Examples and SolutionsExamples for routing jobs to the cloud.

Burst when On-premise Nodes are Not AvailableYou want to run as many jobs as possible on on-premise nodes, and only burst to the cloud when nomore on-premise nodes are available.

Solution: Create two execution queues, one for local jobs and one for cloud jobs. Set up local peeringso that the local queue pulls from the cloud queue whenever local nodes are available.

Send Small Jobs to the CloudYou have big machines for on-premise nodes and want to reserve those big machines for big jobs. Youwant to send smaller jobs to the cloud.

Solution 1:

• Create three queues: a routing queue, a local queue, and a cloud queue.

• Use the routing queue to collect jobs on submission.

• Set resource gating on the local queue to filter out smaller jobs.



• Allow smaller jobs into the cloud queue.

Solution 2:

• Create two queues: a local queue, and a cloud queue.

• Use a queuejob hook to route jobs into the appropriate queue.

Send Specific Jobs Only to the CloudYou want to send specific jobs to the cloud because:

• an application needed by some jobs runs well in the cloud.

• a resource that is available in the cloud is not available locally.

• a department has exhausted its share of local resources, and wants to send its jobs to the cloud.

Solution :

• Create two queues: a local queue, and a cloud queue.

• Use a queuejob hook to route jobs into the appropriate queue.

Charge Departments for Resources UsedYou have multiple departments and each department should be charged for the resources it uses.

Solution :

• Create a cloud queue for each department.

• Set separate limits on each cloud queue.

8.16.8 Disabling Cloud BurstingDisable cloud bursting by disabling the PBS Professional periodic hook.

1. Login to the PBS Server as root.

2. Enter the following command and press ENTER.

qmgr -c "set hook cloud_burst enabled=False"


PBS ProfessionalConsiderations 9

9 PBS Professional Considerations


• 9.1 Upgrading or Reconfiguring PBS Professional (p. 226)

Control Administrator's Guide9 PBS Professional Considerations p.226

9.1 Upgrading or Reconfiguring PBS ProfessionalStop the Data Collectors before upgrading or reconfiguring PBS Professional.

Before upgrading PBS Professional, performing maintenance, or re-configuring execution hosts wherethe system is drained of running jobs and the PBS Professional Server and MoMs are shut down, it isrecommended to stop the Analytics Data Collectors.

1. Login to each machine where a Data Collector is installed.

2. Stop the Data Collector:systemctl stop pbsworks-dc


Advanced Configuration 10

10 Advanced Configuration


• 10.1 Advanced Configuration by Component (p. 228)

• 10.2 Advanced Configuration for the Web Application (p. 230)

• 10.3 Advanced Configuration for Monitoring (p. 237)

• 10.4 Advanced Configuration for Analytics (p. 239)

• 10.5 Advanced Configuration for Cloud (p. 242)

Control Administrator's Guide10 Advanced Configuration p.228

10.1 Advanced Configuration by ComponentOverview of advanced configuration that is available for each Control component and associatedconfiguration files.

Web Application

Table 14: Advanced Configuration and Associated Files for the Web Application

Configuration Category Configuration File

Single Sign-on Authentication PC_HOME/config/shared/auth-config/auth_details.json

PC_HOME/config/shared/auth-config/krb5.conf

PC_HOME/config/pcgui/tomcat/conf/web.xml

Multi-Factor Authentication PC_HOME/config/shared/auth-config/auth_details.json

SSH Server PC_HOME/config/ams/jaas-config/AA_jaas.config

Session Timeout PC_HOME/config/ams/AA_Service.properties

Gateway HTTPS Port PC_HOME/config/api_gateway/nginx/nginx.conf

License Server PC_HOME/config/license/environment.json

PC_HOME/pbs-control-simulator/etc/pbssim/environment

PC_EXEC/pbsa/scripts/server-start.sh

Analytics

Table 15: Advanced Configuration and Associated Files for Analytics

Configuration Parameter Configuration File

Location of Accounting Logs PC_HOME/config/dc/modules/pbsa/config/config.xml

Analytics Hostname PC_HOME/config/api_gateway/nginx/nginx.conf

PC_HOME/pbs-control-simulator/etc/pbssim/environment




PC_HOME/pbs-control-simulator/etc/pbssim/pbssim.conf

Monitor

Table 16: Advanced Configuration and Associated Files for Monitor


Concurrent Threads for Retrieving JobDetails

PC_HOME/config/monitor/services/monitor.conf

Job and Node Details Refresh Interval PC_HOME/config/monitor/services/monitor.conf

Cloud

Table 17: Advanced Configuration and Associated Files for Cloud


Cloud Hostname PC_HOME/config/api_gateway/nginx/nginx.conf

PC_HOME/config/api_gateway/nginx/conf.d/model.conf

PC_HOME/config/shared/token_handler_config/token_validator/RequestValidatorConfig.json

DNS Nameservers /etc/docker/daemon.json



10.2 Advanced Configuration for the WebApplication

10.2.1 Advanced Configuration for Authentication

Configuring Single Sign-On

Enabling Single Sign-OnEnable Single Sign-On (SSO) so that clients who are logged into a domain can access Control withoutproviding credentials a second time.

The browser that is used to access Control has to be configured for Kerberos and NT LAN Manager(NTLM). View How to configure supported browsers for Kerberos and NTLM for a list of supportedbrowsers and platforms, as well as steps for configuring the supported browser.

Additionally, every service that uses Kerberos authentication needs to have an SPN set for it so thatclients can identify the service on the network. For information on how to register a SPN see Registeringa SPN. If a SPN has already been registered, verify the SPN before enabling SSO - Verifying SPNRegistrations.


Note: By default SSO is disabled.

1. Login to the machine where the Web Interface is installed.

2. Edit the file PC_HOME/config/shared/auth-config/auth_details.json.

3. Enable SSO by changing the value of the sso key to true."sso": { "enable_sso": "true" },

4. Edit the file PC_HOME/config/shared/auth-config/krb5.conf.

5. Replace all instances of $SSO_DOMAIN$ with the domain name (in uppercase) against which theusers are authenticated.[libdefaults] default_realm=$SSO_DOMAIN$ default_tkt_enctypes = aes128-cts rc4-hmac des3-cbc-sha1 des-cbc-md5 des-cbc-crc default_tgs_enctypes = aes128-cts rc4-hmac des3-cbc-sha1 des-cbc-md5 des-cbc-crc permitted_enctypes = aes128-cts rc4-hmac des3-cbc-sha1 des-cbc-md5 des-cbc-crc

[realms] $SSO_DOMAIN$ = { kdc = $SSO_DOMAIN$ default_domain = $SSO_DOMAIN$}

[domain_realm] .$SSO_DOMAIN$ = $SSO_DOMAIN$


https://ping.force.com/Support/%20PingFederate/Integrations/How-to-configure-supported-browsers-for-Kerberos-NTLM


On Windows, find the domain name by right-clicking My Computer and selecting Properties.

Figure 72: Windows Domain Name

On Linux, find the domain name using the command dnsdomainsname.

6. Edit the file PC_HOME/config/pcgui/tomcat/conf/web.xml.

7. Perform a search for the following text: ALTAIR PBSControl SSO Configurations.

8. Uncomment the XML just below this text by removing  that surround the XML.

9. Update SPN_USER_NAME and SPN_USER_PASSWORD with the credentials used when registeringthe SPN. <init-param> <param-name>spnego.preauth.username</param-name> <param-value>{SPN_USER_NAME}</param-value> </init-param>

<init-param> <param-name>spnego.preauth.password</param-name> <param-value>{SPN_USER_PASSWORD}</param-value> </init-param>

10. Restart Control:



pbs-control-manager restart

Registering a SPNRegister a Service Principal Name (SPN) so the you can implement Kerberos authentication.

SPNs are unique identifiers for services running on servers. Every service that uses Kerberosauthentication needs to have an SPN set for it so that clients can identify the service on the network.

1. Click this link: http://spnego.sourceforge.net/spnego_tomcat.html.

2. Scroll down until you find the section titled "Registering an SPN".

3. Follow the instructions to register a SPN.

Verifying SPN RegistrationsCheck that a SPN is registered for a specific computer.

1. Click the Windows Start button.

2. Enter command in the search box.

3. Right-click Command Prompt.4. Click Run as administrator.5. Enter the command: setspn -T -F -Q http/<DNSNAME>.

<DNSNAME> is the DNS hostname used when registering the SPN.

Disabling Single Sign-OnDisable Single Sign-On so that users are forced to enter their credentials when logging into Control.




3. Disable SSO by changing the value of the sso key to false."sso": { "enable_sso": "false" },

4. Edit the file PC_HOME/config/pcgui/tomcat/conf/web.xml.

5. Perform a search for the following text: ALTAIR PBSControl SSO Configurations.

6. Comment the XML just below this text by surrounding the XML with .


Enabling Multi-Factor AuthenticationEnable DUO multi-factor authentication to strengthen access security by requiring two methods to verifyyour identity.

You must have already implemented DUO multi-factor authentication at your site.


Multi-factor authentication adds a second layer of security to your site. Verifying your identity using asecond factor (like your phone or other mobile device) prevents anyone but you from logging in, even ifthey know your password.



3. Enable multi-factor authentication by changing the value of the enable_multi_factor key to true. "multifactor": { "enable_multi_factor": "true",

4. Replace %INTEGRATION_KEY%, %SECRET_KEY%, and %APPLICATION_KEY% with theappropriate values obtained when you chose to protect a new application using the DUO AdminPanel."integration_key":"%INTEGRATION_KEY%","secret_key":"%SECRET_KEY%","application_key":"%APPLICATION_KEY%",

5. Replace %DUO_SCREEN_URL% with the URL that is hosting DUO authentication.

Note: Control sends requests to this URL and communication must be done overHTTPS - i.e. https://someURL. Additionally, the request that Control sends is a signedrequest, therefore you must be prepared to accept a signed request. For an example,see the HTML in step 5.

"duo_screen_URL":"%DUO_SCREEN_URL%"

6. Make sure your post_action in the DUO web page points to the following URL:https://PC_HOSTNAME/auth/validate/mf



where PC_HOSTNAME is the hostname of the machine where Control is installed.

Example of DUO HTML page:<!DOCTYPE html><html> <head> <meta charset="utf-8"> <title>Duo Security 2-factor auth Tut - Enable Two-Factor Auth</title> <link rel="stylesheet" href="css/main.css"> </head><body> <iframe id="duo_iframe" width="100%" height="500" frameborder="0"></iframe> <script src="js/Duo-Web-v2.js" type="text/javascript"></script>    <script> var urlParams = new URLSearchParams(window.location.search); var myVar = urlParams.get('sig_request'); console.log(urlParams.get('sig_request')); console.log(urlParams.toString()); Duo.init({ 'host': 'api-45d973a0.duosecurity.com', 'sig_request': myVar, 'post_action': 'https://PC_HOSTNAME/auth/validate/mf' }); </script> Welcome to DUO login</body></html>

Changing the SSH ServerPoint to a new SSH server.



2. Edit the file PC_HOME/config/ams/jaas-config/AA_jaas.config.

3. Change the value of host to the hostname of the new SSH server.LoginModule { com.altair.jaas.module.SSHUnixLoginModule Required host="localhost" port="22" cache_ttl="60" debug=false; com.altair.jaas.module.UsersAccessControl Required;};




10.2.2 Advanced Configuration for Licensing

Configuring the License ServerChange the location of the license server.


1. Login to the machine hosting the Web Interface.

2. Back up the file PC_HOME/config/license/environment.json.

3. Edit the file environment.json.

4. Update the value of ALTAIR_LICENSE_PATH to the new license server in the formatport@hostname."ALTAIR_LICENSE_PATH":"6200@dm-10",

5. Save the file.


7. Login to the machine hosting the Simulator.

8. Back up the file PC_HOME/pbs-control-simulator/etc/pbssim/environment.

9. Edit the file environment.

10. Update the value of ALTAIR_LICENSE_PATH to the new license server in the [email protected]_LICENSE_PATH=6200@localhost

11. Save the file.

12. Update the PBS Professional license server.qmgr -c 's s pbs_license_info=LIC_PATH'

Where LIC_PATH is the new license server in the format port@hostname.

13. Restart PBS Professional.service pbs restart



16. Back up the file PC_EXEC/pbsa/scripts/server-start.sh.

17. Edit the file server-start.sh.

18. Update the value of LICENSE_SERVER_PATH to the new license server in the [email protected]_OPTS="$JAVA_OPTS -DLICENSE_SERVER_PATH=6200@dm-10"

19. Save the file.

20. Restart Analytics.systemctl restart pbsworks-pbsa



10.2.3 Configuring the Session TimeoutChange when a session closes as a result of inactivity over the client connection.


The default session timeout value is 5 hours.


2. Edit the file PC_HOME/config/ams/AA_Service.properties.

3. Change the value of java.sessionTimeout to a new value in hours.java.sessionTimeout=5


10.2.4 Changing the Gateway HTTPS PortChange the port that the Gateway HTTPS service listen on.



2. Edit the file PC_HOME/config/api_gateway/nginx/nginx.conf.

3. Change the Gateway HTTPS default port to a new port by updating the default value of 9500 oflisten of server. server { listen 9500 ssl default_server;

server_name _; ... ... }




10.3 Advanced Configuration for Monitoring

10.3.1 Configuring Concurrent Threads for Retrieving JobDetailsChange the number of concurrent threads that are created to retrieve PBS Professional job details whensystem memory issues arise.

Important: Do not change the number of concurrent threads unless your site isexperiencing system memory issues.

Monitor retrieves job details from PBS Professional using concurrent threads. By default, the number ofthreads created is based on the number of system cores multiplied by four (an 8 core system will create8 * 4 = 32 concurrent threads). This multiplication factor has been well tested and provides optimalresults. If your site is experiencing system memory issues, then the number of concurrent threads canbe reduced to free up memory.

CAUTION: The value of concurrent threads should always be set to a positive (great thanzero) number. Setting the number of concurrent threads to a low value will negatively affectjob refresh time. Assigning a higher value when there is a small number of PBS Professionaljobs will not improve performance.


1. Login to the machine where the Monitor component is installed.

2. Edit the file PC_HOME/config/monitor/services/monitor.conf.

3. Update the value of QSTAT_THREADS.QSTAT_THREADS=0


10.3.2 Configure the Job and Node Details RefreshIntervalChange how often Control requests job and node details from PBS Professional.

Important: Do not change the refresh interval unless your site is experiencingperformance issues.

Monitor runs a background process which retrieves job and node detail information from PBSProfessional. By default, this information is updated every 300 seconds (5 minutes). This interval can bemodified when performance issues are being observed.



CAUTION: Setting the refresh interval to a low value may overwhelm PBS Professional withcontinuous requests.

Thef ollowing steps must be done as root or as a user with sudo permissions using the sudo command.

1. Login to the machine where the Monitor is installed.

2. Edit the file PC_HOME/config/monitor/services/monitor.conf.

3. Update the value of REFRESH_SCHEDULER_WAIT_TIME_IN_SECONDS.REFRESH_SCHEDULER_WAIT_TIME_IN_SECONDS=300




10.4 Advanced Configuration for Analytics

10.4.1 Changing the Location Where the Data CollectorGathers the Accounting LogsChange the location of the PBS Professional accounting logs without reinstalling the Data Collector.


The main responsibility of the Data Collector is to make the PBS Professional accounting logs and nodeinformation (pbsnodes -av) accessible to Analtyics. The Data Collector has two functions. It first copiesthe accounting logs and node information from your site’s PBS Professional accounting log location toan intermediate holding area. Secondly, it transports the information stored in the holding area to themachine where Analytics is installed.

You can change the location where the PBS Professional accounting logs are stored without having toreinstall the Data Collector.


2. Edit the file PC_HOME/config/dc/modules/pbsa/config/config.xml.

3. Change the value of the XML element <entry> with the attribute @key="accounting-log-dir" tothe new location of the PBS Professional accounting log directory.<entry key="accounting-log-dir">/var/spool/PBS/server_priv/accounting</entry>

4. Restart the Data Collector.systemctl restart pbsworks-dc

10.4.2 Changing the Analytics HostnameChange the hostname of the Analytics component when it is moved to a new host.




3. Change the Analytics component hostname by updating the existing value of server for upstreamAnalyticsHost in the format of hostname:port.upstream AnalyticsHost{ server dm-04:9000; }

4. Change the Analytics component hostname by updating the existing value of server for upstreamAnalyticsBAHost in the format of hostname:port.upstream AnalyticsBAHost{ server dm-04:9443; }

5. Save the file.

6. Skip to Step 9 if the Web Application and Simulator are installed on the same machine.




8. Login to the machine where the Simulator component is installed.

9. Edit the file PC_HOME/pbs-control-simulator/etc/pbssim/environment.

10. Change the value of PBS_SIM_ANALYTICS_HOST to the new hostname of the Analytics component.

11. Save the file.

12. Edit the file PC_HOME/pbs-control-simulator/etc/pbssim/pbssim.conf.

13. Change the value of analytics_host to the new hostname of the Analytics component.

14. Save the file.


10.4.3 Resetting the Analytics Accounting Log DataDiscard all parsed accounting log data and reset all Data Collectors.

Before resetting the Analytics accounting log data, all simulation and model data must be deleted - seeDeleting All Simulation and Model Data.

The following commands must be executed as root or as a user with sudo privileges using the sudocommand.


2. Stop Analytics by issuing the command:systemctl stop pbsworks-pbsa

3. Execute the following command:pbsa-data-reset

All parsed accounting log data is discarded and Data Collectors are reset. The Analytics databasepassword to is also reset to “pbsworks@123”.

4. Use the pbsa-data-password command to change it to a site specific password:pbsa-data-password OLD_PASSWORD NEW_PASSWORD

Deleting All Simulation and Model DataRemove simulation and model data before resetting the PBSA dataset.


When the PBS Professional accounting log data is discarded, all simulation and model data must bedeleted as well.

1. Login to the machine where the Simulator component has been installed.

2. Execute the following command to remove all simulation and model data from Control:pbs-sim-agent --operation delete_all --auth-token placeholder

Simulation and model data is removed when the following message is displayed:



PBS Sim data reset complete.



10.5 Advanced Configuration for Cloud

10.5.1 Accessing Cloud AnalyticsView graphs of Docker container monitoring through a Grafana interface.

Note: When accessing the Cloud Monitoring interface, if you see an unauthenticated page,you will need to login to through the URL http://hostname:9980/pclm/ with the credentialssupplied when starting and initializing Cloud. If you installed Cloud using the Controlinstaller, the default credentials are admin@altair/Altair@123. If you manually installedCloud, use the credentials provided in step 3 of Starting and Initializing Cloud.


2. Enter the URL: http://hostname:9980/pclm/#/monitoring

Where hostname represents the hostname of the machine where Cloud is installed.

The Cloud Monitoring interface is displayed.

10.5.2 Accessing the CloudFlow InterfaceAccess the CloudFlow interface to follow the application workflows.


2. Enter the URL: http://hostname:CloudFlowPort

Where hostname represents the hostname of the machine where Cloud is installed andCloudFlowPort is the Cloud Server port + 10000 (default port is 9980 so CloudFlow port would be19980).

10.5.3 Changing the Hostname of the Cloud ComponentChange the hostname of the Cloud component when it is moved to a new host.


1. Login to the machine where the Control Web Interface is installed.


3. Change the Cloud component hostname by updating the existing value of server for upstreamPBSCloud in the format of hostname:port. upstream PBSCloud{ server dm-01:9980; }

4. Save the file.

5. Edit the file PC_HOME/config/api_gateway/nginx/conf.d/model.conf.


http://hostname:9980/pclm/#/monitoring


6. Change the value of $CLOUD_IO_URL to the new hostname in the format hostname:port.set $CLOUD_IO_URL "http://dm-01:9980/pclm";

7. Save the file.

8. Edit the file PC_HOME/config/shared/token_handler_config/token_validator/RequestValidatorConfig.json.

9. Change the value of pbscloud_url to the new hostname in the format hostname:port."pbscloud_url":"http://dm-01:9980/pclm"

10. Save the file.



Troubleshooting 11

11 Troubleshooting


• 11.1 Logging (p. 245)

• 11.2 Tracing an Issue (p. 254)

• 11.3 Troubleshooting Port Conflicts (p. 256)

• 11.4 Troubleshooting Cloud Bursting DNS Errors (p. 257)

Control Administrator's Guide11 Troubleshooting p.245

11.1 LoggingGeneral logging information.

Logging behavior for the following components and services can be modified:

• Web Application

◦ Authentication

◦ Gateway

◦ Licensing

◦ Notifications

• Monitor

• Analytics

• Simulator and Configure

Default Logging LevelThe default logging level is set to informational providing messages that highlight the progress of theservice at a coarse-grained level.

11.1.1 Location of Log FilesUse log files to troubleshoot problems with Control.

Installer Log FilesObtain detailed logging information for the installation process by navigating to PC_INSTALLER.

A log file is written to this directory whenever Control is installed, therefore each installation log fileis timestamped. During the installation process, information is written to this file as each serviceis installed. Service installation is done asynchronously, so messages pertaining to all services areinterspersed throughout this log file. Installation messages are also written to a log file that is specificto each service. These log files are located at:

PC_HOME/logs/SERVICE>/SERVICE_node_config.log

Where SERVICE is the name of the service.

Installer Report FileEvery time Control is installed or uninstalled a file called pbswinstall_op_report is appended withparameter information used during the installation/uninstallation process. The file is placed in the /etc directory. The contents of this file may be valuable for diagnosing issues with the installation/uninstallation of Control.

Shared Resources Log FilesLog files for shared resources (Java libraries, Apache Web Server, Tomcat) can be found at: PC_HOME/logs/shared



Component and Service Log File Location

Component Service Log File Location

Web Interface PC_HOME/logs/pcgui/tomcat

Message Router PC_HOME/logs/pbsworks-messagerouter

Gateway PC_HOME/logs/api_gateway

Licensing PC_HOME/logs/license

Notification PC_HOME/logs/wsbridge

Web Application

Authentication PC_HOME/logs/ams

Analytics Server PC_HOME/logs/pbsa/logs

Data Collector PC_HOME/logs/dc/logs

Simulator PC_HOME/logs/pbsworks-admin

Monitor PC_HOME/logs/monitor/services

Configure PC_HOME/logs/pbsworks-admin

Cloud http://hostname:5601 where hostname is thehostname of the machine where the Cloud componentis installed.

11.1.2 Web Application Logging

Changing the Authentication Service Logging LevelChange the logging level to troubleshoot authentication.



2. Edit the file PC_HOME/config/ams/logging-config/log4j.properties.

3. Find the following line in the file:log4j.logger.com.altair=info, R

4. Change the value of log4j.logger.com.altair from "info" to the appropriate logging level.

• OFF - designated the highest possible ranking and is intended to turn off logging.



• FATAL - designates very severe error events that will presumably lead the application toabort.

• ERROR - designates error events that might still allow the application to continue running.

• WARN - designates potentially harmful situations.

• INFO - designates informational messages that highlight the progress of the application atcoarse-grained level.

• DEBUG - designates fine-grained informational events that are most useful to debug anapplication.

• TRACE - designates finer-grained informational events than the DEBUG.

• ALL - designates the lowest possible rank and is intended to turn on all logging.


Changing the Gateway Service Logging LevelChange the logging level to troubleshoot the Gateway.



2. Edit the file PC_HOME/config/api_gateway/logging-config/log4j.properties.

3. Find the following line in the file:log4j.logger.com.altair=info, R

4. Change the value of log4j.logger.com.altair from "info" to the appropriate logging level.










Changing the Licensing Service Logging LevelChange the logging level to troubleshoot licensing.





2. Edit the file PC_HOME/config/license/log4j.properties.

3. Find the following line in the file:log4j.logger.com.altair.pbsworks.license=INFO, licenseActivityLogAppender

4. Change the value of log4j.logger.com.altair.pbsworks.license from "INFO" to theappropriate logging level.









5. Find the following line in the file:log4j.logger.com.altair.pbsworks.license.api=INFO, licenseApiActivityLogAppender

6. Change the value of log4j.logger.com.altair.pbsworks.license.api from "INFO" to theappropriate logging level.










Changing the Notification Service Logging LevelChange the logging level to troubleshoot the Notification Service.





2. Edit the file PC_HOME/config/wsbridge/log4js.json.

3. Find the following lines in the file:"categories": { "default": { "appenders": [ "wsbridge" ], "level": "info" } }

4. Change the value of level from "info" to the appropriate logging level.










11.1.3 Configure and Simulator Logging

Changing the Configure and Simulator Logging LevelChange the logging level to troubleshoot the Configure and Simulator components.


1. Login to the machine where the Configure and/or Simulate Simulator is installed.

2. Edit the file PC_HOME/pbsworks-admin/etc/environment.

3. Add the following line to the file:PBSWADMIN_LOG_LEVEL=

4. Set the value of PBSWADMIN_LOG_LEVEL to the appropriate logging level.

• DEBUG - fine-grained informational events that are most useful to debug an application.

• INFO - informational messages highlighting the progress of the application at a coarse grainedlevel.

• WARNING - potentially harmful situations such as failures and exceptions.




11.1.4 Monitor Logging

Monitor LoggingMonitor logging levels are set for specific Java packages. By default, the logging level is set to “info” forall packages and provides a nominal amount of logging.

If you want increase the details of logging for a package, then you need to change the logging level forthe corresponding package.

Below is a list of the Monitor Java packages and the types of classes the package contains.

Package Type of classes

com.altair.monitor.nextgen Base package of Monitor

com.altair.monitor.nextgen.cache Cache refreshers classes

com.altair.monitor.nextgen.command Action, Jobs and Nodes commands

com.altair.monitor.nextgen.core

com.altair.monitor.nextgen.data Data providers base class

com.altair.monitor.nextgen.data.cache App to EHCache cache managers classes

com.altair.monitor.nextgen.data.services Job and node summary parsers classes

com.altair.monitor.nextgen.model

com.altair.monitor.nextgen.notifications Notification and refresh all classes

com.altair.monitor.nextgen.queue Action response update queues classes

com.altair.monitor.nextgen.services Service classes for endpoints

com.altair.monitor.nextgen.session Session handling classes

com.altair.monitor.nextgen.utils General utility clases

com.altair.monitor.nextgen.ws Service endpoint classes

Changing the Monitor Logging LevelChange the logging level to troubleshoot the Monitor component.




1. Login to the machine where Monitor is installed.

2. Edit the file PC_HOME/config/monitor/services/loggerConfig.xml.

3. Scroll down until you find the Java package for which the logging level is going to be changed.In this example, the logging level for the base package of Monitor is being changed.<category name="com.altair.monitor.nextgen"> <priority value="info" />  </category>

4. Change the logging level for the Java package to:

Log Level Action

debug 1. Comment out <<priorityvalue="info" />> by prefixing the linewith <!– and suffixing the line with --> .

2. Remove the  surrounding<priority value="debug">

trace 1. Comment out <<priorityvalue="info" />> by prefixing the linewith <!– and suffixing the line with --> .

2. Remove the  surrounding<priority value="trace">

Tip: The logging level can be changed to any valid log4j logging level in addition to"debug" or "trace".









This example changes the logging level from INFO to DEBUG.<category name="com.altair.monitor.nextgen">  <priority value="debug" />



</category>


11.1.5 Analytics Logging

Changing the Analytics Log File SizeDefine the Analytics maximum log file size.


By default, Analytics generates a new log file once the size of the current log file exceeds 100MB.

1. Login to the machine where the Analytics component is installed.

2. Edit the file PC_HOME/config/pbsa/config/log4j.properties.

3. Locate the following line in the file:log4j.appender.frameworkRollingFile.MaxFileSize=100MB

4. Set the value of log4j.appender.frameworkRollingFile.MaxFileSize to the appropriate log filesize in megabytes.

5. Restart Analytics.systemctl restart pbsworks-pbsa

Changing the Analytics Logging LevelChange the logging level to troubleshoot Analytics.




3. Locate the following line in the file:log4j.rootLogger=INFO, frameworkRollingFile

4. Change the value of log4j.rootLogger from "INFO" to the appropriate logging level:











5. Restart the Analytics.systemctl restart pbsworks-pbsa

Changing the Maximum Number of Analytics Log Backup FilesDefine the maximum number of Analytics log backup files to keep.


By default, Analytics keeps a maximum of 1000 files as backup.



3. Locate the following line in the file:log4j.appender.frameworkRollingFile.MaxBackupIndex=1000

4. Set log4j.appender.frameworkRollingFile.MaxBackupIndex to the appropriate number ofbackup files to save.

5. Restart the Analytics.systemctl restart pbsworks-pbsa

11.1.6 Cloud Logging

Accessing Cloud LogsMonitor Cloud by viewing its logs through a Kibana interface.


2. Enter the URL: http://hostname:5601

Where hostname represents the hostname of the machine where Cloud is installed.



11.2 Tracing an Issue

11.2.1 Diagnosis ScriptA script that captures Control log and system data to be used for diagnostics.

Namepbswinstall_agent.py

Syntaxpbswinstall_agent.py diagnosis

DescriptionCapture log data for all Control components and services, as well as system data, running processesand installer log files. This information is useful when troubleshooting issues with Control. By default thelast 24 hours worth of log files are collected.

This command must be executed as root or as a user with sudo privileges using the sudo command onthe machine where Control is installed.

Parameter FileA parameter file PC_INSTALLER/pbswinstall.conf can be updated to configure the number of days oflog files to collect. The file contains the two parameters:[diagnosis]journalctl_start = todayjournalctl_end = tomorrow

To collect logs for additional days, these parameters can be set to a date in the format YYYY-MM-DD.

Table 18: Parameter Examples and the Number of Days of Log Files Collected

journalctl_start journalctl_end Days Log Files are Collected

today tomorrow Last 24 hours of log files are collected.

today today Only today's logs files are collected.

2018-11-15 2018-11-05 Log files from Nov 5th to Nov 15th are collected.

Example

PC_INSTALLER/pbswinstall_agent.py diagnosis



11.2.2 Using the Diagnosis ScriptUse the Diagnsosis script to gather log files and system data to help troubleshoot issues with Control.

1. Navigate to the PC_INSTALLER directory.

2. Enter the following command:source source_install_env && ./framework_init

3. Run the script:./pbswinstall_agent.py diagnosis

A tarball file is created in the current working directory calleddiagnosis_report_DATE_TIME.tar.gz where DATE and TIME is the file creation timestamp.



11.3 Troubleshooting Port ConflictsResolve port conflicts encountered while installing Control.

Default ports must be free to successfully install Control. When a port conflict is encountered duringinstallation, a message is logged and also displayed to the user who is installing the product. Theerror message that is displayed is Cannot start <servicename> service, port <port number> iscurrently in use.

Currently, there is no way to change the port used by all of the Control services, so the process usingthe port must be stopped.

1. Determine which process is using the port using an appropriate Linux command such as netstator lsof.

2. Stop the process that is using the port.

3. Uninstall Control.

4. Re-install Control.

See AlsoControl Components, Services and PortsUninstalling Control



11.4 Troubleshooting Cloud Bursting DNS ErrorsTry to force Docker to use your site's DNS server when DNS errors are encountered during cloudbursting.

DNS errors encountered while cloud bursting are displayed in the Cloud Notification center (accessed byclicking located in the upper right-hand of the Cloud tab). You will see errors like:

Failed to establish a new connection: [Errno -2] Name or service not known

Try the following to resolve the DNS issue:

1. Login to the machine hosting Cloud.

2. Create the file /etc/docker/daemon.json.

3. Edit the file.

4. Add the following lines to the file:{ "dns": ["YOUR_NAMESERVER1", "YOUR_NAMESERVER2"]}

5. Save the file.

6. Restart the Docker daemon:systemctl restart docker


Installing Control WithoutInternet Connectivity A-1

A-1 Installing Control Without Internet Connectivity

Install Control on a machine that does not have access to the internet. This requires a manualinstallation of Cloud.


• A-1.1 Installing Docker (p. 259)

• A-1.2 Installing MongoDB (p. 260)

• A-1.3 Cloud Prerequisite Script (p. 261)

• A-1.4 Installing Control (p. 262)

• A-1.5 Installing Cloud Manually (p. 264)

• A-1.6 Installing the Analytics Data Collector (p. 274)

Control Administrator's GuideA-1 Installing Control Without Internet Connectivity p.259

A-1.1 Installing DockerInstall Docker from a package or binaries.


1. Login to a machine with internet connectivity.

2. Download the Docker package or binaries using these instructions:

• For RHEL/CentOS, use these instructions: Install Docker CE from binaries.

• For SLES/SUSE, use these instructions: Install Docker EE from a package.


4. Copy the Docker package or binaries from the machine with internet connectivity using acommand such as scp.

5. Choose one of the following options to install and start Docker:

• For RHEL/CentOS, use these instructions: Install Docker CE from binaries.

• For SLES/SUSE, use these instructions: Install Docker EE from a package.






Recommended Docker Logging ConfigurationSet the maximum size of the Docker log file and the maximum log files to be kept.

To avoid consuming a lot of disk space and to rotate Docker logs, configure your Docker logging systemfollowing this documentation: https://docs.docker.com/config/containers/logging/configure/.

It is recommended to use the following configuration where the maximum size of the log before it isrolled is 10MB and the maximum number of files that are kept is 10.{ "log-driver": "json-file", "log-opts": { "max-size": "10m", "max-file": "10" }}


https://docs.docker.com/install/linux/docker-ce/binaries/

https://docs.docker.com/install/linux/docker-ee/suse/#install-from-a-package

https://docs.docker.com/install/linux/docker-ce/binaries/

https://docs.docker.com/install/linux/docker-ee/suse/#install-from-a-package

https://docs.docker.com/config/containers/logging/configure/


A-1.2 Installing MongoDBInstall MongoDB for use by the Web Interface, Analytics, Configure, and Simulator component for datastorage.

Attention: If the Monitoring component is being installed stand-alone, then MongoDB doesnot need to be installed. Otherwise, if you are installing any of the other components (WebInterface, Analytics, Configure, Simulator), MongoDB must be installed.


Restriction: Only MongoDB 3.4 and 3.6 are supported.

MongoDB is used by the Web Interface, Analytics, Configure and Simulator components for datastorage. Due to legal reasons, its installation could not be incorporated into the 2018.4 installer,therefore it must be manually installed, wherever the Web Interface, Analytics, Configure, andSimulator components are installed. Once MongoDB is installed, no other configuration is necessary.

1. Login to a machine with internet connectivity.

2. Download the MongoDB tarball.


4. Copy the MongoDB tarball from the machine with internet connectivity using a command such asscp.


• Follow these instructions to install MongoDB 3.4: Install MongoDB Community Edition FromTarball 3.4

• Follow these instructions to install MongoDB 3.6: Install MongoDB Community Edition FromTarball 3.6


https://docs.mongodb.com/v3.4/tutorial/install-mongodb-on-linux/





A-1.3 Cloud Prerequisite ScriptPrerequisite check for Cloud.

Run the below script to check the following prerequisites before installation:

• Presence of Docker and its version

• Space for the images (/var/lib/docker)

• Connectivity with the Cloud Docker repository

Note: The script will complain that it cannot connect to the Cloud Docker repository if youare installing Cloud manually because the machine where you are installing Control does nothave internet connectivity.

#!/bin/bashset -e

# check dockercommand -v docker >/dev/null 2>&1 || { echo >&2 "[ERROR] docker is required but is not installed or unavailable for this user."; exit 1; }command -v docker-init >/dev/null 2>&1 || { echo >&2 "[ERROR] docker-init is required but is not installed or unavailable for this user."; exit 1; }docker info >/dev/null 2>&1 || { echo >&2 "[ERROR] Cannot connect to the Docker daemon. Check that docker is running and that $USER belongs to docker group."; exit 1; }docker_server_version=$(docker info | grep -i "Server Version" | awk -F ": " '{print $2}' | cut -d '.' -f1)if [ $docker_server_version -lt 17 ]; then echo >&2 "[ERROR] Docker version is $docker_server_version. Only versions >= 17 are supported." exit 1fi

# check disk space on docker's root directorydocker_root=$(docker info | grep -i "Docker Root Dir" | awk -F ": " '{print $2}')if [ $(df -P $docker_root | awk 'NR==2 {print $4}') -lt 31457280 ]; then echo >&2 "[ERROR] Not enough space on $docker_root. Only $(df -Ph $docker_root | awk 'NR==2 {print $4}') available, 30GB is required." exit 1fi

# check registry accesscommand -v curl >/dev/null 2>&1 || { echo >&2 "[ERROR] curl is needed to check registry connectivity."; exit 1; }curl --silent --fail -X GET https://a882d237-5d07-48a7-837e-9bd8652d79ec:kq9ZKzrToLGm78xGBDv5lKx5hg5YOB8JDz3usbp+2eo=@pbscloudio.azurecr.io/v2/_catalog &> /dev/null \|| { echo >&2 "[ERROR] Cannot access to docker registry. Please check connection to https://pbscloudio.azurecr.io"; exit 1; }

# prerequisites metecho "[OK] All prerequisites met. You can proceed to Cloud installation."



A-1.4 Installing ControlInstall all Control components except Cloud using the Advanced installation type.

Review the installation prerequisites before proceeding with the installation.


Several components use MongoDB for data storage, therefore there will be multiple database instancesrunning once this installation is complete. The Monitor component and PBS Professional use Postgres fordata storage, therefore there will be two Postgres database instances running after this installation iscomplete.
















11. Choose to install all components except for Cloud by entering 1,2,3,4,5 and pressing ENTER.

12. Enter the hostname of the machine where Cloud will be installed and press ENTER.




14. Choose to install both the Analytics component and the Data Collector by entering 3 and pressingENTER.





18. Review the installation summary and then press ENTER to install the product.Message similar to those below are displayed as the installation progresses.Installing...-------------

This may take a while...Unzipping files into home, exec, and installer folders....

Progress is: 0% Web Interface Beginning PBSWorks Shared InstallationProgress is: 12% Web Interface Completed updating PBS Works Shared config files.........All milestones reached!!

19. Once the installation is complete, set the Service User's PAS password by issuing the command:su - SERVICE_USER -c "PAS_EXEC/pas-commands/bin/pas-passwd"

Where SERVICE_USER is the username of the Service User.



A-1.5 Installing Cloud ManuallyManually install Cloud when there is no access to the internet.

When the machine where Control is installed is not connected to the internet, a manual installation ofCloud is necessary. Docker must be installed on a machine with internet connectivity. Cloud images arethen downloaded and saved using an installer script. The images are then copied to the machine whereControl is hosted and loaded. Cloud containers are installed, Cloud is configured, started and initialized.

Installing DockerInstall Docker which is used to host Cloud.

Installing Docker for RHEL and CentOSInstall Docker for RHEL and CentOS platforms.


• You must have access to a machine that has internet connectivity and it must have the sameoperating system as the one where Control is going to be installed.

• 20GB free space on /var/lib/docker.

• For more information about installing Docker CE see Get Docker CE for CentOS.

1. Login to a machine that has internet connectivity.

2. Install required packages.yum install -y yum-utils

3. Setup the docker repository:yum-config-manager --add-repo \https://download.docker.com/linux/centos/docker-ce.repo

4. Install the packets and its dependencies:yum install docker-ce








https://docs.docker.com/install/linux/docker-ce/centos/


Installing Docker for SLES and SUSEInstall Docker for SLES and SUSE platforms.


• You must have access to a machine that has internet connectivity and it must have the sameoperating system as the one where Control is going to be installed.

• 20GB free space on /var/lib/docker.

• Docker CE is not supported on SLES, therefore Docker EE must be installed.

1. Login to a machine that has internet connectivity.

2. Install and start the latest version of Docker EE by following these instructions Get Docker EE forSLES.

3. Verify that the file /usr/bin/docker-init exists.

4. If the file does not exist:

a) Download the tarball for the latest version of Docker from the Docker repository:

https://download.docker.com/linux/static/stable/x86_64/

b) Extract the archive to /tmp:tar zxvf docker-18.06.1-ce.tgz -C /tmp

c) Execute the command:cp /tmp/docker/docker-init /usr/bin/







Download and Save the Cloud ImagesUse a Cloud download script to download and save Cloud images.

The installer script must be run on a machine having internet connectivity. This machine must haveDocker installed.

Use PBScloud_images_installer.sh to download and save the Cloud images.

1. Login to the machine with internet connectivity where Docker was installed.




https://download.docker.com/linux/static/stable/x86_64/docker-18.06.1-ce.tgz


2. Download the Cloud installer script.

3. Executed the following command:./PBScloud_images_installer.sh save

The Cloud images are downloaded and saved to a directory called PBScloud_images.

Load the Cloud ImagesUse a Cloud download script to load the Cloud images.


2. Copy the Cloud installer script from the machine with internet connectivity using a command suchas scp.

3. Copy the PBScloud_images directory from the machine with internet connectivity using acommand such as scp.

Make sure to copy the PBScloud_images directory to the same location asPBScloud_images_installer.sh (same level).

4. Executed the following command:./PBScloud_images_installer.sh load

The Cloud images are loaded.

Installing the Cloud ContainersUse Pocker to install the Cloud containers.

• The following steps must be done as root or as a user with sudo permissions using the sudocommand.

Pocker is a Docker wrapper, designed to setup, install and manage the Cloud containers.


2. Run the Pocker container used to setup and install Cloud:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v DIR/config:/config \-v DIR/kard:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash


3. Create your deployment environment definition:pkr kard create $CLOUD_VERSION -e pbscloud \--extra tag=$CLOUD_VERSION ui_port=9980 \ui_url_prefix='/pclm'




Linking Cloud with the PBS Works License ServerConfigure Cloud so that it can access the PBS Works license server.


Licensing for cloud bursting is done through the Control licensing service. This service is installedwhen the Web Interface component is installed. If Control has not been installed yet, you will need thehostname or IP address of the machine where it is going to be installed to configure Cloud for licensing.

1. Navigate to the directory DIR/kard/current.


2. Create/edit the file meta.yml and add the following lines to the file:license_server: host=http://PC_WEB_INTERFACE_HOSTNAME:9100 auth_url=http://CLOUD_IP_ADDRESS:9980/pclm

Replace PC_WEB_INTERFACE_HOSTNAME with the hostname/IP address of the machine wherethe Control Web Interface is installed. Replace CLOUD_IP_ADDRESS with the hostname/IP addressof the machine where Cloud is installed.

3. Save the file.



Replace DIR with the location where Cloud is installed.




Configuring a SMTP Server for Cloud Bursting Quota AlertsSet up a SMTP server so that when a cloud bursting resource alert limit is reached an email is sent tothe Cloud Administrator.






• If you installed Cloud using the Control installer and installed in the default location:export KARD_PATH=\/var/spool/pbsworks/$CLOUD_VERSION/pbscontrol/home/cloudio/kard

• If you installed Cloud using the Control installer and installed in a non-default location:export KARD_PATH=DIR/cloudio/kard

Replace DIR with the path to the Control home directory.



3. Navigate to the directory $KARD_PATH/current.

4. Create/edit the file meta.yml and add the following lines to the file:smtp: host: SMTP_HOST_URL port: SMTP_PORT login: SMTP_LOGIN password: SMTP_PASSWORD

The SMTP Server port number defaults to 25.

Note: The value of login and password can be left blank if a username and passwordare not necessary to use the SMTP server.

5. Save the file.

6. Launch a pkr container:docker run -ti --rm -e CLOUD_VERSION=$CLOUD_VERSION \-v /run/docker.sock:/run/docker.sock \-v $KARD_PATH/config:/config \-v $KARD_PATH/kard:/pkr/kard pbscloudio.azurecr.io/pkr:$CLOUD_VERSION bash




Starting and Initializing CloudStart and initialize Cloud using pkr.


1. Generate your environment:pkr kard make

2. Start Cloud:pkr start



3. Initialize Cloud by entering the command:pkr ext init full --domain_name <domain_name> \--username <admin_username> --password '<admin_password>' \--user_email '<admin_email>'

The values between the <> are used during the Control installation process to link Cloud andControl. You will need to replace these four values:

The domain_name does not have to be a true DNS domain, but it a group name for your project.It cannot include a period ( . ).

The admin_username and admin_password identify the credentials of the Cloud Administrator. Theadmin_username should not contain an asterisk ( * ). The admin_password must be at least 6characters in length and must contain at least one number.

Note: Email alerts are sent to the Cloud Administrator when a cloud bursting resourcealert limit is reached.

Below is an example:

pkr ext init full --domain_name chrysler --username tsmith --password 'password@123' --user_email '[email protected]'The output of this command contains keys to unlock the vault and main passwords. Copy theoutput of this command and store it in a location where you can find it easily.

4. Navigate to the directory: PC_HOME/data/api_gateway.

5. Edit the file: Cloud_io_Credentials.json.

This file contains a key-value pair for "username" and "password" "username":"admin@altair", "password":"TM9RHX9nNLX0riiPM3/T6w=="

6. Change the value of username to those values entered while initializing Cloud (step 3 above) inthe format admin_username@domain_name.Ex. tsmith@chrysler

7. Encrypt the password (admin_password) entered while initializing the Cloud component (step 3above):PC_EXEC/api_gateway/scripts/password_encryption.sh encrypt <admin_password>

Note: The output of this script is the encrypted password.

8. Copy the output from the previous step and paste it into the Cloud_io_Credentials.json file asthe value of password.

9. Save the file Cloud_io_Credentials.json.



Restoring Cloud DataRestore the Cloud data so that cloud accounts and bursting scenarios do not have to be recreated.

Note: For the 2018.4 version of Control, Pocker a tool for generating Docker images hasbeen replaced with the Open Source tool pkr.

The procedure for restoring the Cloud data is different depending upon the version of Control previouslyinstalled at your site.

Restoring Cloud Data for a Migration from 2018.2 to 2018.4Restore the Cloud data after migrating from Control 2018.2 to 2018.4.














8. Update the kard version:



cd /pkr/kard/$CLOUD_VERSIONsed -i \-e "s#^project_name:.*#project_name: '$(echo $CLOUD_VERSION | sed 's/\.//')'#" \-e "s#^src_path:.*#src_path: /pkr/kard/$CLOUD_VERSION/./src#" \-e "s#^tag:.*#tag: '$CLOUD_VERSION'#" \-e "s#^driver:.*#driver:\n name: compose#" \-e "\$alicense_server:\n \auth_url: $(grep âuth_url site_config/pacioli.conf | cut -d "=" -f2-)\n \host: $(grep ^host site_config/pacioli.conf | cut -d "=" -f2-)" \meta.yml


10. Start MongoDB:pkr start -s mongodb

11. Update the MongoDB database:docker exec -it mongodb mongo -u admin --authenticationDatabase admin \notification-center -p $(cat /pkr/kard/$CLOUD_VERSION/meta.yml | \grep db_mongo | awk '{print $2}') \--eval "db.createUser({ user: 'notification-center', \pwd: '$(cat /pkr/kard/$CLOUD_VERSION/meta.yml \| grep db_notification-center \| awk '{print $2}')', roles: \[ { role: 'readWrite', db: 'notification-center' } ] });"














20. Update Keystone:pkr ext init base

21. Update the database:docker exec -ti bootstrapper-api bashpython -m pclm_common.dbmigrate bootstrapper listpython -m pclm_common.dbmigrate bootstrapper migrate


Restoring the Cloud Data for a Migration from 2018.3 to 2018.4Restore the Cloud data after migrating from Control 2018.3 to 2018.4.














8. Update the kard version:



cd /pkr/kard/$CLOUD_VERSIONsed -i \-e "s#^project_name:.*#project_name: '$(echo $CLOUD_VERSION | sed 's/\.//')'#" \-e "s#^src_path:.*#src_path: /pkr/kard/$CLOUD_VERSION/./src#" \-e "s#^tag:.*#tag: '$CLOUD_VERSION'#" \-e "s#^driver:.*#driver:\n name: compose#" \-e "\$alicense_server:\n \auth_url: $(grep âuth_url site_config/pacioli.conf | cut -d "=" -f2-)\n \host: $(grep ^host site_config/pacioli.conf | cut -d "=" -f2-)" \meta.yml











18. Update the database:docker exec -ti bootstrapper-api bashpython -m pclm_common.dbmigrate bootstrapper migrate






A-1.6 Installing the Analytics Data CollectorInstall the Analytics Data Collector on each PBS Professional complex.



• The Data Collector must have access to the machine hosting Analytics.

If your site has clusters with the same fully qualified domain name (FQDN), then the Analytics data isaggregated for both clusters under a single FQDN. To allow Analytics data for clusters having the sameFQDN to be tracked separately, a unique name must be provided while installing each Data Collector.



1. Login to the PBS Professional Server.
















11. Choose the Analyze component by entering 5 and pressing ENTER.

12. Choose to install the Data Collector by entering 2 and pressing ENTER.

13. Enter the hostname of the machine where Control Web Application is installed and press ENTER.


15. Enter the hostname of the machine where Analytics was installed.

16. Enter a unique name (FQDN) for the Data Collector.





20. Review the installation summary and then press ENTER to install the product.Message similar to those below are displayed as the installation progresses.Installing... -------------This may take a while...Unzipping files into home, exec, and installer folders.... Progress is: 0% Analyze Beginning PBSWorks Shared InstallationProgress is: 12% Analyze Completed updating PBS Works Shared config files.........All milestones reached!!

21. Update the system path to include the path to the Analytics Commands:export PATH=$PATH:PC_EXEC/dc/scripts/pbsa-commands/bin

Perform post-installation configuration.


IndexAauthentication

multi-factorDUO 233enable 233

Service Principal NameKerberos 232register 232verify 232

Single Sign-Ondisable 232enable 230Kerberos 230NT LAN Manager 230NTLM 230Service Principal Name 230SPN 230

CCloud 10, 16, 36, 36, 36, 41, 48, 70, 259, 262, 264, 264, 265, 274CloudFlow

Cloud 242configure 41, 48, 262, 274conventions

PC_EXEC 8PC_HOME 8

Ddeployment 12, 16, 24Docker 36, 36, 36, 259, 264, 264, 265DUO 233

HHyperWorks 10, 16

Iinstallation

Cloud 16, 36, 36, 36, 259, 264, 264, 265configure 41, 48, 262, 274Control 16, 41, 48, 262, 274HyperWorks 16installers 16

276

MongoDB 16, 38, 260monitor 41, 48, 262, 274PBS Application Services 16PBS Professional 16PBSA 16simulator 41, 48, 262, 274simulator application definitions 41, 48, 262, 274simulator solver 41, 48, 262, 274single machine 16web application 41, 48, 262, 274

KKerberos 230, 232

Llogin access

default 24granting 60

logsCloud 242, 253

MMongoDB 16, 38, 41, 48, 260, 262, 274monitor 41, 48, 262, 274

NNT LAN Manager 230NTLM 230

PPAS , See PBS Application ServicesPBS Application Services 10, 16PBS Professional 10, 16PBSA 10, 16, 38, 260PC_EXEC 8, 41, 48, 262, 274PC_HOME 8, 41, 48, 262, 274pkr 36, 36, 36, 259, 264, 264, 265

SService Principal Name 230, 232, 232Service User 24, 38, 41, 48, 60, 260, 262, 274services

analytics 12Cloud 12

277

configure 12Control 72MongoDB 12monitor 12PBSA 72restart

Control 73, 79simulator 12start

Cloud 76Control 73, 77, 77, 77, 79

statusControl 74, 77, 80

stopCloud 76Control 73, 79

web application 12simulator 38, 41, 48, 260, 262, 274simulator application definitions 41, 48, 262, 274simulator solver 41, 48, 262, 274SPN 230, 232, 232SSO 230, 232, 232supported

browsers 10operating systems 10products

Cloud 10HyperWorks 10PBS Application Services 10PBS Professional 10PBSA 10

Uuninstall

Cloud 70Control 68, 69

Wweb application 41, 48, 262, 274

278