Disaster Recovery using Hitachi Universal Replicator on ...docshare01.docshare.tips/files/30620/306203529.pdfThis reference architecture guide depicts the necessary components and

November 2015

Reference Architecture Guide

Disaster Recovery using Hitachi Universal Replicator on Hitachi Unified Compute Platform for the SAP HANA Platform using Hitachi Compute Blade 2500, Hitachi Virtual Storage Platform G1000, and Hitachi NAS Platform 4060

Abhishek DhanukaBy

FeedbackHitachi Data Systems welcomes your feedback. Please share your thoughts by sending an email message to [email protected]. To assist the routing of this message, use the paper number in the subject and the title of this white paper in the text.

mailto:[email protected]?subject=Document%20AS-440-00%20

Contents

Solution Overview................................................................................................................ 3

Hardware Elements......................................................................................................................... 7Software Elements ........................................................................................................................ 11

Solution Design................................................................................................................... 12

Hitachi Compute Blade 2500 Chassis Configuration ................................................................... 12520X B2 Server Blade Architecture .............................................................................................. 14Fibre Channel SAN Architecture ................................................................................................... 14Storage Architecture ..................................................................................................................... 15Hitachi NAS Platform 4060 Architecture....................................................................................... 23Network File System Design for Shared Binaries ......................................................................... 25Management Server...................................................................................................................... 26Network Architecture .................................................................................................................... 27SAP Storage Connector API Fibre Channel Client ....................................................................... 33Journal Sizing................................................................................................................................ 33Inter-site Configuration ................................................................................................................. 35Disaster Recovery and Replication Components ......................................................................... 36

Engineering Validation ...................................................................................................... 43

Test Methodology ......................................................................................................................... 43Test Results .................................................................................................................................. 43

1Disaster Recovery using Hitachi Universal Replicator on Hitachi Unified Compute Platform for the SAP HANA Platform using Hitachi Compute Blade 2500, Hitachi Virtual Storage Platform G1000, and Hitachi NAS Platform 4060

Reference Architecture Guide

This reference architecture guide depicts the necessary components and overall layout of a disaster recovery solution using Hitachi Universal Replicator on Hitachi Unified Compute Platform for the SAP HANA platform (UCP) in a scale-out configuration. This guide documents how to deploy this configuration using the following:

Hitachi Compute Blade 2500 (CB 2500) chassis with 520X B2 server blades

Quanta Cloud Technology QuantaPlex T41S-2U server with one node

Hitachi Virtual Storage Platform G1000 (VSP G1000)

Hitachi NAS Platform 4060 (HNAS 4060)

System management unit

Hitachi Universal Replicator (HUR)

SAP HANA

Two operating systems:

SUSE Linux Enterprise Server for SAP Applications

Red Hat Enterprise Linux (Alternate)

The testing of this solution in the lab was only on a 2+1 configuration. However, this reference architecture supports the scale-out configurations as listed in Hitachi Unified Compute Platform for the SAP HANA Platform using 3 TB SAP HANA Nodes in a Scale-out 48 TB Configuration of 16 Active Nodes and 3 Standby Nodes with Hitachi Compute Blade 2500 Chassis, 520X B2 Server Blades, and Hitachi Virtual Storage Platform G1000 Reference Architecture Guide (AS-399-02). For information concerning your implementation, please contact your GSS representative for more details.

This document supports understanding the example architecture of disaster recovery using Hitachi Universal Replicator on Unified Compute Platform for SAP HANA in a scale-out configuration.

The scale out environment for Unified Compute Platform for SAP HANA is a preconfigured analytical appliance that provides real-time access to operational data for use in analytic models. Changes to this architecture require approval from the following at Hitachi Data Systems:

Sales

Solution Engineering and Technical Operations

Solution product management

1

https://www.hds.com/assets/pdf/hitachi-ucp-for-sap-hana-with-compute-blade-2500-and-vsp-g1000.pdf



2Failure of SAP HANA results in revenue loss. For protection from this loss, use two sites in the disaster recovery strategy. In addition to failover production, the second site can handle the quality assurance environment of the SAP HANA landscape.

The primary business problem this solution answers is disaster recovery for SAP HANA. This solution performs asynchronous replication of SAP HANA data volumes and log volumes on Hitachi Virtual Storage Platform G1000 to the secondary site. It also performs asynchronous replication of the SAP HANA binaries and other configuration files stored in the /hana/shared file system on Hitachi NAS Platform 4060 to the secondary site.

Data centers at each site must have almost identical hardware for this disaster recovery solution. Implementing Hitachi Universal Replicator on this environment permits the additional use of the secondary site as a quality assurance environment. This additional use requires adding additional disk drives on the secondary site storage to run the quality assurance system.

Hitachi Virtual Storage Platform G1000 technology permits maintaining sufficient performance for the SAP HANA production instance and site-to-site replication.

This technical paper assumes you have familiarity with the following:

Storage area network-based storage systems

Network attached storage (NAS) systems

General storage concepts

General storage replication skills and concepts

General network and virtual IP knowledge

General WAN knowledge

Advanced SAP HANA skills

Disaster Recovery scenarios

Common IT storage practices

Note — Testing of this configuration was in a lab environment. Many things affect production environments beyond prediction or duplication in a lab environment. Follow the recommended practice of conducting proof-of-concept testing for acceptable results in a non-production, isolated test environment that otherwise matches your production environment before your production implementation of this solution.

2

3Solution OverviewThe primary site A contains a production SAP HANA database instance. Implement this site as a scale-out configuration of Hitachi Unified Compute Platform for the SAP HANA platform.

The secondary site is an exact replica of primary site, with the exception of optional additional storage disks. Site B houses the following:

Failover production instance

(Optional) Non-production instances, such as the following that requires additional disks:

Quality assurance

Development

Test, In-memory database of SAP HANA.

The design of the SAP HANA in-memory database enables this solution to use the same set of Hitachi Compute Blade 2500 nodes for production and non-production instances.

In the test environment for this solution, the Hitachi Virtual Storage Platform G1000 at the secondary site housed two sets of disks for data, log, and /hana/shared LUNs for the following:

The replicated production instance

(Optional) The quality assurance system

In this solution, install command control interface on the management servers on the primary site and secondary site, which performs the data replication operations within Hitachi Virtual Storage Platform G1000 at each site using Hitachi Open Remote Copy Manager instances. Each instance at the primary site and secondary site management server has its own Hitachi Open Remote Copy Manager configuration file that lists the following for replication between two sites:

SAP HANA data volumes

SAP HANA log volumes

Hitachi NAS Platform LUNs

Hitachi Virtual Storage Platform G1000 works with Hitachi Universal Replicator software to enable SAP HANA disaster recovery. This solution uses a Fibre Channel over IP (FCIP) switch for the wide area network connections between each Virtual Storage Platform G1000 at the primary site and the secondary site. Take the options for long distance data replication in the landscape using the pre-existing enterprise network infrastructure and service provider into consideration for the SAP HANA disaster recovery setup.

With Hitachi Universal Replicator, updates on the SAP HANA nodes to the primary production volume on the primary site Virtual Storage Platform G1000 are copied to a local journal volume at primary site storage. The secondary site Virtual Storage Platform G1000 “pulls” data from the primary site journal volume across the inter-site wide area network connection to the secondary volume. The local system is free to perform its role as a transaction processing resource rather than as a replication engine.

For the asynchronous replication of Hitachi NAS Platform LUNs hosting the /hana/shared file system on Virtual Storage Platform G1000, perform a one-time initial configuration after completing the initial pair copy operations to configure the RAID mirror relationships between the NAS Platform LUNs at both sites. Make the NAS Platform clusters at both the sites aware of the Universal Replicator LUN relationship between them during this initial configuration by using the RAID mirroring command (sd-mirror-remotely).

3

4In case of an outage or any component failure in the primary site, an administrator initiates a manual failover to the secondary site using customized scripts. There are two different possibilities for enabling client connection recovery, using either virtual IP failover or DNS failover configuration. The actual implementation differs, based on the network and cluster management capabilities.

This solution supports the following four different disaster recovery options.

1. Add a Disaster Recovery Site

This option is for a primary site for production and a secondary site for disaster recovery.

2. Add a Quality Assurance or Development Instance

This option is for a primary site for production and a secondary site for disaster recovery site and a single quality assurance or development SAP HANA instance.

3. Disaster Recovery Connectivity Bundle

This option is for a Fibre Channel over IP (FCIP) switch between two sites.

4. Only Disaster Recovery Site

This option is for adding a secondary site for a disaster recovery solution to an existing SAP HANA landscape. (The primary site for production already exists.)

To perform the replication of Hitachi Unified Compute Platform for the SAP HANA platform, the reference solution uses the following:

Hitachi Compute Blade 2500 with 520X B2 Server Blades

Quanta Cloud Technology QuantaPlex T41S-2U Server

An ultra-dense design equipped with four independent nodes, it creates the flexibility to set up different workloads independently in one 2U shared infrastructure, providing optimal data center performance. This solution uses only one node out of the four nodes.

Hitachi Virtual Storage Platform G1000

This is a storage platform that scales for all data types that flexibly adapts for performance, capacity, and multi-vendor storage.

Hitachi NAS Platform 4060

This is a network-attached storage solution used for file sharing, file server consolidation, data protection, and business-critical NAS workloads.

System management unit software

Providing front-end server administration and monitoring tools for Hitachi NAS Platform, this supports clustering and acting as a quorum device in a cluster.

Hitachi Universal Replicator

With asynchronous replication over any distance, this provides for business continuity, optimization of resource usage, and improves efficiency as well as resilience.

4

5 SAP HANA Platform

This is a multi-purpose, in-memory database appliance to analyze transactional and analytical data.

Brocade VDX 6740-48 Switch

This 48-port switch provides 10 GbE connectivity to the appliance.

Brocade ICX 6430-48 Switch

This 48-port 1 GbE switch provides management network to the appliance.

Brocade ICX 6430-24 Switch

A 24-port 1 GbE switch that provides a NAS platform private network

Brocade 7800 Extension Switch

Providing up to sixteen 8 Gb/sec Fibre Channel ports and six 1 Gigabit Ethernet (GbE) ports for scalable bandwidth, port density, and throughput, this switch extends and optimizes storage area network (SAN) fabric connectivity over a distance to support business continuity and disaster recovery applications.

The Brocade 7840 switch can also be used instead of the Brocade 7800 switch. The 7840 switch has 24 ports of auto-sensing 2 Gb/sec, 4 Gb/sec, 8 Gb/sec, and 16 Gb/sec interfaces in a single 2U enclosure.

Figure 1 on page 6 shows the configuration of this solution.

5

6

Figure 1

6

7Hardware ElementsTable 1 describes the hardware used to deploy a three active nodes and one standby node configuration at each site. All the configurations of Hitachi Unified Compute Platform for the SAP HANA Platform support Hitachi Universal Replicator.

Table 1. Hardware Elements

Hardware Quantity Configuration Role

Hitachi Compute Blade 2500 chassis

4 8 server blade chassis

2 management modules

10 cooling fan modules

5 power supply modules

2 × FIVE-FX 16 Gb/sec 2-port Fibre Channel PCI-Ex card for each SAP HANA node

4 × 2-port 10GBASE-SR PCI-Ex card for each SAP HANA node

Server blade chassis.

One chassis is required for two SAP HANA nodes.

520X B2 server blade 24 2 × 18-core processors

768 GB RAM

2 pass through mezzanine card on Mezzanine Slot 2 and Slot 4 of each server blade, except for Server Blade 15

Server blade for the SAP HANA nodes.

Four server blades are required for each SAP HANA node.

SMP connection board for 520X server blade

6 4 server blade SMP connection board

SMP expansion module

SMP connector cover

SMP connector turns four physical blades into one SAP HANA node with 8 × 18-core processors and 3 TB of memory.

One SMP connector is required per SAP HANA node.

Hitachi NAS Platform 4060 4 For each NAS Platform server

2 cluster ports

2 × 10 Gb ports

2 Fibre Channel ports

2 Ethernet ports

Provide NFS shared file system for the following:

SAP HANA binaries

Cluster-wide configuration files

Storing one in-memory backup

7

8

System management unit software

2 2 Intel Core 2 Duo E7500 processor, 2.93 GHz CPU, 4 GB RAM

NAS Platform cluster management


2 MPB - 2 pairs

DKA (BED) - 2 pairs

CHA (FED) - 3 pairs

Block storage for SAP HANA nodes and NAS Platform

Quanta Cloud Technology QuantaPlex T41S-2U server

2 For T41S-2U server:

Intel Xeon E5-2620 v3 processor, 2.4 GHz CPU, 32 GB RAM

2 × 500 GB 7200 RPM SATA drives

1 dual port 10GbE Intel 82599ES SFP+OCP mezzanine card

1 dual port 1 GbE Base-T Intel i350 mezzanine card

Emulex Dual Port 8 Gb/sec Fibre Channel HBA

Management server runs the following:

NTP

Hitachi Command Suite

Hitachi Hi-Track Remote Monitoring system

SAP HANA Studio

Command control interface

Brocade VDX 6740-48 port switch

8 Two distinct VLANs, each dedicated to NFS and SAP HANA inter-cluster network

Two switches with one VLAN to provide uplink network to local area network infrastructure

10 GbE NFS and inter-cluster network

10 GbE client network

Brocade ICX 6430-48 port switch

2 1 GbE

48 ports

1 GbE Management Network

Brocade ICX 6430-24 port switch

2 1 GbE

24 ports

NAS Platform private network

Brocade 7800 switch 4 16 x 8Gb/sec Fibre Channel

6 x 1 GbE FCoE Switch

FCoE switch for storage connections

Table 1. Hardware Elements (Continued)

Hardware Quantity Configuration Role

8

9Hitachi Compute Blade 2500Hitachi Compute Blade 2500 delivers enterprise computing power and performance with unprecedented scalability and configuration flexibility. Lower your costs and protect your investment.

Flexible I/O architecture and logical partitioning allow configurations to match application needs exactly with Hitachi Compute Blade 2500. Multiple applications easily and securely co-exist in the same chassis.

Add server management and system monitoring at no cost with Hitachi Compute Systems Manager. Seamlessly integrate with Hitachi Command Suite in Hitachi storage environments.

This configuration uses 24 520X B2 server blades in the Hitachi Compute Blade 2500 chassis. Table 2 has the specifications for the 520X B2 server blades used in this solution.

Table 2. 520X B2 Server Blade Configuration

Feature Configuration

Processors Intel Xeon processor E7-8880

2 processors per server blade

Processor SKU Intel Xeon processor E7-8880 v3

Processor frequency 2.3 GHz

Processor cores 18 cores

Memory DIMM slots 48 slots populated

Memory 768 GB RAM

16 GB DIMMs

Network ports 1 USB 3.0 port

KVM connector (VGA, COM, USB 2.0 2-port)

Other interfaces Intel Xeon processor E7-8880

2 processors per server blade

9

http://www.hds.com/products/compute-blade/compute-blade-2500.html

10Hitachi NAS Platform 4060Hitachi NAS Platform is an advanced and integrated network attached storage (NAS) solution. It provides a powerful tool for file sharing, file server consolidation, data protection, and business-critical NAS workloads.

Powerful hardware-accelerated file system with multi-protocol file services, dynamic provisioning, intelligent tiering, virtualization, and cloud infrastructure

Seamless integration with Hitachi SAN storage, Hitachi Command Suite, and Hitachi Data Discovery Suite for advanced search and index

Integration with Hitachi Content Platform for active archiving, regulatory compliance, and large object storage for cloud infrastructure

This solution uses NAS Platform 4060 servers for file system sharing of the global binary and configuration SAP HANA files. There are four NAS Platform 4060 server nodes.

The system management unit provides front-end server administration and monitoring tools for NAS Platform. It supports clustering and acts as a quorum device in a cluster.

Hitachi Virtual Storage Platform G1000Hitachi Virtual Storage Platform G1000 provides an always-available, agile, and automated foundation that you need for a continuous infrastructure cloud. This delivers enterprise-ready software-defined storage, advanced global storage virtualization, and powerful storage.

Supporting always-on operations, Virtual Storage Platform G1000 includes self-service, non-disruptive migration and active-active storage clustering for zero recovery time objectives. Automate your operations with self-optimizing, policy-driven management.

The following reside on this storage device:

Operating system LUNs

Data LUNs

Log LUNs

LUNs for the Hitachi NAS Platform cluster

This solution uses two Hitachi Virtual Storage Platform G1000 storage platforms.

To properly size the storage, refer to Hitachi Unified Compute Platform for the SAP HANA Platform using 3 TB SAP HANA Nodes in a Scale-out 48 TB Configuration of 16 Active Nodes and 3 Standby Nodes with Hitachi Compute Blade 2500 Chassis, 520X B2 Server Blades, and Hitachi Virtual Storage Platform G1000 Reference Architecture Guide (AS-399-02).

At any time, the secondary site only has one live SAP HANA instance. The secondary site is for production or quality assurance. Normally the quality assurance instance is available for use at the secondary site in case of a service outage, when it becomes the production instance.

Use the server priority manager at the secondary site to do the following:

Designate the prioritized ports (replication) and non-prioritized ports (quality assurance).

Set the upper limits and thresholds for the I/O activity of these ports to prevent low-priority activities from negatively affecting the high priority activities, such as replication.

Additional information is available in the Hitachi Virtual Storage Platform G1000 Performance Guide.

10

http://www.hds.com/products/file-and-content/network-attached-storage/

http://www.hds.com/products/storage-software/hitachi-command-suite.html

http://www.hds.com/products/storage-software/data-discovery-suite.html

http://www.hds.com/products/file-and-content/content-platform/

http://www.hds.com/products/storage-systems/hitachi-virtual-storage-platform-g1000.html?WT.ac=us_mg_pro_hvspg1000

11Software ElementsTable 3 describes the software products used to deploy this three active node and one standby node configuration.

Table 3. Software Elements

Software Version

Hitachi Storage Navigator Modular 2 Microcode dependent

SMU software 12.4.3924.05

Hitachi NAS Platform firmware 12.4.3924.11

Hitachi Universal Replicator Microcode dependent

Command control interface 01-34-03/04

Hitachi Dynamic Link Manager 8.0.1

Operating System

Used on Hitachi Compute Blade 2500

SUSE Linux Enterprise Server for SAP Applications

SUSE Enterprise Linux 11 SP3

Alternate: Red Hat Linux Enterprise RHEL 6.6

SAP HANA 1.0 SPS09, Rev. 97 or later

SAP HANA Studio 1.0 SPS09, Rev. 97 or later

Microsoft® Windows Server® 2012 R2

Used on QuantaPlex T41S-2U

Standard Edition

Brocade VDX 6740 port switch 4.1.3b or later

Brocade ICX 6430-48 port switch 7.4.00 or later

Brocade 7800/7840 switch 7.4.00 or later

11

12Solution DesignThe detailed design for using Hitachi Universal Replicator with this Hitachi Unified Compute Platform for the SAP HANA platform solution is based on specifications from SAP and is a 3+1 node configuration that includes the following:

“Hitachi Compute Blade 2500 Chassis Configuration,” starting on page 12

“520X B2 Server Blade Architecture” on page 14

“Fibre Channel SAN Architecture,” starting on page 14

“Storage Architecture,” starting on page 15

“Hitachi NAS Platform 4060 Architecture,” starting on page 23

“Network File System Design for Shared Binaries,” starting on page 25

“Management Server,” starting on page 26

“Network Architecture,” starting on page 27

“SAP Storage Connector API Fibre Channel Client” on page 33

“Journal Sizing,” starting on page 33

“Inter-site Configuration,” starting on page 35

“Disaster Recovery and Replication Components,” starting on page 36

Hitachi Compute Blade 2500 Chassis ConfigurationThis solution uses four Hitachi Compute Blade 2500 chassis. Each chassis has of a total of eight 520X B2 server blades. Each SAP HANA node is four server blades connected using a four-blade SMP interface connector. This means that each chassis has two SAP HANA nodes.

The PCIe slots on each of the chassis have the following components.

One 2-port 10GBASE-SR PCI-Ex card for each of the following IOBD slots:

01B

03A

09B

11A

02A

04B

10A

12B

One FIVE-EX 16 Gb/sec 2-port Fibre Channel PCI-Ex card for each of the following IOBD slots:

01A

09A

04A

12A

12

13Figure 2 shows the front and back view of Chassis 1 for Hitachi Compute Blade 2500. Use the same configuration for the remaining chassis.

Figure 2

Table 4 on page 14 shows the Hitachi Compute Blade 2500 chassis configuration of all the SAP HANA nodes.

13

14

520X B2 Server Blade ArchitectureThis solution uses 12 520X B2 server blades for each node in the 3 TB SAP HANA configuration with two active nodes plus one standby node for each disaster recovery site.

Each SAP HANA node has four server blades connected using the four-blade SMP interface connector. This creates a single eight-socket SMP node with 144 cores and 3 TB of memory.

Table 5 lists the 3-node server blade configuration per site.

Fibre Channel SAN ArchitectureEach of the 520X B2 server blades has a pass-through mezzanine card on Mezzanine Slot 2 and Mezzanine Slot 4. These slots connect to the 16 Gb/sec Hitachi FIVE-FX Fibre Channel PCI-Ex card and to the 10GbE NIC card installed in the PCI-Ex slots through the backplane within the server chassis.

For each SAP HANA node, there are two dedicated Fibre Channel ports on Hitachi Virtual Storage Platform G1000. The Fibre Channel SAN architecture of a SAP HANA node consists of two Fibre Channel cables. Each Fibre Channel cable connects the 16 Gb/sec Hitachi FIVE-FX Fibre Channel PCI-Ex cards on the chassis with the designated Virtual Storage Platform G1000 ports.

Table 6 shows the storage port mapping, which is identical for both sites.

Table 4. Hitachi Compute Blade 2500 Chassis Configuration of SAP HANA Nodes for Each Site

Chassis Server Blades SAP HANA Node Name

Role of SAP HANA Node

Chassis 1 1, 3, 5, and 7 SAP HANA Node 1 Master

Chassis 1 9, 11, 13, and 15 SAP HANA Node 2 Worker

Chassis 2 1, 3, 5, and 7 SAP HANA Node 3 Standby

Table 5. Server Blade Configuration at Each Site

Server Blades Total 12 server blades

3 nodes with 4 SMP-connected server blades per node

Total Number of CPU Cores 432

Total Memory 9 TB

Table 6. Storage Port Mapping

SAP HANA Node

Chassis PCI-Ex slot, Port Virtual Storage Platform G1000 Ports

Node1 Chassis 1 IOBD 01A, Port 0 1C






HNAS HNAS1 Hitachi NAS Platform Server 1, FC Port 1 1A

14

15

Set the port properties for the direct connection between Hitachi Compute Blade 2500 and Hitachi Virtual Storage Platform G1000, as shown in Table 7.

The Hitachi FIVE-FX 16 Gb/sec HBA can emulate FC-SW virtually. Set the BIOS for the Hitachi FIVE-FX HBA to enable for Multiple Port ID. Use the Fabric storage port setting of ON to set FC-SW virtual mode.

Storage ArchitectureThe central storage system for the SAP HANA scale-out cluster is a Hitachi Virtual Storage Platform G1000 storage platform. Several usage aspects divide the space provided by Virtual Storage Platform G1000, as follows:

Operating system LUN provisioning for SAP HANA nodes

Log device provisioning for SAP HANA database

Data device provisioning for SAP HANA database

Block storage provisioning for Hitachi NAS Platform shared file system to store the SAP HANA binaries and cluster-wide configuration files.

We follow a design that utilizes a building block approach in multiples of four active nodes. So we use a four active node building block design to provision storage for this setup

Figure 3 on page 17 shows the RAID group configuration for the Virtual Storage Platform G1000 architecture used in the SAP HANA appliance with four active nodes and one standby node.

Each SAP HANA node has its own data volume and log volume. Only active SAP HANA nodes need data volumes and log volumes. Standby nodes do not require these volumes.

HNAS HNAS1 Hitachi NAS Platform Server 1, FC Port 3 2A

HNAS HNAS2 Hitachi NAS Platform Server 2, FC Port 1 1B

HNAS HNAS2 Hitachi NAS Platform Server 2, FC Port 3 2B

Table 7. Port Properties

Property Value

Port Attribute Target

Port Security Disabled

Port Speed Auto

Fabric ON

Connection Type P-to-P

Table 6. Storage Port Mapping (Continued)

SAP HANA Node

Chassis PCI-Ex slot, Port Virtual Storage Platform G1000 Ports

15

16This design follows a four-node building block approach for the SAP HANA data volumes, log volumes, and shared binaries. Provision the parity groups in Figure 3, as follows.

Operating System Boot LUN

Each node has its own 100 GB LUN on Virtual Storage Platform G1000 for the operating system volumes.

A single parity group configured as RAID-6 (6D+2P) on 900 GB drives provisions the operating system LUN for SAP HANA nodes 1 to 5 (up to a maximum of 19) on Virtual Storage Platform G1000.

From this parity group, create five LDEVs, each with a capacity of 100 GB.

Map each LDEV exclusively to the corresponding SAP HANA node as follows: LUN ID 00.

The installation of SUSE Linux Enterprise Server for SAP Applications or Red Hat Enterprise Linux resides on the operating system LUN.

Hitachi NAS Platform 4060 Block Storage

The block storage for Hitachi NAS Platform consists of three parity groups on Virtual Storage Platform G1000, each configured as RAID-6 (6D+2P) on 24 × 900 GB drives to store the shared binaries and configuration files of the SAP HANA database.

In each of the three parity groups, create two LDEVs of 2400 GB each.

Create a dynamic provisioning pool named HNAS_HDP_pool. Assign all the created Hitachi NAS Platform LDEVs to this pool. This allows the use of all the disks concurrently on NAS Platform for better performance.

For a 4-node building block, create six virtual volumes, each with 2400 GB in HNAS_HDP_pool. Do the LUN path assignment for these virtual volumes to the ports in Table 6 on page 14 connected to Hitachi NAS Platform.

SAP HANA Log Volumes

For the SAP HANA log volumes, create two parity groups configured as RAID-6 (6D+2P) on 16 × 900 GB drives.

In each of the parity groups, create two LDEVs of 600 GB each.

Map each SAP HANA log volume to all SAP HANA nodes at each port with the LUN ID 1 of the specified host.

SAP HANA Data Volumes

For the SAP HANA data volumes, create four parity groups, each configured as RAID-6 (14D+2P) on 64 × 900 GB drives.

Create four LDEVs with a capacity of 2816 GB per each parity group. Table 8 on page 18 shows the parity groups and LDEVs created for data volumes.

Assign four LDEVs (with LUN ID 2-5) for use as data volumes to each SAP HANA node, as shown in Table 8.

16

17

Figure 3

Table 8 on page 18 shows the parity groups and LDEV assignment of boot volumes, the Hitachi NAS Platform volumes, the SAP HANA log volumes, and the SAP HANA data volumes for production system at both the sites.

17

18

Table 9 on page 19 shows the dynamic provisioning pool IDs and virtual volume LDEV IDs for Hitachi NAS Platform.

Table 8. Groups and LDEV Assignment of Operating System Boot, Hitachi NAS Platform, SAP HANA Log Volumes, and SAP HANA Data Volumes for Production System at Primary and Secondary

Parity Group ID

Parity Group RAID Level and disks

LDEV ID Name LDEV Size

1 RAID-6 (6D+2P) on 900 GB 10k RPM SAS drives

00:01:00 HANA_OS_LUN_N1 100 GB





00:00:01 HNAS_VOL_1 2400 GB

00:00:02 HNAS_VOL_2 2400 GB


00:00:03 HNAS_VOL_3 2400 GB

00:00:04 HNAS_VOL_4 2400 GB


00:00:05 HNAS_VOL_5 2400 GB

00:00:06 HNAS_VOL_6 2400 GB


00:01:01 HANA_LOG_N1 600 GB

00:02:01 HANA_LOG_N2 600 GB


00:03:01 HANA_LOG_N3 600 GB

00:04:01 HANA_LOG_N4 600 GB


00:01:02 HANA_DATA_N1_01 2816 GB

00:02:02 HANA_DATA_N2_01 2816 GB

00:03:02 HANA_DATA_N3_01 2816 GB

00:04:02 HANA_DATA_N4_01 2816 GB


00:01:03 HANA_DATA_N1_02 2816 GB

00:02:03 HANA_DATA_N2_02 2816 GB

00:03:03 HANA_DATA_N3_02 2816 GB

00:04:03 HANA_DATA_N4_02 2816 GB


00:01:04 HANA_DATA_N1_03 2816 GB

00:02:04 HANA_DATA_N2_03 2816 GB

00:03:04 HANA_DATA_N3_03 2816 GB

00:04:04 HANA_DATA_N4_03 2816 GB


00:01:05 HANA_DATA_N1_04 2816 GB

00:02:05 HANA_DATA_N2_04 2816 GB

00:03:05 HANA_DATA_N3_04 2816 GB

00:04:05 HANA_DATA_N4_04 2816 GB

18

19

While mapping the LUN path assignment for each node, add the LUNs in the following order:

1. Map the OS LUN for the specific SAP HANA node.

2. Map the log volume and data volume of each SAP HANA node.

Table 10 shows an example configuration of the LUN path assignment for node01 on primary site. The LUN assignment would be the same for all nodes except for the first LUN, which would be the operating system LUN of that specific node.

Table 9. Dynamic Provisioning Pool IDs and Virtual Volume LDEV IDs of Hitachi NAS Platform for Production System at Both Sites

Dynamic Provisioning Pool ID

Dynamic Provisioning Pool Name

Virtual Volume Names Virtual Volume LDEV ID for Hitachi NAS Platform Shared Binaries

Virtual Volume Size for Hitachi NAS Platform Shared Binaries

MPB Assignment

0 HNAS_HDP_Pool HNAS_HANA_VVOL_1 00:0A:01 2400.00 GB MPB0

HNAS_HANA_VVOL_2 00:0A:02 2400.00 GB MPB4





Table 10. LUN Path Assignment at Primary Site

LUN ID LDEV ID LDEV Name

0000 00:01:00 hananode01

0001 00:01:01 LOG_1

0002 00:01:02 DATA_1_01

0003 00:01:03 DATA_1_02

0004 00:01:04 DATA_1_03

0005 00:01:05 DATA_1_04

0006 00:02:01 LOG_2

0007 00:02:02 DATA_2_01

0008 00:02:03 DATA_2_02

0009 00:02:04 DATA_2_03

0010 00:02:05 DATA_2_04

0011 00:03:01 LOG_3

0012 00:03:02 DATA_3_01

0013 00:03:03 DATA_3_02

0014 00:03:04 DATA_3_03

0015 00:03:05 DATA_3_04

0016 00:04:01 LOG_4

19

20

Figure 4 shows the LUN and port assignment for the maximum SAP HANA server nodes 19 on each Hitachi Virtual Storage Platform G1000.

Figure 4

0017 00:04:02 DATA_4_01

0018 00:04:03 DATA_4_02

0019 00:04:04 DATA_4_03

0020 00:04:05 DATA_4_04

Table 10. LUN Path Assignment at Primary Site (Continued)


20

21Table 11 shows the parity groups and LDEV assignment the Hitachi NAS Platform volumes, the SAP HANA log volumes, and the SAP HANA data volumes for quality assurance system at the secondary site.

Table 11. Groups and LDEV Assignment of Operating System Boot, Hitachi NAS Platform, SAP HANA Log Volumes, and SAP HANA Data Volumes for Quality Assurance System at the Secondary Site

Parity Group ID


LDEV ID LDEV Name LDEV Size


00:00:25 HNAS_VOL_QA_1 2400 GB

00:00:26 HNAS_VOL_QA_2 2400 GB


00:00:27 HNAS_VOL_QA_3 2400 GB

00:00:28 HNAS_VOL_QA_4 2400 GB


00:00:29 HNAS_VOL_QA_5 2400 GB

00:00:30 HNAS_VOL_QA_6 2400 GB


00:01:06 HANA_LOG_QA_N1 600 GB





16 RAID-6 (14D+2P)RAID-6 (14D+2P) on 900 GB 10k RPM SAS drives

00:01:07 HANA_DATA_QA_N1_01 2816 GB

00:02:07 HANA_DATA_QA_N2_01 2816 GB

00:03:07 HANA_DATA_QA_N3_01 2816 GB

00:04:07 HANA_DATA_QA_N4_01 2816 GB


00:01:08 HANA_DATA_QA_N1_02 2816 GB

00:02:08 HANA_DATA_QA_N2_02 2816 GB

00:03:08 HANA_DATA_QA_N3_02 2816 GB

00:04:08 HANA_DATA_QA_N4_02 2816 GB


00:01:09 HANA_DATA_QA_N1_03 2816 GB

00:02:09 HANA_DATA_QA_N2_03 2816 GB

00:03:09 HANA_DATA_QA_N3_03 2816 GB

00:04:09 HANA_DATA_QA_N4_03 2816 GB


00:01:10 HANA_DATA_QA_N1_04 2816 GB

00:02:10 HANA_DATA_QA_N2_04 2816 GB

00:03:10 HANA_DATA_QA_N3_04 2816 GB

00:04:10 HANA_DATA_QA_N4_04 2816 GB

21

22Table 12 shows the dynamic provisioning pool IDs and virtual volume LDEV IDs for Hitachi NAS Platform.

Table 13 shows an example configuration of the LUN path assignment for Node001 on the primary site. The LUN assignment would be the same for all nodes except for the first LUN, which would be the operating system boot LUN of that specific node.

Table 12. Dynamic Provisioning Pool IDs and Virtual Volume LDEV IDs of Hitachi NAS Platform for Quality Assurance System at the Secondary Site



Virtual Volume Names Virtual Volume LDEV ID for NAS Platform Shared Binaries

Virtual Volume Size for NAS Platform Shared Binaries

MPB Assignment

1 HNAS_QA_HDP_Pool

HNAS_QA_VVOL_1 00:0A:25 2400.00 GB MPB0






Table 13. LUN Path Assignment at the Secondary Site


0000 00:01:00 hananode01

0001 00:01:01 LOG_1

0002 00:01:02 DATA_1_01

0003 00:01:03 DATA_1_02

0004 00:01:04 DATA_1_03

0005 00:01:05 DATA_1_04

0006 00:02:01 LOG_2

0007 00:02:02 DATA_2_01

0008 00:02:03 DATA_2_02

0009 00:02:04 DATA_2_03

0010 00:02:05 DATA_2_04

0011 00:03:01 LOG_3

0012 00:03:02 DATA_3_01

0013 00:03:03 DATA_3_02

0014 00:03:04 DATA_3_03

0015 00:03:05 DATA_3_04

0016 00:04:01 LOG_4

0017 00:04:02 DATA_4_01

0018 00:04:03 DATA_4_02

0019 00:04:04 DATA_4_03

22

23

Hitachi NAS Platform 4060 ArchitectureThis describes the architecture for Hitachi NAS Platform 4060 and related systems.

This reference architecture uses two Hitachi NAS Platform 4060 file system modules. They are part of Hitachi Virtual Storage Platform G1000 in the cluster configuration. The two NAS Platform servers are cluster interconnected with two 10 Gb/sec network links.

Each NAS Platform server directly connects to the Virtual Storage Platform G1000 target ports using two Fibre Channel cables, from each NAS Platform 4060 node.

System Management UnitWeb Manager, the graphical user interface of the system management unit, provides front-end server administration and monitoring tools. It supports clustering and acts as a quorum device in a cluster. This solution uses an external system management unit that manages two NAS Platform servers.

Use one of the following browsers to run Web Manager:

Microsoft Internet Explorer®, version 9.0, or later

Mozilla Firefox, version 6.0, or later

0020 00:04:05 DATA_4_04

0021 00:01:01 LOG_QA_1

0022 00:01:02 DATA_QA_1_01

0023 00:01:03 DATA_QA_1_02

0024 00:01:04 DATA_QA_1_03

0025 00:01:05 DATA_QA_1_04

0026 00:02:01 LOG_QA_2

0027 00:02:02 DATA_QA_2_01

0028 00:02:03 DATA_QA_2_02

0029 00:02:04 DATA_QA_2_03

0030 00:02:05 DATA_QA_2_04

0031 00:03:01 LOG_QA_3

0032 00:03:02 DATA_QA_3_01

0033 00:03:03 DATA_QA_3_02

0034 00:03:04 DATA_QA_3_03

0035 00:03:05 DATA_QA_3_04

0036 00:04:01 LOG_QA_4

0037 00:04:02 DATA_QA_4_01

0038 00:04:03 DATA_QA_4_02

0039 00:04:04 DATA_QA_4_03

0040 00:04:05 DATA_QA_4_04

Table 13. LUN Path Assignment at the Secondary Site (Continued)


23

24Private Management NetworkConnect the private management interfaces of the Hitachi NAS Platform 4060 servers and the system management unit to the Brocade ICX 6430-24 port switch. Devices connected to this private ICX 6430-24 port switch are only accessible through the system management unit.

Public Data NetworkThe public data network consists of the public Ethernet port of Hitachi NAS Platform 4060 and system management unit software connected to the Brocade ICX 6430-48 switch.

Storage SubsystemThis solution uses Hitachi Virtual Storage Platform G1000 as the storage subsystem. Hitachi NAS Platform has direct attached Fibre Channel connections with Virtual Storage Platform G1000.

Server ConnectionsFigure 5 shows the back of Hitachi NAS Platform 4060.

Figure 5

Port C1 and Port C2 are the NAS Platform 4060 cluster ports. To enable clustering, do the following:

Connect Port C1 of first NAS Platform server to Port C1 of the second NAS Platform server.

Connect Port C2 of first NAS Platform server to Port C2 of the second NAS Platform server.

Port tg1 and Port tg2 are 10 GbE ports. Link aggregate and connect these ports to the Brocade VDX 6740-48 switches.

Connect Fibre Channel Port FC1 and Port FC3 directly to the Hitachi Virtual Storage Platform G1000 ports, as follows:

1A and 2A for the first NAS Platform server

1B and 2B for the second NAS Platform server

Connect port eth0 of the NAS Platform server to the management network on the Brocade ICX 6430-48 port switch. Connect port eth1 of the NAS platform server to the private network on the Brocade ICX 6430-24 port switch.

24

25For the direct connection between Hitachi NAS Platform and Hitachi Virtual Storage Platform G1000, set the port properties as shown in Table 14.

Network File System Design for Shared BinariesThis solution requires a network file system to store cluster-wide SAP HANA binaries and configuration files of the in-memory database. Host this shared file system, called /hana/shared/<SID>, on Hitachi NAS Platform 4060. Mount this file system on all SAP HANA nodes. <SID> is the system ID for the SAP HANA production database instance.

Table 8 on page 18 and Table 9 on page 19 show the parity group setup and the twenty four LDEVs and virtual volumes used for NAS Platform in this configuration.

This solution uses six virtual volume LDEVs, as listed in Table 9 on page 19.

Refer to each LDEV as a system drive.

With these system drives, create a single storage pool called HANABIN_<SID>.

Configure two EVS on the NAS Platform nodes, as follows:

EVS on NAS Platform node 1 as HNASEVS1

EVS on NAS Platform node 2 as HNASEVS2

Create the shared file system hana_shared_<SID> using the storage pool HANABIN_<SID> with the following:

Capacity based on the number of SAP HANA nodes:

For up to 4 active HANA nodes: 13 TB




Block size of 32 KB

Auto expansion disabled

Mount and then export the file system. Mount the NFS export /hana_shared_<SID> on the file system path /hana/shared/<SID> on all nineteen SAP HANA nodes.

Set the MTU size to 9000 on both NAS Platform nodes.

Follow similar procedure for the QA system on the secondary site.

Table 14. Hitachi Virtual Storage Platform G1000 Port Properties

Property Value

Port Attribute Target

Port Security Disable

Port Speed Auto (8 Gb/sec)

Fabric OFF

Connection Type FC-AL

25

26Management ServerThis solution uses one node of a four-node Quanta Cloud Technology QuantaPlex T41S-2U server for the management server. The management server acts as a central device for managing the SAP HANA platform.

Manage the following from the management server:

Hitachi Compute Blade 2500 chassis

520X B2 server blades

Brocade ICX 6430 — 48 port switch

Brocade VDX 6740 switches

System management unit

SAP HANA nodes

Hitachi NAS Platform 4060 servers


NTP configuration

Hi-Track Remote Monitoring system from Hitachi Data Systems

Hitachi Command Suite and management of the server blades

SAP HANA Studio

Figure 6 on page 27 shows the management server network ports using one dual port 1 GbE Base-T Intel i350 mezzanine card.

Slot 01 Port 2 — Connect this port to the local area network. It provides 1 GbE network to the management server.

Slot 01 Port 1 — Connect this port to the Brocade ICX 6430-48 port switch that provides 1 GbE network to all other switches, chassis, and Hitachi NAS Platform nodes.

The management server has the following additional components:

One dual port 10 GbE Intel 82599ES SFP+ OCP mezzanine card

One Emulex 2-port 8 Gb/sec Fibre Channel HBA on the PCIe slot

Connect the 10 GbE network ports to two different Brocade switches, VDX 6740-A and VDX 6740-B, to provide management access to the SAP HANA nodes from the Quanta 2U4N server using the NFS network.

26

http://www.quantaqct.com/catalog/microsite/Product/Servers/Rackmount-Servers/Multi-node-Servers/QuantaPlex-T41S-2U-4-Node-p282c77c70c83c185

27

Figure 6

Install the following software on the management server:

Hitachi Command Suite

Hitachi Compute Systems Manager

Command control interface

Hitachi Dynamic Link Manager

NTP server service

PuTTY

Teraterm

JRE version jre-7u51-windows-i586 (no 64 bit)

Adobe Flash Player

WinSCP

SAP HANA Studio

Network ArchitectureFor the client network, two Brocade VDX 6740 switches provide external connectivity. In the solution rack, the switch placement is as follows:

The switch at rack unit 40 is the VDX 6740-D switch 54

The switch at rack unit 39 is the VDX 6740-C switch 53

27

28The scale-out solution for SAP HANA internally uses the NFS network and the inter-cluster network only. They can share a switch-pair, given that traffic can be strictly separated. Accomplish this by using VLANs on two ISL-paired Brocade VDX 6740 switches. In the solution rack, refer to the switches as follows:

The switch at rack unit 37 is the VDX 6740-B switch 52

The switch at rack unit 36 is the VDX 6740-A switch 51

The scale-out SAP HANA solution requires five separate networks.

SAP HANA Inter-Cluster Network

This network provides communication between the SAP HANA instances on the cluster.

Connect the SAP HANA inter-cluster network to Switch 51 and Switch 52 (Brocade VDX 6740). Set an MTU size of 9100, in accordance with Brocade best practices.

Isolate using a VLAN of 100.

Each of these ports utilize active-active bonding mode:

Switch 51 and Switch 52 (VDX 6740 switches)

Port eth9901 and Port eth9902 of the Linux operating system.

SAP HANA Client Network

This network is dedicated to traffic between the SAP HANA database and its clients.

Connect the SAP HANA client network to Switch 53 and Switch 54 (Brocade VDX 6740). Set with an MTU size of 9100, in accordance with Brocade best practices.

By default, isolate using a VLAN of 200. However, a different VLAN can be used to isolate this network.


Switch 53 and Switch 54 (Brocade VDX 6740 switches)

Port eth9921 and Port eth9922 of the Linux operating system.

Connect the up-link ports of the Brocade VDX 6740-48 switches to the local area network.

SAP HANA NFS Network

This network provides access to SAP HANA shared binaries and configuration files using the NFS protocol.

Connect the SAP HANA NFS network to Switch 51 and Switch 52 (Brocade VDX 6740). Set with an MTU size of 9100, in accordance with Brocade best practices.

Isolate using a VLAN of 150.


Port eth9911 and Port eth9912 of the Linux operating system

28

29 Hitachi NAS Platform Private Network

This network is used for the NAS Platform heartbeat.

A 1 GbE Brocade ICX6430-24 port switch supplies the heartbeat.

Management Network

This network is used with the management server. See “Management Server,” starting on page 26.

This resides on a 1 GbE Brocade ICX 6430-48 port switch.

The management network does not need to have a VLAN assigned to it.

The Brocade ICX 6430-48 port switch uses the default switch configuration.

The SAP HANA inter-cluster network, client network, and NFS network are required to have the following:

No single point of failure

Provide at least 10 GbE equivalent throughput

To meet these requirements, use four dual-port 10 GbE PCIe cards per SAP HANA node. Bond two ports from different PCIe network adapters at the operating system level using link aggregation, following the IEEE 802.3ad Link Aggregation standard for each of the three networks:

SAP HANA inter-cluster

NFS

Client network

Connections of each bond need to go to physically different VDX 6740 switches, so when one switch fails there is still another route to the corresponding host. This solution connects two switches together using inter switch link. It lets both switches act together as one single logical switch with the characteristics that, if one switch fails, there still is a path to the hosts.

The complete network setup uses the ports on the 10 GbE PCI-Ex cards listed in Table 15 as an example configuration for chassis 1. Follow the similar configuration for SAP HANA nodes on other chassis as well.

Table 15. Network Setup Using 10 GbE PCI-Ex NIC Cards

PCI-Ex Slot Number Port Operating System Level eth Port

Network Description

IOBD 01B 0 eth9921 Client Network for Node01

IOBD 01B 1 eth9911 NFS Network of Node01

IOBD 02A 0 eth9902 Inter-cluster Network for Node01



IOBD 04B 1 eth9912 NFS Network for Node01


IOBD 09B 1 eth9911 NFS Network of Node02


29

30

SAP HANA Inter-cluster Network Connection for Node 1 Figure 7 shows the SAP HANA inter-cluster network connection for Node 1.

Figure 7



IOBD 12B 1 eth9912 NFS Network for Node02

Table 15. Network Setup Using 10 GbE PCI-Ex NIC Cards (Continued)

PCI-Ex Slot Number Port Operating System Level eth Port

Network Description

30

31Configure the SAP HANA inter-cluster network using operating system-level bonding on every node. Make the following 10 GbE connections, as listed in Table 16.

SAP HANA NFS Network Connection for Node 1 Figure 8 shows the SAP HANA NFS network connection for Node 1.

Figure 8

Table 16. SAP HANA Inter-Cluster Port Mapping

Chassis, PCI-Ex Slot Number, Port VDX 6740 Switch and Port Bond

Chassis 1, IOBD 03A, Port 0 VDX 6740-A, Port #1 Bond 0 of Node 1

Chassis 1, IOBD 02A, Port 0 VDX 6740-B, Port #1 Bond 0 of Node 1

Chassis 1, IOBD 11A, Port 0 VDX 6740-A, Port #2 Bond 0 of Node 2

Chassis 1, IOBD 10A, Port 0 VDX 6740-B, Port #2 Bond 0 of Node 2

31

32Configure the SAP HANA NFS network using operating system-level bonding on every node. Make the 10 GbE connections using the mappings in Table 17.

SAP HANA Client Network ConnectionFigure 9 shows the SAP HANA client network connection.

Figure 9

Table 17. SAP HANA NFS Network Port Mappings


Chassis 1, IOBD 01B, Port 1 VDX 6740-A, Port #21 Bond 1 of Node 1

Chassis 1, IOBD 04B, Port 1 VDX 6740-B, Port #21 Bond 1 of Node 1

Chassis 1, IOBD 09B, Port 1 VDX 6740-A, Port #22 Bond 1 of Node 2

Chassis 1, IOBD 12B, Port 1 VDX 6740-B, Port #22 Bond 1 of Node 2

32

33Configure the SAP HANA client network using operating system-level bonding on every node. Make the following 10 GbE connections using the port mappings shown in Table 18.

SAP Storage Connector API Fibre Channel ClientThe SAP HANA storage connector API Fibre Channel client defines a set of interface functions called during the following:

Normal SAP HANA cluster operation

Failover handling

Storage connector clients implement the functions defined in the storage connector API.

The scale-out of Hitachi Unified Compute Platform for the SAP HANA platform uses the fcClientLVM implementation, which supports the use of logical volume manager. SAP supports this solution to enable the use of high-performance Fibre Channel devices for a scale-out installation.

The fcClientLVM implementation uses standard Linux commands, such as multipath and sg_persist. Install and configure these commands.

The fcClientLVM implementation is responsible for mounting the SAP HANA volumes. It also implements a proper fencing mechanism during a failover by means of SCSI-3 persistent reservations.

Configuration of the SAP Storage Connector API is contained within the SAP global.ini file in /hana/shared/<SID>/global/hdb/custom/config.

Journal SizingThe following are maximum throughput numbers, per journal:

If the distance between the sites is less than 2000 km (about 1200 miles), then the maximum throughput is about 750 MB/sec.

If the distance is greater than 2000 km, then the maximum throughput is about 500 MB/sec.

If the distance is greater than 5000 km (about 3100 miles), then the maximum throughput is about 350 MB/sec.

With one RAID group configured as RAID-6 (6D+2P) using 900 GB SAS drives, testing in the Hitachi Data System lab achieved a throughput of 69 MB/sec.

For the base configuration of the solution, it is recommended to use at least two RAID groups configured as RAID-6 (6D+2P) with 900 GB SAS drives in a dynamic provisioning pool for use as a journal group.

Use this is the formula used to calculate the journal capacity:

Journal Capacity = Peak hour megabytes change × Number of hours

Table 18. SAP HANA Client Network Port Mappings


Chassis 1, IOBD 01B, Port 0 VDX 6740-C, Port #1 Bond 2 of Node 1

Chassis 1, IOBD 04B, Port 0 VDX 6740-D, Port #1 Bond 2 of Node 1

Chassis 1, IOBD 09B, Port 0 VDX 6740-C, Port #2 Bond 2 of Node 2

Chassis 1, IOBD 12B, Port 0 VDX 6740-D, Port #2 Bond 2 of Node 2

33

34Table 19 provides the journal sizing information.

Using the two RAID groups configured as RAID-6 (6D+2P), create LDEVs as in Table 20.

Create a dynamic provisioning pool named HUR_Journal_Group and assign all the LDEVs created in Table 20 to this pool. Create the four virtual volumes in Table 21, each with 2400 GB for use as a journal volume.

To achieve a zero RPO, for a distance lower than 2000 km (about 1240 miles) we need the maximum of 750 MB/sec journal throughput which is only possible by the use of 11 RAID-6 (6D+2P) parity groups.

Table 19. Journal Sizing

Number of RAID groups configured as RAID-6 (6D+2P)

Hitachi Universal Replicator Peak Write Throughput in MB/sec

Hitachi Universal Replicator Peak Write Throughput in GB/hour

Hitachi Universal Replicator Journal Space in GB

Hours for journal space: (Journal space in GB)/(peak GB/hour)

2 138 485 9600 19.7

3 207 727 14400 19.8

4 276 970 19200 19.7

5 345 1212 24000 19.8

6 414 1455 28800 19.7

7 483 1698 33600 19.7

8 552 1940 38400 19.7

9 621 2183 43200 19.7

10 690 2425 48000 19.7

11 759 2668 52800 19.7

Table 20. LDEVs for Use with HUR_Journal_Group Dynamic Provisioning Pool

Parity Group ID


LDEV ID LDEV Name LDEV Size MPB Assignment


00:0B:01 Journal_1 2400 GB MPB0





Table 21. Dynamic Provisioning Pool IDs and Virtual Volume LDEV IDs



Virtual Volume Names

Virtual Volume LDEV ID

Virtual Volume size

MPB Assignment

1 HUR_Journal_Group_1 Journal_VVOL_1 00:0C:01 2400.00 GB MPB0

Journal_VVOL_2 00:0C:02 2400.00 GB MPB4

2 HUR_Journal_Group_2 Journal_VVOL_3 00:0C:03 2400.00 GB MPB0

Journal_VVOL_4 00:0C:04 2400.00 GB MPB4

34

35During implementation of the Hitachi Universal Replicator service, from Hitachi Data Systems Global Services needs to right-size the journal volumes based on the actual workload, throughput, and RPO requirement. Then, the following needs to happen:

Add additional disks to create a journal volume of RAID groups configured as RAID-6 (6D+2P) on 900 GB drives.

Create additional LDEVs and add them to the existing dynamic provisioning pool HUR_Journal_Group.

For every single RAID group added to the pool, create two virtual volumes of 2400 GB size to use as a journal volume.

Use this is the formula used to calculate the journal capacity:

Journal Capacity = Peak hour megabytes change × Number of hours

Inter-site ConfigurationSAN switch and long distance amplifiers must be between the primary site and the secondary site, as and when applicable.

If you have the SAN switches and long distance amplifiers in the existing infrastructure, you can utilize the same infrastructure.

Table 22 lists the target and initiator ports of the two storage systems at each site along with zoning alias. Both sites need one zoning configuration.

Table 23 has the details on the zoning between the management server and Hitachi Virtual Storage Platform G1000.

Create a single zone named MGMT_Pri_Sec.

Add all four Virtual Storage Platform G1000 ports aliases and all of the four management server Emulex port aliases in the MGMT_Pri_Sec zone.

With this configuration, command devices on the primary system and secondary system are accessible on the management servers at the primary site and secondary site.

Table 22. Zoning Configuration

Initiator Port RCU Target Port Zone Name

PrimA_Port3B SecB_Port7B PrimA_Port3B_SecB_Port7B

PrimA_Port4B SecB_Port8B PrimA_Port4B_SecB_Port8B

SecB_Port3B PrimA_Port7B SecB_Port3B_PrimA_Port7B

SecB_Port4B PrimA_Port8B SecB_Port4B_PrimA_Port8B

Table 23. Hitachi Compute Rack Management Server Zoning

Virtual Storage Platform G1000 Alias

Hitachi Compute Rack Management Server Alias

Zone Name

PrimA_Port5B MGMT_PRI_Port0 MGMT_Pri_Sec

SecB_Port5B MGMT_SEC_Port0 MGMT_Pri_Sec

PrimA_Port6B MGMT_PRI_Port1 MGMT_Pri_Sec

SecB_Port6B MGMT_SEC_Port1 MGMT_Pri_Sec

35

36Disaster Recovery and Replication ComponentsIn this reference architecture for SAP HANA disaster recovery, Hitachi Universal Replicator setup requires doing the following:

“Install Command Control Interface” on page 36

“Install Hitachi Dynamic Link Manager” on page 36

“Configure Command Devices” on page 36

“Configure Replication using Command Control Interface,” starting on page 36

“Setup Hitachi Universal Replicator,” starting on page 39

Install Command Control InterfaceCommand control interface enables you to perform storage system operations by issuing commands to Hitachi Virtual Storage Platform G1000.

In this solution, install command control interface on the management server used as a management server for the primary site and the secondary site. Command control interface has the following components residing in these locations:

Storage System — Command devices and Hitachi Universal Replicator volumes (P-VOLs and S-VOLS)

Quanta Cloud Technology QuantaPlex T41S-2U Management Server — Hitachi Open Remote Copy Manager

Install Hitachi Dynamic Link ManagerInstall the Hitachi Dynamic Link Manager software on the management server for the primary site and the secondary site. Dynamic Link Manager manages the access paths to the storage system.

Dynamic Link Manager provides the ability to distribute loads across multiple paths and switch to another path if there is a failure in the path currently being used, improving system availability and reliability.

Configure Command DevicesA command device is a dedicated logical volume on the storage system that functions as the interface to the storage system from the host. The command device accepts commands from the host that are executed on the storage system.

In this solution, create a 100 MB command device logical volume on the local and remote storage system. Each management server has one dual-port Emulex HBA card, connected through the Brocade 7800 or Brocade 7840 switch to Hitachi Virtual Storage Platform G1000. Perform zoning and add the LUN path for the command devices in such a way that the primary site management server and the secondary site management server can access the command devices on both storage systems.

Configure Replication using Command Control InterfaceA key aspect of this reference architecture on Hitachi Virtual Storage Platform G1000 is defining the volume pair relationship for replication between storage systems. Define and manage storage replication relationships through the Hitachi Storage Navigator graphical user interface or on the management server running Hitachi Open Remote Copy Manager.

The Hitachi Open Remote Copy Manager operates as a daemon process on the host. When activated, Hitachi Open Remote Copy Manager refers to its configuration files. The Hitachi Open Remote Copy Manager instance communicates with the storage sub-system and remote servers.

36

37Two instances of Open Remote Copy Manager are required for Hitachi Universal Replicator replication to be operational.

A Hitachi Open Remote Copy Manager instance on the primary management server manages the P-VOLs.

A Hitachi Open Remote Copy Manager instance on the secondary management server manages the S-VOLs.

The Hitachi Open Remote Copy Manager configuration file defines the communication path and the logical units to be controlled. Each instance has its own configuration file. The configuration file lists the following for replication:

SAP HANA data volumes

Log volumes

Hitachi NAS Platform LUNs

Figure 10 shows the content of the configuration file (horcm04.conf) used by the Hitachi Open Remote Copy Manager instance on the primary management server.

Figure 10

37

38Figure 11 shows the configuration file (horcm06.conf) files used by the Hitachi Open Remote Copy Manager instance on the secondary management server.

Figure 11

Figure 12 lists the entries that need for adding to the services file of the management server on both sides for Open Remote Copy Manager to function.

Figure 12

38

39Setup Hitachi Universal ReplicatorThe reference architecture setup is as follows:

For the failover to the secondary site, this solution uses two initiator ports on Hitachi Virtual Storage Platform G1000 at the primary site connected to two RCU target ports on Hitachi Virtual Storage Platform G1000 at the secondary site.

For the failback to the primary site, it uses two initiator ports from the storage system at the secondary site connected to the two RCU target ports on the storage system at the primary site.

The initiator and RCU target ports on Hitachi Virtual Storage Platform G1000 on the primary site connect to the Brocade 7800 Fibre Channel switch at the primary site.

The initiator and RCU target ports on Hitachi Virtual Storage Platform G1000 on the secondary site connect to the Brocade 7800 Fibre channel switch at the secondary site.

Define the port attributes for initiator and target ports on Hitachi Virtual Storage Platform G1000. Configure the storage system for Hitachi Universal Replicator replication by defining the logical paths for replication. Refer to the Hitachi Virtual Storage Platform G1000 Hitachi Universal Replicator User Guide for the setup of logical paths for replication and defining port attributes.

Configure Hitachi Universal ReplicatorConfiguring Hitachi Universal Replicator for Hitachi Unified Compute Platform for the SAP HANA platform requires the following steps. The detailed implementation steps are in Hitachi Virtual Storage Platform G1000 - Hitachi Universal Replicator User Guide. Contact your Hitachi Data Systems representative if you need this document.

1. Define the Fibre Channel port attributes.

Define the port attributes for initiator and target port on Virtual Storage Platform G1000.

2. Setup the logical path between the primary system and the secondary system.

Associate the primary system and the secondary system in the Universal Replicator relationships and define the logical paths between them. This is required in order to perform the Universal Replicator pair operations.

3. Create parity groups and LDEVs for journal groups.

Create journal group RAID groups and LDEVs for use as journal volumes, as described in Journal Sizing.

4. Register journal volumes in a journal.

Register every journal volume in the journal group.

Start the Hitachi Open Remote Copy Manager InstanceTo start up the Hitachi Open Remote Copy Manager instance on the primary management server, issue the following command:

C:\horcm\etc\horcmstart 04

To start up the Open Remote Copy Manager instance on the secondary management server, issue the following command:

# C:\horcm\etc\horcmstart 06

39

40Initial Pair Copy OpearationPerform the initial data transfer, called the pair copy operation, between the primary site and the secondary SAP HANA node volumes in data/log/HNNAS with the following command, issued on the primary side:

# C:\horcm\etc\paircreate -IH04 –g HANADR -vl –f async 0

When executing the paircreate command, the initial copy happens. The primary storage system copies all the data in sequence from the P-VOL directly to the S-VOL.

During the initial copy process, the status of the P-VOL and S-VOL is COPY. On completion of the initial copy process, the status of the P-VOL and S-VOL changes to PAIR.

Initial Configuration for Replication of Hitachi NAS Platform 4060 LUNsTo use Hitachi Universal Replicator to replicate the SAP HANA LUNs on Hitachi NAS Platform, perform the following initial configuration steps. This is a one-time configuration to setup the RAID mirror relationship between LUNs on Hitachi NAS Platform. The steps only list a high-level overview of the important RAID mirroring commands on NAS Platform.

To perform the initial configuration, do the following.

1. Complete the initial pair copy (pair create) operation between the primary site and the secondary site for SAP HANA these volumes:

Data

Log

NAS Platform

2. Perform the initial NAS Platform mirroring.

(1) On any instance of command control interface, validate that the primary site and the secondary site are correctly paired and synchronized at "PAIR" status.

(2) On the console of NAS Platform for the secondary site, do the following:

i. Verify HANA_SHARED_<SID> shared file system is unmounted.

ii. Copy the cluster information (cluster UUID and the cluster name) in a text editor window and keep it for later.

iii. Log on as root to the SMU of the secondary site.

iv. Verify that you are in the root folder.

v. Run the commands to get the luids and cod information of the NAS Platform disks to be replicated.

vi. Copy these files to the SMU on the primary site.

(3) On the console of NAS Platform for the primary site, do the following:

i. Log on as root to the SMU of the primary site.

ii. Go to the folder where you copied the files from the secondary site.

iii. Run the command to mirror the disks.

iv. List the span information to confirm the mirror relationship.

v. Add the UUID of the NAS Platform on the secondary site to the span, using the cluster UUID and cluster name using information from Step 2.ii .

40

41 (4) Log on to the secondary site:

i. Rescan the cod for the span.

ii. List the span. Expect a clean output, not showing the production instance span.

3. Test the configuration by performing failover to the secondary site.

(1) Ensure the pair is synchronized completely.

(2) Shutdown Production on the Primary site

(3) Unmount /hana/shared/, the data file system, and the log file system at the primary site for production instance on the SAP HANA nodes as well as Hitachi NAS Platform servers.

(4) Go to the secondary site and do the following:

i. If it exists, shut down the quality assurance instance.

ii. Unmount /hana/shared/, the data file system, and the log file system at the secondary site for quality assurance instance on the SAP HANA nodes as NAS Platform.

iii. Using command control interface, run a pair split with a read write.

iv. List the file system. This should show HANA_SHARED_<SID> as unmounted.

v. After the first failover, use 'evsfs' to bind each file system to an EVS at the secondary site.

vi. Mount the production shared file system on NAS Platform.

vii. Mount the production shared file system on SAP HANA nodes.

viii.Check at the operating system level of the SAP HANA nodes if the /hana/shared file system is mounted.

ix. Start the SAP HANA system and validate if it successfully comes up.

4. Test the configuration by performing failback to primary site.

(1) Shutdown the production SAP HANA database at the secondary site. Umount the hana_shared_<SID> file system on the secondary NAS Platform cluster.

(2) Do a pair resync with swaps. Once it is 100% completed, start the failback.

(3) At the primary site, do the following:

i. On command control interface, run the pair split with read and write.

ii. On the SMU, list the spans. This ensures the mirror relationship.

iii. Mount the hana_shared_<SID> file system on the primary NAS Platform cluster and HANA nodes. Check at the operating system level of SAP HANA nodes if the /hana/shared file system is mounted.

iv. Start the SAP HANA system and validate if it successfully comes up.

v. Once validation completed, run the pair resync with swaps.

If one site is permanently lost and the surviving LUs are promoted into the SSWS state, it is necessary to run 'sd-peg -up' on the surviving NAS Platform cluster to make it treats the S-VOLs as P-VOLs. Otherwise, 'sd-peg' should never be used.

41

42In addition, registry changes made on one cluster while it is in production always need recording when made, and then copied to the other cluster after the next failover:

If creating any new file systems have, bind them to EVSs using the 'evsfs' command, and then exported.

If deleting any file systems on one cluster, delete them from the registry of the other cluster by the filesystem-forget-and-delete-nv-data command.

If any exports have been created, deleted, or modified on one cluster, make the same changes on the other cluster.

42

43Engineering ValidationValidation of this reference architecture was conducted in the Hitachi Data Systems laboratory. The steps to failover to the secondary site and then to failback to primary site with soft and hard techniques using Hitachi Universal Replicator were tested.

Test MethodologyTo test the setup, the following scenarios were executed in the lab:

Planned failover to the secondary site

Planned failback to the primary site

Planned failover to the secondary site during mid-flush and then failback to the primary

Unplanned failover to the secondary site

Automated failover and failback using Hitachi Disaster Recovery Manager for SAP HANA without quality assurance on the secondary site

Test ResultsAll the tests passed without issues. The RTO was less than an hour.

43

For More InformationHitachi Data Systems Global Services offers experienced storage consultants, proven methodologies and a comprehensive services portfolio to assist you in implementing Hitachi products and solutions in your environment. For more information, see the Hitachi Data Systems Global Services website.

Live and recorded product demonstrations are available for many Hitachi products. To schedule a live demonstration, contact a sales representative. To view a recorded demonstration, see the Hitachi Data Systems Corporate Resources website. Click the Product Demos tab for a list of available recorded demonstrations.

Hitachi Data Systems Academy provides best-in-class training on Hitachi products, technology, solutions and certifications. Hitachi Data Systems Academy delivers on-demand web-based training (WBT), classroom-based instructor-led training (ILT) and virtual instructor-led training (vILT) courses. For more information, see the Hitachi Data Systems Services Education website.

For more information about Hitachi products and services, contact your sales representative or channel partner or visit the Hitachi Data Systems website.

http://www.hds.com/services/

http://www.hds.com/corporate/resources/

http://www.hds.com/services/education/

http://www.hds.com/services/

http://www.hds.com/corporate/resources/

http://www.hds.com/services/education/

http://www.hds.com/

http://www.hds.com/

1

Corporate Headquarters2845 Lafayette StreetSanta Clara, CA 96050-2639 USAwww.HDS.com community.HDS.com

Regional Contact InformationAmericas: +1 408 970 1000 or [email protected], Middle East and Africa: +44 (0) 1753 618000 or [email protected] Pacific: +852 3189 7900 or [email protected]

HITACHI is a trademark or registered trademark of Hitachi, Ltd. © Hitachi Data Systems Corporation 2015. All rights reserved. Hi-Track is a trademark or registered trademark of Hitachi Data Systems Corporation. Microsoft, Windows Server, and Internet Explorer are trademarks or registered trademarks of Microsoft Corporation. All other trademarks, service marks, and company names are properties of their respective owners.

Notice: This document is for informational purposes only, and does not set forth any warranty, expressed or implied, concerning any equipment or service offered or to be offered by Hitachi Data Systems Corporation.

AS-440-00, November 2015,

www.hds.com

community.HDS.com

mailto:[email protected]



tel:+14089701000

tel:+4401753618000

+85231897900

https://twitter.com/HDScorp

https://www.linkedin.com/company/hitachi-data-systems

https://www.facebook.com/HitachiDataSystems

https://www.youtube.com/user/HDScorp

Documents

Disaster Recovery using Hitachi Universal Replicator on ...docshare01.docshare.tips/files/30620/306203529.pdfThis reference architecture guide depicts the necessary components and