Achieving Zero Downtime for SQL

Achieving Zero Downtime for SQL

©2017 ScaleArc. All Rights Reserved. 2

Downtime is disruptive and costly

• Poor user experience• Decreased productivity• Wasted time and IT resources• Lost revenue


Two Sources of Downtime

• Unplanned Outages– A strong failover architecture is imperative

– Instantly identify and automatically reroute traffic away from a failed database server, even across data centers

– Avoid application errors and user dissatisfaction

• Planned Maintenance– Carefully examine your database maintenance strategy

• Are you bringing servers completely offline for maintenance?

• Are you disabling applications associated with those databases during maintenance?


Want Guaranteed Downtime?

Keep your applications directly connected to your databases

This 1:1 tie will guarantee that any database hiccup will make your application

throw up

Or…


Leverage the Fact that Databases are Getting Smarter

Modern databases offer:• Scale out• Better replication• Automatic failover

But make one key change:End the 1:1 tie between apps and databases by inserting an

abstraction layer.


An Abstraction Layer is Critical for Zero Downtime

• Database Load Balancing Software acts as a buffer

PrimaryPrimary SecondaryPrimary

replicationSecondary

• Leverages database replication and failover

• Makes failover invisible to the app by queing writes until the new primary is ready

• Enables servers to go offline for maintenance without taking down the app

Database Load Balancing Software


ScaleArc’s Database Load Balancing Software

• Intelligently directs traffic into the database on behalf of the app• Deploys transparently, with no changes to app or database• Leverages and augments database’s replication and failover


Avoid Downtime during Unplanned Outages


replicationSecondary



replication


Secondary

1 ScaleArc performs read/write split



replication


Secondary


RRR

WR W R R

W


Primary Secondary

replication


Secondary


R

WR W R R

W2 Upon database failure,

ScaleArc queues writesW

W


Primary Secondary

replication


Secondary


R

WR W R R

2 Upon database failure, ScaleArc queues writes

3 After secondary is promoted, ScaleArc directs queries to new primary server

Primary

WWW


Primary Secondary

replication


Secondary


R

WR W R R

2 Upon database failure, ScaleArc queues writes

3 After secondary is promoted, ScaleArc directs queries to new primary server

Primary

WWW

4 Apps see delay, but do not restart


SecondarySecondary

replication

Enable Zero Downtime Maintenance

SecondaryPrimary

WRRWR

RW

WR R


Secondary

replication


Secondary

1 Mark a database to be taken offline

Primary

WRRWR

RW

WR R


Secondary

replication


Secondary


Primary

WRRWR

RW

WR

2 ScaleArc automatically reroutes queries away from offline server


Secondary

replication


Secondary


Primary

WRRWR

RW

WR


3 Patch the server and bring it back online


SecondarySecondary

replication


SecondaryPrimary

WRRWR

RW

WR R

4 After health check, ScaleArc auto restores traffic to updated server

3 Patch the server and bring it back online




ScaleArc Keeps Business Data Flowing

• Avoid lost productivity and business profitability• Ensure customer satisfaction• Keep critical data flowing


Achieving Zero Downtime for SQL

Want to learn more? Click here.

Technology

Achieving Zero Downtime for SQL