SQL Server on Linux Journey – Part 4: High Availability
On Part 3 of this series we covered SQL Server on Azure, including Azure VMs running on SUSE and the “Bring-Your-Own-Subscription” route.
Now, we’ll move on to cover how to add High Availability with SUSE Linux Enterprise Server HA Extension to the mix.
High Availability for SQL Server on Linux
The big question is: how much does downtime cost and how much a customer can afford to lose?
Enterprises use SUSE Linux for their 24/7 mission-critical workloads to leverage our reliability/scalability, and database administrators (DBAs) and architects must possess extensive high availability understanding to architect a highly available platform.
Balaji Barmavat, a former SUSE Public Cloud Solution Architect and Sourabh Agarwal, a Sr. Program Manager for the Microsoft Data Platform Group described SQL Server HA on Linux components and terminology as below:
Components include pacemaker (the core clustering component, that does things like coordinate across the clustered machines. It includes CoroSync, a framework and set of APIs that provides things like quorum, the ability to restart failed processes etc. managing communication between nodes); fence method (there are different methods to take a cluster out if a node is unresponsive so we can protect data); libQB (library for logging); SQL Server HA resource agent (specific functionality to integrate with Pacemaker and monitor availability groups); external cluster type (new clustering mechanism introduced to communicate with Linux distributions).
Terminology includes node (any computer); resource (any type of service or application that is known to Pacemaker, such as IP address, a file system, or a database); resource group (holds together cluster resources and when one fails, the entire cluster group is taken offline and failed over to another available cluster node); constraints (how you determine the behavior of a resource); fencing (isolating resources and locking those out of a node whose status is uncertain); and quorum (number of voting elements that must be part of active cluster membership for that cluster to start properly or continue running).
You can watch a demonstration on how to configure SUSE Linux Enterprise Server Pacemaker HA on Azure at this
What about HA on Azure?
If you choose Azure VMs, while users will be able to set up a pacemaker cluster on SUSE SLES VM on Azure, high availability in the SUSE repository will only be enabled in a future SQL Server community update.
As mentioned above, to have a highly available environment with SUSE on Azure today, customers need to Bring their own subscriptions to Azure SLES 12 SP5 and SUSE HA Extension and configure SLES Cluster for SQL Server Availability Group. You can watch a demonstration at this PASS Session starting at 35:25.
Next Stop… Part 5
On Part 5 of this series we talk about containers and later we’ll move on to close with Big Data Clusters. Stay tuned!