SNAP HA on Azure

SnapReplicate™ allows for manual replication of volumes to another node in the event of a problem or a planned maintenance. SNAP HA allows replication to be triggered automatically in such a case, by establishing a heartbeat between linked instances. If the heartbeat fails to register for more than a few moments, the other instance takes over, ensuring seamless access to the provisioned data.

To set up SNAP HA, start the process by clicking Add SNAP HA.

Configuration of SnapReplicate™ is a prerequisite to setup of SNAP HA™. If SnapReplicate™ is not configured, the Add SNAP HA™ button will be grayed out.

If you have not yet configured a notification email, the opportunity to provide one will be presented prior to continuing SNAP HA™. Provide an email address for support reports and logs to be sent to, and click OK.

The Add High Availability wizard opens, providing an Instruction window. Click Next.

Next, the Azure credentials tab will appear, with the credentials auto-populated from choices made in virtual machine creation process.

Verify the credentials as necessary and click next.

Here you will create and add an IP Address that is within the same CIDR block as the instances. In simplest terms, ensure that the IP address starts with the same numbers as the two instances, but which is not in use. To avoid conflicts, list the private IPs of your virtual machines within the subnet, a well as any virtual IPs used for other HA pairings.

The update all subnets option is an advanced option that can be useful if you do not have a large number of subnets. If hundreds of subnets are in use, it is recommended not to select this option.

The Virtual IP on same subnet option should only be checked if using VNET peering. This too is an advanced option, and not within the scope of this guide.

Click Next once the virtual IP has been provided.

Next, you can tune the behavior of your HA pairing.
1. You can determine the max number of retries before your virtual machine fails over.
2. You can determine the max time (in seconds) that storage can be unavailable before a failover is triggered.
3. You can also set a default for max ioping request time, to ensure that a failover is triggered more quickly in event of failure.
4. Finally, you can determine the behavior of the failed node during a failover.
  1. Reboot - this is the default option, allowing for quicker recovery and re-establishment of high availability, as the failed node will reboot, and SNAP HA will be reactivated, with the original node set as secondary.
  2. Shutdown - The failed node will remain shut down. You will need to reboot the instance manually to re-establish high availability.
  3. None (No action taken) - This option is only for debug or support use. The failed node will remain in its current state.
Click Next to continue.

Click Finish on the Finish HA Setup screen.

To test, shut down one of the instances. The other will become primary after a few moments. Alternatively, select Actions, and Takeover to simulate a failover. Follow the instructions found in Recovering from a High Availability Failure to re-establish your highly available configuration.