If the primary cluster fails, the mirror cluster can take over its operations after a small manual intervention. The manual procedure makes the mirror instance into the primary.
In the event that the production cluster is destroyed, monitoring and alerting will notify the administrator. The administrator can then make the mirror into the new primary, by stopping it from pulling backups generated by the old primary. These steps define the mirror cluster as the new primary. Then a new mirror can be deployed.
- If it is still running, disconnect the primary from the network.
-
Log in to the mirror, and issue the command to stop it from acting as a mirror:
$ tscli backup stop-mirror -
It may take some time for the mirror to finish recovering the last backup from the primary. Issue the command to verify that the cluster has stopped running in mirror mode before continuing:
$ tscli backup mirror-status -
Start up the periodic backups on the mirror (which is now the new primary):
$ tscli backup set-periodic --at <hour1,hour2, ...> --directory <shared_backup_directory> [--num_backups <num_backups>] - Deploy a new mirror appliance when possible.