Unplanned failover of applications
When the primary site of the given application is unreachable, you can run unplanned failover of applications to your local cluster.
About this task
-
Important: In this document, the site where the application is running is the primary site and peer site is the secondary site of an application.
- Run the unplanned failover from any of the following racks:
- In case of rack shutdown, from the surviving rack.
- In case of network disconnection, from the rack where the applications must be redeployed.
- In case of AFM shutdown, from the rack AFM that is in good state.
- To prepare the cluster to host the selected applications, the failover procedure creates the PVs, which are needed for the application to access data.
- When you create applications and PVCs on a primary site, its associated PVs or filesets have a
quota limit. You can replicate the entire data on the primary to the secondary as no quota limit
exists in the secondary. The following issues can occur because of quota limit:
- Some data may not get replicated when you do a failover from the new primary without quota limits to the new secondary with quota limits.
- In the new secondary site, the Regional status of some of these failover applications may show as RPO missed
- The Snapshot may go missing for few applications post restart replication
Procedure
What to do next
After forced failover, restart replication option is provided on the old primary, that is, for applications that were forced failed over to the peer cluster. For the procedure to restart, Restarting replication.