Follow

About

SRM allows running the recovery plans using one of the following modes:

Running recovery plans is performed directly from the VMware SRM user interface.

  • When VMware SRM carries out the selected recovery plan it sends the relevant commands to the InfiniBox system on the Recovery / Protect site through the InfiniBox SRA.
  • The commands are described below. 

Test

Running a plan in Test mode will attempt to recover the virtual machines in a test environment on the recovery site.

If "Replicate recent changes to recovery site" is used, the Test mode validates that the recent changes were successfully synced from the protected to the recovery site. 

The Test mode performs the following actions on the InfiniBox in the Recovery site:

  1. For each volume in the protection group - Create a snapshot. (on replication target)
  2. Create a child snapshot(s) on the previously created snapshot(s).
  3. Make the child snapshot(s) Write-enabled. 
  4. Map the child snapshot(s) to an ESXi host.

Cleanup

Cleanup performs the following actions on the InfiniBox in the Recovery site:

  1. Unmap the child snapshot(s) from the ESXi host(s).
  2. Delete the snapshot(s) and the child snapshots(s).

Planned migration and Disaster recovery

Running a plan in one of these modes will attempt to shut down the VMs at the protected site and recover the VMs at the recovery site.

Planned migration

Both sites and the InfiniBox replication must be connected and available.

  • Validates that the recent changes were replicated to the Recovery site.
  • If errors are encountered the recovery process ceases.

Disaster recovery

Attempt to replicate recent changes to the recovery site, but otherwise use the most recent synchronized data that exists on the recovery site.

  • The recovery process continues even if errors are encountered.

Recovery process for Async & Sync replication

Both modes performs the following InfiniBox actions:

On the Protected site

  • Make the volume(s) in the protection group Write-protected. (if possible)

During Planned migration the replication state of the volume(s) in the protection group will become Auto Suspended.

On the Recovery site

  1. For each volume in the protection group (replication target) - Suspend the replication. 
    • Volume's replication state becomes Suspended.
  2. Make the volume(s) Write-enabled.
  3. Map each volume to an ESXi host.

Reprotect

Running reprotect will commit the results of the recovery, and configure protection in the reverse direction.

  • Therefore, the InfiniBox replication is resumed in the reverse direction.

Reprotect performs the following InfiniBox actions:

On the Protected site

  • Unmap each volume in the protection group from the ESXi host(s). 

On the Recovery site

  • Resuming the replication in the reverse direction.
    • The volume(s) are now the replication source.

At the end of the Reprotect process, the Protected and the Recovery sites are swapped.

Recovery process for Active-Active replication

Due to the nature of Active-Active replication, SRM does not need not perform any operation directly on the InfiniBox systems. (by default) 

On the Recovery site

SRM will:

  1. Validate that each volume in the protection group is mapped to the ESXi hosts.
    • If an Active-Active peer is not mapped, it will be mapped. 
  2. Register and power-on (if configured) the protected VMs.

Isolating the Recovery site

In case SRM cannot power-on the protected VMs at the recovery site due to file locks, SRM tries to isolate the Active-Active peer on the Recovery site by unmapping the corresponding Active-Active peer on the Protected site.

Reprotect

Running reprotect will commit the results of the recovery, and configure protection in the reverse direction.

  • No operations will be performed directly on the InfiniBox systems.

At the end of the Reprotect process, the Protected and the Recovery sites are swapped.


Failback as soon as possible to the original Protected Site

As the Active-Active replication Preferred system is set for the InfiniBox system in the original Protected site, it is important to failback as soon as possible. (As soon as the original Protected site is up and running)

Was this article helpful?
0 out of 0 found this helpful

0 out of 0 found this helpful

Comments