Result of Leaving a VPG in the Failover Test, Failover-Before-Commit, or Move-Before-Commit States for a Long Duration

KB Number:
00000135

Solution:

Purpose


With Zerto Virtual Replication, an administrator may wish to test a VPG in one of the following manners, without risking losing VPG CDP History journal data:

  • Failover Test
  • Live Failover, in the Failover-Before-Commit state
  • Move, in the Move-Before-Commit state
In each of these instances, it is important to keep in mind that the testing should not go on for a long duration.  This solution article explains the reason why this is the case.

 

 

Affected Versions


All Versions

 

Complexity


Low

Solution


When in any of the following states, the recovered VM test IOs are written to a separate volume, instead of directly to the target disks:
  • Failover Test
  • Live Failover, in the Failover-Before-Commit state
  • Move, in the Move-Before-Commit state
In versions of Zerto Virtual Replication prior to 2.0 Update 4 (Zebra), the volume used is the CDP history journal.  Starting with Zerto Virtual Replication 2.0 Update 4 (Zebra), a separate scratch volume is generated in each instance, and is removed after the VPG is no longer in a testing state.
The primary effect of this is that, while in these states, the recovered VMs have read-only access to the target disks, and the administrator can safely check the application for consistency, without risking writing to the target disks, or losing other journal checkpoints for the VPG.
As all IO writes during these states are written to a scratch volume, it is important to limit the duration of these states, as when the volume becomes full, the applications can no longer write IOs, which will lead to corruption and other errors.  At this point, the only option is to stop the Failover Test, or Rollback the Live Failover/Move, and start again.

1 Star2 Stars3 Stars4 Stars5 Stars (2 votes, average: 1.00 out of 5)
Loading...