- This topic has 3 replies, 3 voices, and was last updated January 16, 2020 by Bar A.
Failback after loss of Primary Datacentre
Shira LJanuary 13, 2020 11:40:22 AM
We are planning for a DR Test in which we will need to fail our VPGs to the secondary datacentre whilst the primary datacentre is taken offline.
This means we will need to commit the failover and will not be able to enable reverse replication (as the primary dc will uncontactable)
Once we finish the test, we will re-establish communication to the primary datacentre
How would we then fail the VPGs back to the primary datacentre
Do we need to re-create VPGs and if yes what is the impact in terms of data having to replicated vs using pre-seeded disks
Bar from Zerto here.
Once a VPG will be failed over with no reverse protection configured, the VPG will enter “needs configuration state”.
Once communication between the 2 sites will be established again, you will be able to configure the VPG settings pointing to the site that was disconnected and by that failing back.
Important to note that the source VMs in this case, will still exist on the production site so you will need to delete them manually (make sure to remove from inventory and not deleting from disk ) in order to utilize the disks as preseed.
The 2nd option would be simply going over an initial sync with keeping the VMs on both sites.
Does the VPG in a needs config state retain all the information about where the VM was before failover along with all the network/storage info? My guess is not since that information is only defined for the recovery site when setting up a VPG. If you failed over several hundred VM’s it could take days to reconfigure unless I’m missing something.
1. In a scenario that the VPG has been failed over when there is no connection between the sites, the VPG will enter “needs configuration state” in that state the VPG does not have any configuration of the recovery site ( which was the protected site before the Failover ).
Hence in that scenario, we will need to configure each VPG with the new recovery settings.
2. In a scenario that there is a connection between the sites once the recovery operation is initiated, and Reverse-Protection was enabled during the recovery operation we will expect the following to occur:
a. What stood for the protected settings will now stand for the recovery settings.
b. What stood for the recovery settings will now stand for the protected settings.
c. What used to be the protected VM will be deleted in order to utilize the disks as preseed.
d. The VPG will enter Delta-sync to match the state of the volumes in the recovery site. (Delta-sync is required to verify the consistency of the volumes on both ends, regardless of the chosen checkpoint is 1 week old or 5 secs old.)
3. In a scenario that there is a connection between the sites once the recovery operation is initiated, and Reverse-Protection was not enabled the VPG will enter “needs configuration state” thus resulting in the need to configure each VPG with the new recovery settings.
You can find additional information about recovery operation in our ZVM administration guide, Page 268 – Chapter 13.