Zerto offers a feature that should allow to perform maintenance on vmware hosts without having impact on running VPGs. All inbound VPG/VMs can be moved to other esxi host with, so called “Change VM recovery VRA”.
This feature works well if you are not performing changes on your infrastructure. Unfortunately logic that is used in code in certain scenarios breaks that feature.
Here are the details:
Zerto site had been configured with datastores that were purposely deployed for customer DR workloads and zerto journals. Then some datastores were decommissioned alongside zerto journal ones. Journal vmdks were moved with storage vMotion and we had no errors on UI. Following that we started to have issues with moving VPG to other DR VRA.
We’ve open a case with support and hand over logs for review. It has been found that some datastores that we decommissioned need to be in site settings/ cloud settings/ Datastore configuration, for us to be able to move VPG as we needed. Unfortunately, we also hit a wall when we tried to move VPGs for which journals were on datastore that had been permanently decommissioned.
The case had been escalated to development team. They’ve responded with acknowledgement that they are aware about the issue and currently there is no plan to fix the problem.
Now we are in situation where for number of VPGs we cannot move them as error says that there is missing datastore. We cannot add that datastore to site configuration as it doesn’t exist anymore.
Support team is not really supportive as this problem may not be covered by internal KB articles and they don’t know what to do.
Thus my questions:
Do you think re-configuring site with zvm tools would remove the deleted datastores from zvm database/config files and allow “Change VM recovery VRA”?
Should I try to remove all VPGs and create those from scratch as new VPG would use existing/enabled datastores and not deleted ones?
Would there be a way to update zerto database/config and update journal default datastore to new datastore, so we could make change in back end and avoid huge work with re-creating VPGs?
All comments welcome…
We are experience the same problem when we delete datastores (after emptying it). What we do is deleting VPGs and preseed the disks and create the VPGs again. If there is plenty of VPGs, you can try Importing VPGs from zerto tools.