Hi. What’s the effect of snapshotting a protected VM? We are looking to do an in place OS upgrade on a VM but need a rollback plan.
For protection, having a snaphot has no impact on Zerto. It just keeps replicating IO to the destination. When you delete the snapshot, again, no impact to Zerto. It’s already written that IO and VMWare is simply applying it to the base disk. Everything is good.
Where you can get into big trouble though, is if you have to rollback the snapshot. You will have out of sync disks in your production and DR datastores. To fix it, you must perform a “full sync” on the VPG and let it get the disks back in sync. Failure to do this will cause corruption if you ever try to failover that VPG.
I have raised this issue with Zerto and asked them to build in a feature that detects snapshot rollbacks and executes a full sync automatically. So far that feature has not been implemented. I have created my own scripts to do this automatically for me. It’s not real time, but it runs once a day so my window of risk is very low.
Hope that helps.
Just as I thought – thanks!
Full sync is happening. System automatically issued full sync and fixed the issue with snapshot rollback.
But by doing so it zero out whole VPG history. Normally I have 10 days history and it went to 0.
Also I did restored only one VM,but whole VPG got zero-out Journal History. To make matter worse I have few VPG that should be close in time to do DR. When one lost history that means all other become non-recoverable.
So the fact is once you do rollback of snapshot evan on one VM you loose all Journal History for your environments.
I guess long time retention might help a bit to keep at least few points of recovery.