While performing a VRA Target host operation I get the following error:
VPG col-pcs-col-vas-zt3 rollback failed. Error is: Grain pool size expected value 314572800 but was 354418688.
Now my only options are Force Sync or Delete for the VPG and Force Sync is not working. Any ideas or suggestions?
did u solve this? i’m having same issue.
I ended up deleting the VPG. The root issue was trying to use the preseed option and the names are not logical and I apparently picked the wrong one causing that error. Sometimes I think the go to thing is deleting the VPGs…
AH ok. I was just changing recovery hosts in order to do vmware patches and put them into maint mode, and a large server got this error. I had to do the same thing. Shame because it takes 3 days to replicate again.
I also ran into the same issue with the delta/bitmap sync’ing taking a long time to catch up. I did find a fix that seemed to correct that. Support noticed a lot of “delete” in the log files and recommended the following:
I’m writing to follow up on case 117868. Here’s the summary I sent on May 14:
Our Development team is still working to determine the root cause of the issue. They’re seeing many messages like this for VPG col-pcs1-COL-EMPFS1-zt6:
DeleteCheckpointsInBatches,Failed to DeleteCheckpoints
That means that the ZVM produced checkpoints to its ShortTerm table and after that decided to delete them directly from the ShortTem (hard delete). However, the ZVM isn’t keeping up with the pace necessary to delete the checkpoints. This operation may be the reason of the write locks. In addition, no new checkpoints are being added because the operations are failing at the delete stage.
As a workaround, our Development team recommends the following steps:
1) Stop the ZVM services
2) As Administrator, add the following line:
t_RepositorySettings_DeleteCheckpointsBatchSize = 20
To the following file:
C:Program FilesZertoZerto Virtual Replicationtweaks.txt
3) Restart the ZVM services
This will slow down the rate of checkpoint operations by reducing the number that are handled in one batch (the default is 250). Our Development team believes that this will allow the delete operations to catch up and then allow checkpoint insertions to progress.
Since I made the change I have not had any issues when we move VPGs and put our hosts into maintenance mode.
That may help you too.