sometimes, while performing backups of 7-8 TB VMs, I find errors like this one in the logs, especially when such jobs take a very long time
Network timeout. DDB Engine did not respond within the timeout period. DDB: MI_GDP_LOCAL_VMs_6, DDB Id: 6. Source: MI-MEDIA, Process: vsbkp
this example comes from Jan 28 at 5:31:13 am
At the same time, in the Events tab of the Job Details I see the following event
Failed to remove the backup snapshot for virtual machine [MI-SERVER], the snapshot may need to be manually removed. [The operation has timed out.]
I think it is reasonable to assume they are related, correct?
Given such cases, should I extend the timeout for snapshot deletion?
Thank you!
Gaetano
Best answer by Damian Andre
It seems likely - especially if you are not seeing the same error on other backups.
7/8 TB VMs very large - as soon as the snapshot is in place, the VM freezes the base disks and starts accumulating changes to a child disk. When the backup is complete (probably many, many hours later), the snapshot removal tasks causes VMware to merge all the changes from the snapshot back into the base disk.
This is a very taxing I/O operation and for such a large VM or a VM with high change rate, can take a long time to process disk consolidation. Even though the snapshot removal times out, VMware should still be processing it in the background and the next VM backup job will check that it completed before it takes a new snap.
The backup should be completed with errors rather than failed, correct?
It seems likely - especially if you are not seeing the same error on other backups.
7/8 TB VMs very large - as soon as the snapshot is in place, the VM freezes the base disks and starts accumulating changes to a child disk. When the backup is complete (probably many, many hours later), the snapshot removal tasks causes VMware to merge all the changes from the snapshot back into the base disk.
This is a very taxing I/O operation and for such a large VM or a VM with high change rate, can take a long time to process disk consolidation. Even though the snapshot removal times out, VMware should still be processing it in the background and the next VM backup job will check that it completed before it takes a new snap.
The backup should be completed with errors rather than failed, correct?
