Ten minute backup delay, with VM snapshots

[Andreas Neufert]

We are still discussing this topic with VMware, but the first signals are that there will be no change or enhancement on the VMware side. VMware gave us some ideas on how to change the architecture completely to the point that there will be no storage snapshots anymore to workaround the issue and as well to the point that it would not work at scale. This is not acceptable for us. We will continue to investigate and work with VMware on this, but the process is not that smooth with what is happening currently within VMware.

[Spex]

Thanks for the update. What about the easy way - shouldn't be that complicated to implement:
- Option to automatically skip VMs with snapshots from BfSS
(there are also several other approaches that Veeam can implement without VMware)

This is really a problem for us in daily operations...

[Andreas Neufert]

Yes, I asked the responsible team already where we stand with this. Idea is currently to not plaster the UI with another checkbox, but to automate things based on timing.

[TWuser]

Had one recently that hung for 60 hours. It wouldn't let me cancel (whether Graceful or Immediate) so I had to disable all other jobs, wait for running jobs to completely, and reboot (service restart would be similar but only save ~10 minutes).

Now I'm trying to figure out if I have fragmentation or something else (yes I checked for snapshots), but either way I wish it would simply time out after a certain amount of time.

[Andreas Neufert]

Can you please share ticket number and job name, so that my team can have a look?

[foggy]

The upcoming 12.2 will have a configurable timeout value for this operation. We'll share the details once it is published.

[Spex]

We are now using 12.2 and are using this new function. It is better to have it, but it is still far from satisfactory.

Last night we had the situation again where a backup job with hardware snapshots did not start due to a vcenter snapshot for all VMs. At least all VMs were processed after the default wait time. The VM with the snapshot was marked with fail and then tried several times with the same method, which of course did not work. Why is there not a failback to no hardware snapshot support (e.g. nbd or whatever). So I have to keep intervening, change the job manually and send another retry. Every day until the customer has deleted his authorized snapshot.

I still see an urgent need for action here and would like the last missing step to be implemented as soon as possible.

[foggy]

Hi Stefan, this last missing step will be available with the next update. Thanks.

[chalkynz]

Yeah this is pretty major risk when taking on storage-integration backups (which are recommended). I haven't tested it for a couple of years, I will try again. Anyone else had progress/workarounds?

[Andreas Neufert]

I would say the majority of the storage integration environments run fine and having snapshot present at the VMs is something that most of the customer avoid. Those that need the functionality use the vCenter integration of our Quick Backups to give the admins the capability of performing very fast real backups instead of snapshots.

In the meantime we are making some progress with VMware on this topic, but it still takes some time. Current idea is to replace the used VMware API with something new as the old used method is limited on the technology side. But this will take some time. Again try to avoid snapshots by using quick backups instead and monitor the situation with our Veeam ONE (Snapshot lifetime alerts). You can use in 12.2 and 12.3 the above mentioned workarounds.