I know stopping a running backup is roughly equivalent to this but as I said, it's rough. You get a lot of warnings and no assurance the backup stopped right (or will stop at all). What I'm asking for is a pause feature in which any currently-running VM backups complete before the backup pauses, and then of course a matching resume feature will also be required.
This would be really useful when engaged backup resources are needed for a one-off operation, for example, or when some part of the infrastructure needs to be brought down for repair, rebooted for updates or hangs, etc.
I reckon it's probably a lot simpler to implement this for backups using per-machine backup files so if it's a question of just implementing pause for that kind of backup versus waiting six years to get engineering time to implement on giant-blob backup files, maybe make it a per-machine option only
'If you truly love Veeam, then you should not let us do this ' --Gostev, in a particularly Blazing Saddles moment
You can certainly stop the backup job "right" and without any warnings by selecting the graceful stop option.
Pausing backup jobs is not a good idea, because it will leave the VM snapshots open and growing.
Oh I fully agree - once a snapshot has been taken the backup should complete. I'm more in mind of "let the N vms currently backing up complete but don't snapshot any more."
Also stopping a backup job gracefully or no generates an alert in our monitoring systems as it shows up as "Failed" in the backup list. Rather see them in state "Paused" so I can configure monitoring to only alert if the backup has been paused for longer than a specified amount of time.
'If you truly love Veeam, then you should not let us do this ' --Gostev, in a particularly Blazing Saddles moment
I see, perhaps we can have the dedicated Windows event log for the graceful stop, so you can filter one out. Although if a malicious insider or a hacker stops the jobs like that, and your monitoring systems do not generate alert as the result, then this is just asking for a trouble, no?
I believe it is only correct to show unprocessed VMs as Failed, as no backups were created - so the restore is not possible. Their backups are not just "paused", which assumes some sort of auto-resume: they fully rely on someone to a) remember they did this and b) actually restart them. That's a lot of reliance on the human factor, if you ask me and is why those monitoring systems are put in place in the first place.
This would be a great feature. My team regularly reports our weekly and monthly backup metrics to upper management, including backup job success rate, which is often negatively weighted whenever we need to reboot a system for patching or some other maintenance. We already have to manually restart the jobs after forcefully stopping them. But at least with a pause\resume feature, you wouldn't get the 'backup failure' and you wouldn't have wasted the amount of time that the job did process before pausing.
If you wanted to avoid having a bunch of paused jobs that admins forget to resume, you could have an option to choose the duration of the pause, and make it mandatory.
I previously used Commvault and there was a suspend option that did just this. It would allow the job to suspend and the resume when ready. This would be an awesome feature in Veeam instead of having to stop a back up job and then have it start over from scratch.
