VHDX resize needs active full to run

[willrussell]

Hi,

Ref case 02383786

Just wanted to check something regarding https://www.veeam.com/kb2450

The above article was pointed out to me during our case, but I've never come across it before. In this case a VHDX resize had silently caused a client's backups to keep reporting 'success', but unfortunately the disk that was resized (a 64 TB vhdx containing the primary company file data) was no longer being successfully backed up (technically the disk was being backed up, but it was empty). The weekly synthetic fulls suddenly dropped down from 30TB to 1TB in size, but of course no-one was looking at raw files to realise this was an issue - they were being reviewed based on the job log.

Unfortunately this has meant almost a month of completely lost file backups.

I'm surprised I haven't come across this limitation before. Is there some sort of improvement that I can suggest which will means Veeam can see if a disks geometry has changed between backup runs, and force an Active Full?

Will

[Mildur]

Hi Will

There could be a limitation in the vss snapshot process.
Vss can only snapshot volumes until 64TB. Is it possible that this disk was excluded from the backup because of this limitation?

[willrussell]

The job log hasn't thrown up any errors regarding the snapshot - so I assume it worked OK? (Application aware processing is enabled)

Also the VHDX itself is 64TB (the maximum) so the volume inside the VHDX is marginally smaller than 64TB.

I've run an Active Full now (without changing disk size again), which will complete tomorrow, so I'll see if that works or not.

Will

[Mildur]

Yeah, then the limitation does not apply in your case with different volumes inside the vmdk.
Let us wait for the results of the active full.

[isusko]

Hi @willruseell,

any updates on this topic?

Even if I was aware of the geometry change of the disk when a resize takes place, I never thought about this possible issue so thanks for posting it on the forum.
Surely would be nice for an alert when the resize happens, in order to make an active full backup.

Isusko

[willrussell]

Hi there,

Yes, the Active Full backup completed yesterday daytime (24h job) and the missing data is now in there. The .vbk is now ~30TB instead of ~1TB. The one incremental that has run is also fine.

I will now just be very careful to remember in case I ever do a vhdx resize! I have done it plenty of times in the past without issue - the only thing that makes this one different is that the vhdx is the maximum possible size of 64TB. Unfortunately I don't have any more hardware to create a 64TB vhdx on for testing

Will