Size 1.1TB (87GB used) but replica processed 500GB

[homoasapiens]

Hello, that might be a superdumb question. I'm very new in Veeam so I beg your pardon if it's already had been replied and you will send me to RTFM that link

I have VM with 1TB of disk space in thin provisioning, however only 87GB is used. So the actual size of VM is ~90GB. When I've started Veeam replica from 1 VMware ESXi server to another - it took a lot of time and before I've terminated that job(had a lack of time, thought it would be faster a little bit) - spotted that Veeam processed around 500 gigs. It's a little bit strange to me, because the real size of VM is 90 gigs.

Any suggestions? I would like to have this replica because it's a mail server and in case of issues - just want to sping the VM w/o long restore procedure.

Help appreciated!

Thanks!

Regards,
Alex!

[wishr]

Hello Alex,

Are you looking at the "processed" or "transferred" amount? The "processed" value does not necessarily indicate the amount of data that has been transferred.

Thanks

[foggy]

Hi Alex, please check if the job uses CBT (look for the [CBT] tag in the job session log right after the disk read information) since otherwise, it will scan the entire disk looking for changes to backup.

[homoasapiens]

Thank you for your replies.

In the Job settings, I can find that CBT is enabled.

But it's still quite funny because the actual size of VM is only 90G, but Veeam is processing(i.e. reading) 500G and was tended to do even more. Maybe for the first replica, it should process "full size" of VM?

I will try on a smaller one.

[homoasapiens]

Short update: 340G Thin provisioned VM with 2 disks - 40G(actual 2.5G) and 300G(actual 12. dumped in 25 mins. Transferred 14.5G

Seems like either there was a glitch, will try to dump that previous during the weekend.

[wishr]

Hi Alex,

Thanks you for the updates.

We never process empty blocks, however, if the blocks are "dirty" we will process them as usual - maybe this is the case with the other VM. In any way, the "transferred" value will indicate the amount of data passed to the target host.

Please keep us posted.

Regards,
Fedor

[homoasapiens]

Hello Fedor,

So finally found time to start over. And again, what I can see:



10.9GB transferred, processed ~400GB. That's odd, however, maybe it's because of the nature of application? It's a mailserver, so maybe while Veeam is backing up the data - the VM is constantly changing?

Regards,
Aleksejs!

[homoasapiens]

Finally, the replication was finished. But I think I found what's the issue with that VM!

I found out that it has snapshots and because of that CBT isn't available.

Regards,
Aleksejs!

[homoasapiens]

Update: After removing all snapshots on the source - VM replicated successfully in 3 hrs.

[wishr]

Hi Alex,

This is the expected behavior. If there are "non-veeam" snapshots, CBT is not possible - that's why we'll process all the data but transfer only the necessary changes.

Glad to see you've managed to resolve it.

Thank you