Extremely slow Entire VM restore - 1MBPS!! - 00672054

[Dave-Departed]

Hi,

I just tried to restore an entire VM from last night's backup, and was astounded to see that I was only getting 1MBPS restore speed, which would've taken 8hrs or so to restore the VM (140GB in total)!!

I did some research, and noticed that I didn't have a VBR Proxy present on the host I was initially trying to restore to, so I thought this was the cause.

So I tried again, this time restoring to a host that DOES have a VBR Proxy present, and the restore speed was still 1MBPS.

I then tested copying a 5GB file from my VBR repository (the same one where the backup of the VM in question lives) to the VMware datastore that I was using as the restore target, and this copied in less than 5 minutes.

Anyone have any idea what could be causing this? Restores at my DR site are a LOT faster than this (and onsite backups to the repository mentioned above are also at least 50x faster).

Thanks in advance!

[Dave-Departed]

Also, a bit of further information - Performing a Windows FLR from the same repository to the same datastore clocks in at 20MBPS.

[Shestakov]

Hello Dave,
Could you check what transport mode is used for your VM restore? It should be [nbd] or [hotadd] and can be found in the job session log.
You can also try to use Instant VM Recovery and Migration jobs to move VM data back to production storage.
Thanks.

[Dave-Departed]

Hi Nikita,

It's using [hotadd], as expected, seeing as there is a proxy on the same host.

I am trying to restore a copy, so will IR work for that?

[Shestakov]

1MBps looks slow for hot-add restore.
Let`s see what the support team say after the logs review. You can also find some useful info in the similar topic.

I am trying to restore a copy, so will IR work for that?

Could you clarify it? Do you mean you are restoring a VM from the copy of the backup?
Instant VM Recovery restores VMs from backups.

[Dave-Departed]

I'm not sure what was going on, but I tried the restore again later, just out of interest, and it was fine. It was hitting 50-60mbps which is normal for our setup.

There were no other backup/restore jobs going on during either of the restores I tried above, and of course I checked the resources of both the main VBR server and proxies involved in the restore, and they were all well within tolerance. The iSCSI storage which acts as our main VBR repository is solely used for this purpose, and this purpose only, so nothing else would've been accessing it.

Very strange. Perhaps just a transient issue. I will close the case

[Shestakov]

Dave,
Did you do same things, but got different result?
I would recommend you to keep communicating with the support team to find out the reason of the issue.
Thanks.

[Dave-Departed]

Hi Nikita,

Yes, exactly the same priocess.

Ok, no problem, I will follow the call through to the end