Large VM Backup loses speed until it is at 3-4mb/s

[GeneralCode]

I have a large VM (~5.5TB) that is on the local disks of a server running vSphere 5.1 and when I attempt to run a backup job it starts at around ~45mb/s and then after a number of hours slows to 11mb/s and eventually 3-4mb/s. This makes the VM impossible to back up since it will take too long to complete the first backup. I had previously been backing up this server successfully when it was about 2TB, but I managed to damaged the old chain and have not been able to get good backups since.

The backups are being saved to an old EqualLogic device, so I don't think that is the issue.

Any suggestions?

[Gostev]

What B&R version are you using? I would try running a test job to a different storage, even locally to the backup server, and see if you have the same issue there. Thanks!

[GeneralCode]

What B&R version are you using? I would try running a test job to a different storage, even locally to the backup server, and see if you have the same issue there. Thanks!

We are running 7.0.0.690 right now. I don't think I have another device with that much space free on it that I can backup to right now. At least not anything that will be anywhere near fast enough (I think I have an old Drobo Pro around somewhere that might have enough space, but I don't think that it will perform well enough to work). Also, we have other jobs that run to that storage device and work just fine. I have even tried shutting down everything on that VM server and all other backup jobs so there would be 0 load on the box and it still has the same thing happen.

[yizhar]

Hi.

Please provide more information such as:

Backup server (and proxy) configuration?
physical/virtual?
How is it accessing the Equallogic?

I've seen similar issue not related to backup with VMware VM stored on EQL - it was related to VAAI,
so if your backup repository server is virtual it might be related.

Veeam backup report and bottleneck?

Are you using reverse or forward incrementals?

And as much more related info.

Yizhar

[GeneralCode]

The backup server is a physical box with 6GB of memory and a Intel Xeon L5520 Processor. There is no proxy for it. The EqualLogic has a volume mounted as a drive letter that the backup server saves to. It is connected via a gigabit Ethernet card.

I am not sure how to check the Veeam backup report when the job hasn't finished.

I have tried reverse and forward incrementals and have the same issue with both.

The VM that I am trying to backup is very large as I said. It has 6 virtual disks, 1 200GB and the rest 1.9TB each. Currently it has about 5.7TB used.

[Gostev]

What is the OS version on your server? I remember there was infamous Windows issue that caused similar slow down. In a few other similar cases, the problem was with the backup storage. In fact, I remember it was Drobo at least once

If you are really interested to nail this down in a reasonable amount of time, I recommend that you open a support case and have support collect debug performance metrics from our data movers. Alternatively, we can spend the next few weeks guessing over the forum posts - I am fine with that approach too