Version 8: ChannelError: TimedOut

[holbrocr]

I'm having several show stopping problems since standing up a new Version 8 server. One in particular is that I am unable to backup one of my larger VMs (1.1TB). Unfortunately, after several days of working with support (ticket #00712967), my frustrations are growing. Has anyone on the forum experienced the following issue and if so, what did you do to resolve?

Here is the scenario:

My environment consists of:
Backup server in Loveland, Colorado
Proxy VM (vmware) in Ames, IA
Backup Repository - CIFS share in Ames IA - assigned to Proxy VM - 1 gig connection
VM to be backed up - Fileserver with 3 disks (20 gig, 300 gig, and 850 gig)

Backup results:
Disk 1 (20 gig) completes successfully
Disk 2 (850 gig) reads about 64 gig (almost every time) before it fails
Disk 3 (300 gig) reads about 32 gig (again consistently) before it fails

The error then is: 12/23/2014 2:02:48 PM :: Error: ChannelError: TimedOut

I've installed a patch I received from support. Following the patch, the backups lasted *longer* and read more of the drive (was originally failing at only 20 gig or so before it failed). I've also limited the repository combined data rate to only 50MB/s.

I powered back on our Veeam 7 server and the job ran successfully. It's definitely an issue with this version of Veeam (along with several more issues that I have yet to open cases on)

[foggy]

Clarence, this looks like a connectivity issue, I wonder if your engineer can define the stage where the timeout occurs after reviewing the logs. Btw, where the host holding the VM in question is located? How do you populate the backup jobs - via direct host connection or via vCenter Server? What transport mode is used to retrieve VM data?

[holbrocr]

Foggy,

The VM that fails resides on the same host as the proxy. I have 13 other VMs that complete successfully in the same VMware cluster, but they are all 100GB or less in size. Veeam uses Hotadd to connect to the storage.

[holbrocr]

Sorry forgot to add...we populate the job through vCenter server. thanks

[veremin]

Just for testing purpose - can you add the same VMs through a host and try to backup them? Will it make any difference? Thanks.

[JWChen]

I had / have the same issue. Mostly with large virtual machines. The problem appeared after migrating from V7 to V8
But it wasn't the size of the virtual machine doing it. It was the fact that those machines have multiple disks of dissimilar sizes

If I let veeam backup these vm's with hot-add AND parallel processing, it will simultaneously start the small C: drive and the large D: drive (for example)
When it is finished with the C: drive, veeam gives vcenter the command to remove the snapshot. (I see this in the vsphere client). When the snapshot has been removed (takes a while depending on the size) the backup of the D: drive fails obviously.

After disabling parallel processing, all was well!

[foggy]

When it is finished with the C: drive, veeam gives vcenter the command to remove the snapshot. (I see this in the vsphere client).

This should not be the case, command to remove snapshot is triggered after all VM disks are processed. I would recommend to contact support to look at the logs.

[Gostev]

When it is finished with the C: drive, veeam gives vcenter the command to remove the snapshot. (I see this in the vsphere client). When the snapshot has been removed (takes a while depending on the size) the backup of the D: drive fails obviously.

I highly doubt this can be the case, as multiple disks NEVER finish processing at the exact same time due to different sizes and contents. Which means that the issue you are talking about would basically affect every single VM of every single customer using hot add (making it about 50% of our customer base). I am sure that if this was the case, we would hear about the issue by now a lot

[GBUMD]

So,

Did you ever get a resolution for this?

We just updated to version 8 U2, and have noted this same issue.

GB