Veeam6 and Esxi5 backup error.

[btn003]

Hi -

I am running Veeam6 and 2 ESX 4.1 hosts and 1 new Esxi5 host.
Veeam are connected to vCenter5. Backups are stored on a Qnap Turbo NAS, shared folder for all hosts.

Backup of VMs on the 2 ESX 4.1 host are ok. No problems.
But backup of any VMs on ESXi5 host fail with error: The object has already been deleted or has not been completely created

I am noticing SSH connection error on the ESXi5 host:

Code: Select all

Failed to login to "myEsxiHost" by SSH, port 22, user "root", elevateToRoot False, autoSudo False
User authentication failed

I am able to connect to the ESXi host with the same credentials via a SSH client ( putty ), so I know the credentials are good and that SSH is running.

Any idea?

Kind r

Bjørn-Tore

[tsightler]

I'm not sure why it's trying to connect to the ESXi5 host via SSH. That shouldn't be happening. With ESXi all communications should be via the management agent. Very strange.

[btn003]

sorry, my bad writing.

I have no SSH errors, I just notice I cant connect to ESxi5 host myself from Veeam GUI.

Update:
If I remove Vcenter connection in Veeam and add the Esxi5 host as standalone host, backup works.
Confused.

[Gostev]

Well, you do have SSH error above, and as Tom correctly stated, there should not be any attempts for Veeam B&R to connect to ESXi 5 via SSH, unless there is some bug with certain sequences of actions. For example, this ESXi 5 may have been "fat" ESX 4.1 before (just like the other 2 hosts), and so it was remembered as such in the Veeam B&R configuration, along with its SSH credentials. So, even after the host upgrade, Veeam B&R may still be thinking it is the same fat ESX host - and is trying to communicate with it accordingly.

This would also explain why adding the server as standalone helps (when you do this, host type is correctly detected and stored in the configuration, and jobs are no longer trying to connect to the host via SSH).

[btn003]

ok I see.

Do you recommend a full format/reinstall of the host?

[Gostev]

I think just adding vCenter Server back will fix the issue, as all hosts will be scanned and stored correctly.

[btn003]

ok I tried that and it did not work. Actually I will reinstall Veeam server as we anyway need to get it on 64bit R2. Hope this solves the problem. Fresh clean install.

[Vitaliy S.]

Please open a support ticket if it doesn't help, though I would agree with Anton - any full re-scan should help.

[pkennett]

I'm seeing exactly the same issue. Migrated a 4.1 ESX server to ESXi 5 (after upgrading virtual Center). Upgraded Veeam to 6 and applied hf6. Replication jobs targeting the ESXi5 now fail. When I edit the job and choose the server or try to access the server under the Datacenter, I get prompted for SSH credentials as if it is still an full ESX box. I can SSH into the server using these credentials and add it as a standalone ESXi box to Veeam.

Problem with deleting and re-adding VC to Veeam is that it doesn't let you without killing all the jobs and replicas that already exist. Is there not any way to force the rescan?

[Sethbartlett]

When going from an ESX host to ESXi host, you will want to create a new job. If you are running 6, this is a painless, simple process where you can add all your VMs you need and check the "Replica seeding/low bandwidth" box at the beginning of the job and then map the replicas. The issue is that we communicated to an ESX host typically with agents and ESXi does not have the same capability as it is stripped down much more than ESX.