Backup Replicate vCenter VM?

[james411]

Hi all,
I've just tried backing up my vCenter server/Veeam server for the first time and it failed with the noted error: Unable to release guest. Error: VSSControl: Failed to freeze guest, wait timeout

I understand this was typical and I have read the KB article about adding the VM through the standalone host method, but I thought this was no longer needed with 6.5 Patch 3.

Is it still required to add the vCenter server through a standalone host or disable VSS processing on the VM?

Or have I misconfigured my job perhaps?

Thanks!

[veremin]

You have been merged to the existing discussion regarding backup and replication of vCenter VM, so, kindly see the answers provided above. Should any additional help be needed, don’t hesitate to let us know.

Thanks.

[Vitaliy S.]

Is it still required to add the vCenter server through a standalone host or disable VSS processing on the VM?

Yes to both.

[james411]

Thanks for the info. I looked thru this thread but for some reason I thought this was resolved with the latest release. Regardless, I think I will just disable VSS as vCenter/Veeam are the only services on this VM and they are each using a separate SQL Server instance. I have a small shop so there is not a lot of changes going on. I also have Veeam backing up its configuration daily so I figure in worst case scenario I would just fire up a new VM, install Veeam, and restore the backup config. I'd have to reinstall and reconfigure vCenter, but that shouldn't be too bad seeing as I only have 2 hosts. Thanks for the input.

[albertwt]

Hello Jim, yes this is "the fact of life" right now, if you are using Veeam VSS.

So is this still the case for backing up Windows Virtual VMware VCenter server 5.1 and above with Veeam Backup & Replication 8.0.0.917 ?
the VCenter DB is in the same VM with the VCenter.

[veremin]

You don't have to backup vCenter VM via standalone connection any longer; take a look at this post. Thanks.

[albertwt]

You don't have to backup vCenter VM via standalone connection any longer; take a look at this post. Thanks.

Ah I see, So I'll treat the Windows VMware VCenter VM as the other normal VM with SQL Server running in it.

is that correct ?

[veremin]

Correct. Enable AAIP, the VC database will be skipped automatically, thanks be to the special logic. Thanks.

[albertwt]

Correct. Enable AAIP, the VC database will be skipped automatically, thanks be to the special logic. Thanks.

So if the SQL Server DB in the same VM will be skipped, then the restoration wouldn't be complete ?

[foggy]

It is not skipped, but rather excluded from VSS processing and is backed up crash-consistent.

[albertwt]

It is not skipped, but rather excluded from VSS processing and is backed up crash-consistent.

Thanks.

[albertwt]

You don't have to backup vCenter VM via standalone connection any longer; take a look at this post. Thanks.

Hi Veeam Experts,

Somehow I got confused myself here after logging Case # 00875094, the support engineer says that I must backup the VCenter VM through stand alone ESXi host ?

I can understand from this blog post created in 2011: : http://up2v.nl/2011/05/06/backup-of-vce ... plication/

but is that still the case in Veeam Backup & Replication 8.0.0.917 (Patch 1) ?

[Shestakov]

Albert,
As has been mentioned above, in v8 vCenter database is excluded from VSS processing automatically.
So it is not the case.
Thanks.

[albertwt]

Albert,
As has been mentioned above, in v8 vCenter database is excluded from VSS processing automatically.
So it is not the case.
Thanks.

Thanks Nikita,
But somehow the support engineer said otherwise ?

I believe that a great software such as Veeam should intelligently able to backup VMware VCenter with no issue.

[veremin]

But somehow the support engineer said otherwise?

Most likely, he's recommended the obsolete practice that is not required any longer.

[brichardson1991_obk]

Hi All,

Just a quick question.

Would there be any problems with replicating a vCentre?

[PTide]

Hi,

Your post has been merged with an existing topic, please see the thread. If any questions persist feel free to ask.

Thank you.

[Messenjah]

Hi,

I have no problems with backing up the vCenter but with replication. I wanted to replicate the vcenter to an offsite target but now we have database inconsistance and the vcenter service can´t be started. So I want to refresh the question: Is it normaly possible to replicate the vcenter without having trouble?

Best regards, Piet

[PTide]

Not sure that I follow.

Did the inconsistence occur before replication or after? Which server is in inconcistent state - replica or source? Also please note that Veeam automatically excludes vCenter from VSS processing so it gets backed up in a crash-consistent state.

Thank you.

[widmerkarl]

If the backup is consistent, then i would recommend to replicate from backup instead of the live VM, in case you are replicating from the live VM. I don't think that it's a big issue when the vCenter database is a litte out of date (difference between backup and replication).

[Messenjah]

Thanks for the fast reply. Maybe I try to explain a little more:

We have a 2 host cluster with vSphere 6.0 onsite where also the physical backupserver is located. All is connected via 10 Gb Ethernet and we use transportmode san for the backups. Additional there is a third host offsite in another building connected over a 1Gb connection. The vCenter is a virtual maschine which is located within the cluster on one of the two hosts.

We configured a normal backup job to backup all machines which are part of the cluster. That runs very good without any problems.
After that we wanted to replicate some maschines to the other building. That works very good too, if we don´t replicate the vcenter. (I will try to replicate from backup, that seems to be a very goog suggestion. Thank you @kawiMTF)

We had it two times (as we tried to replicate the vCetner) that in that moment the replication job starts (a vmware snapshot created) the eventlog from the live VM lists an error something like that: http://kb.vmware.com/selfservice/micros ... Id=2006849

The job has already replicated the vcenter but some other maschine in the job are sill running. Then the vcenter service crashes ( maybe a result from creating snapshot) and the replication job fails. I cannot start the vcenter service anymore. I need to restore the whole machine. But I can´t use the replicated maschine, because the failure is there already also present. I need to go back 1 restorepoint from my backup which runs without problems.

Now the question: Is it possible to replicate every kind of maschine regardless if it is an SQL DB or the vcenter or something else without getting the data currupted or is that not the recommended way?

[PTide]

Is it possible to replicate every kind of maschine regardless if it is an SQL DB or the vcenter

It is, however some difficulties may occur when trying to perform transaction-consistent backup of linux VM or vCenter, but no corruption of a source VM should normally occur.

So, you say that backups are good but replica is not, correct? Kindly try to exclude the vCenter database manually and see if that helps. If the replication will keep failing and breaking your source VM then please open a case with our support team and post your case ID here.

Btw, which version of VBR are you using?

Thank you

[Messenjah]

Yes, correct. The strange thing is, that in both cases the backups are made from vmware snapshot, which gets created for the job.

I tried to replicate from backup with success. Maybe for now this is the best way.

Honestly I am not sure what a vCenter without a proper DB is worth. I need the vCenter on the replication target for disaster cases so I can start the vCenter without any hands-on. Or is there another mechanism that the replicated vcenter maschine with excluded database is able to manage the environment, at least transitional?

We use B&R 8.0.0.2030

[Messenjah]

Now the problem came up again. But now I excluded the whole vCenter from replication. I wanted to replicate 5 VMs. The replication job worked as it has to. Now I wanted to check if CBT and everyhting works fine if I start the job again.
Then I get this on Veeam:


In the same moment the vCenter Service crashes with following event viewer entry:


I wasn´t sure if the problem comes up only if I want to replicate the vCenter VM. But now its clear that it doesn´t matter. This is the third time within 2 days of testphase....

[PTide]

From the error it seems like there is an issue with ESXi. HTTP web service processing API requests is down. ESX host reboot should help. Also please open a case with support and post your case ID here.

Thank you.

[Messenjah]

No veeam has lost the connection because of the vcenter service crash. And the service crashed because of something happend with the db.

The ESX Hosts are up and running. It´s only the vcenter which has problems. I have restored the whole vCenter from backup again. Know everything is green. Until I perform the next replication jobs...

The caseid is: 01100942

[PTide]

Sorry for asking that many questions, but what's the version of the target host? Have you installed any updates recently?

[Messenjah]

Nothing to be sorry about, I am thankfull for the support.

All ESX hosts has Version 6.0.0. Build 2494585. I havn´t installed any update, but maybe I should consider it.