Backup Job still running despite reboot?

[aeccles]

I'm currently backing up up a 650gb Exchange server over a 10mbps WAN (VPN) link to my data center. This is the first run of the job, so it's huge... been running for about 3 days so far. Due to an oversight on my part, this server had updates installed and was rebooted last night. I of course cried a little bit when I realized this, but then became a bit confused when I checked the Veeam console and saw that the job was still running... at first I thought it was just hung but it is actually still progressing.

So, Two questions.

1) What makes this possible? I've had reboots occur while jobs were running before and the jobs ended (failed.)
2) Will this backup be reliable, especially since updates were installed in the middle of it? I'm guessing/hoping that VSS will only transfer the data as it was when the job started?

Thanks

[foggy]

1) What makes this possible? I've had reboots occur while jobs were running before and the jobs ended (failed.)

Veeam B&R is an image-based solution that leverages VM snapshots to backup VMs. Basically, at the moment when backup starts, VM snapshot is created and VM data is retrieved from this snapshot. After the backup snapshot is committed to the parent virtual disk.

2) Will this backup be reliable, especially since updates were installed in the middle of it? I'm guessing/hoping that VSS will only transfer the data as it was when the job started?

The backup will be reliable, however, it should finish first, which in case of Exchange running from snapshot for several days could take forever. Have you considered to use backup seeding option for remote backup scenario?

[aeccles]

Foggy,

Thanks for the quick response. Great that the job can continue through a reboot!

I know it's not 100% accurate, but at the current processing rate (405gb of 650gb processed over 58 hours... roughly 7gb/hour) I was expecting the job to wrap up in the next 40 hours or so. I foolishly had not considered the VM snapshot commit time/requirement. I just looked at the "Storage Views" in the vSphere Client for my Exchange server and I'm seeing a 33.5gb snapshot. I understand this will probably double since the job is only about 60% complete.

I read through the link you provided and I think I understand, but I would like to try and clarify it... Is it correct that the snapshot will take several hours (depending on my storage) to commit, but the only disruption to users/exchange will be during the few seconds when the commit completes? As long as Exchange isn't down for more than a minute I am OK with that.

[foggy]

Could be more than a minute (there are several examples in the referred topic). Depends on the storage IO performance and the VM load mostly.

[trackstar]

Hello,

Just curious if a machine reboot occurs (patch, etc) during a backup, is there any issue? I am guessing application aware might be if occurs during a command to truncate the logs, etc. Also what if the machine reboots during a snapshot? Been using Veeam almost 4 years and never had issue but just came up on my mind today.

[cby]

We've been using Veeam since version 2 and backing up Citrix XenApp servers which, for historical reasons, are rebooted every night to address memory leaks. Over the years we have never encountered an issue with backups that were not sorted on first or second retry when we suspected that the reboot caused the initial failure. Not a technical explanation but an empirical one

Agree that application-aware may be a problem, but again the retry should address this once the server is back up.

[foggy]

Proper reboots do not affect the backup process, since VM is running on the snapshot. There might be issues in case the VM is forcibly shut off at the moment app-aware processing is working, without letting the running processes to complete (but still the backup will be performed, though will be crash-consistent), but this is not the case with patch application.

[mkirkpatrick]

Once Veeam has created a snapshot (vCenter) of a VM for backup, is it OK to reboot the that VM? I assume that since Veeam is backing up the snapshot it is no longer accessing the VM. Thanks.

[Vitaliy S.]

Yes, it's safe.

[Mgamerz]

What about if it is doing guest processing or indexing? I thought that had to inject something into the VM.

[Vitaliy S.]

VM has to be running to execute indexing and run preparation tasks for application-aware processing.