evaulating veeam - question re failed backup

[markwragge]

Hi, I am evaluating veeam as a repalcement for our current backup product. I am using veeam 4.1 in an esx 4.1 environment. I have setup a backup job containing 9 windows VMs. I have enabled VSS, de-dupe, compression and change block tracking.

The first backup ran and 8 clients were successful. One client failed as the backup disk ran out of space. The failed client had backed up 250gb at the time of the backup. Should this failed data be deleted from the vbk file on the backup disk. I do not think that the 250gb has been deleted so is using space on the backup disk.

Second question is that on the second night backup i expected much faster backup as the backup was not a full. THe 2nd backup file was created and is a vrb file of total size of 34gb. The backups of each client have taken a really long time.It appears that even though there was only 34gb backed up each client backup time was reduced by a max of 30% with several client backups taking longer than the full backup.

On the second night i removed the large client from the backup job. 3 of my 8 clients have warning message that change block tracking is not enabled as VM must be upgraded to version 7.

I am backing up to a data domain device. Is it ok to have de-dupe and compression enabled on the Veeam Job and then backup to the data domain which is also de-dupe.

[Vitaliy S.]

Mark,

You may consider deleting this file for the failed job by using Delete from disk option from the backup console. However, you next job run should remove the VBK file automatically.

That's correct to be able to use CBT you should upgrade all your VMs to hardware version 7. Please make sure you meet all the requirements for CBT to work:
http://www.veeam.com/forums/viewtopic.p ... lock#p8858

As for DataDomain question, please have a look at the F.A.Q. section for more information:
http://www.veeam.com/forums/viewtopic.p ... ain#p18665

Hope it helps!

[markwragge]

If i select delete from disk option it will delete the backup of all servers from the particular night as there was 9 servers in the backup job. Is there a way that the Veeam can delete the backup contents for one server within the job OR in my case delete the contents of one Failed server.

At the moment i have just two nights backups so i can delete the job and lose two nights but free up the disk space.

Otherswise i suppose i can wait 14 days until the retention period for the original backup expires and in this case Veeam would delete the failed server backup (this server is no longer a member of the backup group). Is this correct assumption?

[Vitaliy S.]

Mark,

Yes, that will delete the backup files from the datastore. If you don't want to lose your night backups, you may take existing files away to another storage and if there is a need to restore some VMs later, you will be able to do this by importing VBK file to the backup console.

There is no way to remove failed VMs from the backup file. You should have enough free space on the destination backup target to run the backup jobs, so right now you may want to choose another backup target or increase the free space on the existing one.

As regards retention policy, it is applied at the end of the each job run, thus meaning it is based on the job runs and not on the days that have passed.

Thank you!

[Gostev]

All VMs stay in VBK until the Full Backup is performed (per-VM retention is v5 feature only). v4 retention removes older restore points, but does not affect VBK content. You cannot remove individual VM from VBK either with v4 (again, this is v5 feature only). Right now, you have to remove VM from job and perform Full Backup, that will create new VBK chain with only actually included VMs.

[glaviolette]

There is no way to remove failed VMs from the backup file. You should have enough free space on the destination backup target to run the backup jobs, so right now you may want to choose another backup target or increase the free space on the existing one.

Similarly I'm just starting to evaluate this product..

So, the *only* way to remove any failed VM backup data from the .VBK file is to perform a full backup? Why doesn't it simply mark the data it just wrote as part of that failed job as "zeroed" data? Maybe I'm just being too simplistic..

Does it at least use the failed written data toward DD?

Does selecting "Perform Full Backup" for an existing backup job zero/delete all the .VBK files for the given job? I could probably run a test and find out, but I have your ear at the moment. Thanks!

Enjoying it so far!

One small "nit".. When viewing the Realtime Statistics for a backup job, if I select a previously backed up VM (Success, Failed, Warning..), when the data is updated on the "In Progress" VM, it changes my selection. I'm guessing you're aware of this, maybe just one of those "deferred" bugs?

[Vitaliy S.]

So, the *only* way to remove any failed VM backup data from the .VBK file is to perform a full backup?

Not really, you can pick up a VM and delete it manually, though please note this action will remove this VM from all backup set.

Here is an existing discussion reagarding deleting VMs from backups: Managing Backups and Replicas - Deleting

[glaviolette]

Not really, you can pick up a VM and delete it manually, though please note this action will remove this VM from all backup set.

Here is an existing discussion reagarding deleting VMs from backups: Managing Backups and Replicas - Deleting

Ok, so I went to Backups, selected the VM from the job that had failed previously (had "processed" 66.28GB of 126.28GB when it failed) AND that had a successful backup, I then deleted it. I noticed one of the .VIB files (373GB) was modified at the exact time of my VM deletion, so I'm assuming it was marking space as available? I went ahead and modified the existing job to exclude all other VMs EXCEPT the one I just deleted and then started it.. It immediately created a new.VIB and as of this post, the file is 56GB in size and it's only 47% complete? Shouldn't VBR have had sufficient space available in the existing .VIB file (that it modified/marked as clear space?) from the failed AND at full VM backup job that was deleted?

Am I misunderstanding something??

Thanks!

[Gostev]

Each new job pass will create new incremental backup (VIB) file. Reusing disk space that used to belong to deleted VM is a feature of reversed incremental backup mode (or incremental mode w/transform), and it happens for VBK file only. In other words, this only applies to job where existing VBK file is "reused" when it comes time to create new full backup.