rollback points not removed?

[AlexL]

I'm daily doing a backup of an entire host, using 10 rollback points, this host has 4 VM's, every day the backup completes succesfully. So far so good. The reasoning for backing up an entire host instead of selected VM's is that now I'm certain everything on the host gets in the daily backup without reconfiguring the jobs.

But... when I move a VM from this host to another host (which I'm not backuping) and move a new VM to the first host the rollback point of the moved VM is never deleted and keeps using diskspace. So the next day I have 4 VM's with 10 rollback points (the original VM's) and 1 rollback point for the new VM. When I do this "shuffle" a few more times, which is common in a HA/DRS enabled environment, things get out of hand, I've had 12 different VM's in the backup set. The only to get rid of them seems to be to delete the entire backup set and start over but then I lose my rollback points.

How do I resolve this situation?

[Gostev]

Hello,

In HA/DRS environment, I recommend backing up on cluster level, or using VM folders. Older rollback points (VRB) will be deleted automatically according to your retention policy. To remove old VMs from full backup (VBK) file without loosing old restore points, right-click the job and select Perform Full Backup.

Thanks!

[AlexL]

I've already tried doing a Full Backup, this did NOT delete the restore points. Hence my confusion and my reason to come here for help.

[Gostev]

Reading back through my previous response, Perform Full Backup cleans up full backup (VBK) file from VMs no longer in the jobs. Restore points (VRB) are removed automatically according to your set retention policy. VRB points are not removed immediately, because:
a) They contain data for more than one VM,
b) We cannot know if VM was deleted intentionally, or due to administrative error which you will want to correct later.

[AlexL]

Ok, I understand both reasons but not being very happy with it since this make VEEAM Backup less than ideal for our environment. Doing a cluster backup instead of single hosts would ease some of the pain but I'm afraid if things keep growing I'll won't be able to fit it in our backup window.

But still, intentionallly deleted VM's will forever stay in the backup file and the only way to get rid of them is to also remove all the restore points? This does not seem Enterprise ready.

I would like to ask if this "feature", the possibility to manually remove a deleted VM from a backup set, so be added in a future release.

[Gostev]

Ok, I understand both reasons but not being very happy with it since this make VEEAM Backup less than ideal for our environment. Doing a cluster backup instead of single hosts would ease some of the pain but I'm afraid if things keep growing I'll won't be able to fit it in our backup window.

We provide a very easy way to scale your backup deployment by deploying additional proxy servers and federating them with the Enterprise Manager. There is no such thing as not fitting backup window with Veeam

But still, intentionallly deleted VM's will forever stay in the backup file and the only way to get rid of them is to also remove all the restore points? This does not seem Enterprise ready.

The only way to remove them is to do Full Backup occasionally.This is how you would do this with any other backup product out there, you just do this implicitly, because they all use traditional backup requiring periodic (weekly) fulls. Each new weekly full will not contain intentionally deleted VMs. And you can do exactly the same with Veeam today, with the current version - by performing periodic full backups.

I would like to ask if this "feature", the possibility to manually remove a deleted VM from a backup set, so be added in a future release.

This is already in v5, see stickied pre-release FAQ

[AlexL]

Since we're using SAN mode and not VCB I don't immediately see how I would go about using multiple proxies?

I changed the rolback points to 2, after doing 2 FULL backups the VBK and VRB were indeed down in size and the phantom VM was also gone from the backup list.

Apperently the options and/or terminology used confused me, I had expected after 1 FULL backup to be rid of the phantom VM's.

Thanks for the quick replies, ultimately they did help me resolve my problem.

Still a happy VEEAM user and curious about how v5 will do when released.

[Gostev]

When I said "proxy", I really meant "Veeam Backup server" (nothing to deal with VCB). You can have any amount of Veeam Backup servers backing up different parts of your environment, no matter which processing mode you use.

[AlexL]

Unfortionately we're currently only able to use one server for backups.

[Gostev]

Have you considered using virtual backup servers and virtual applicance processing mode?

[AlexL]

We've got a EMC CX3 SAN connected with FC, no NFS capabilities. For space considerations I prefer to use 1 large LUN with one file system/volume/partition as the backup destination which is currently around 2.5TB. This setup excludes using multiple (virtual) backup servers since the backup volume can't be shared between multiple machines. I'd only be able to do multiple servers if I'd split this large LUN into multiple smaller LUNs or in multiple volumes and only assign 1 volume per backup server. But that way I'd loose flexibility, I'd really have to "guess" how much I'd be needing for a certain backup server and it's jobs. Resizing would be problematic and take a lot of time. Only when we make the entire switch over to VEAAM from our old traditional backup product I'd have 2 extra 2.5TB LUNs for a total of 3. This will probably take a few months.