v7 - Backup copy jobs to offsite tape location

[kristofvm]

Hi,

For a customer, I have the following setup:
- DC 1 has 80 VMs with adequate IOPS available for VMware storage, serveral storage boxes that hold most critical servers. I have a virtual Veeam backup server here and 2 virtual proxies. Veeam backups are made here for quick restores on local repository. Replica job is run as well to replicate couple VMs from DC2 for DR. No tape livrary here as it is a hosted DC.
- DC2 has 10 VMs with a smaller VMware storage solution with limited IOPS. Veeam backups er made for quick restores of those VMs on local repository (SAS drives on physical backup server), also have 2 virtual proxies here for replica's. Replica jobs are defined here for DR of DC1. Currently, all VMs are replicated to this DC2, so they can be backed up to tape afterwards. This solution was chosen previously but is putting extra stress (IOPS) on the storage solution of DC2, also backup of replica's does not support CBT which makes this even worse. Tape rotation is done at DC2, basically for all VMs from DC1 & DC2.
- Ther is a 100Mbit WAN connection on MPLS in between.

With Veeam v7 version, I wanted to redesign to use copy backup job to copy backup restore points. Then, I only have to replicate VMs with short RTO in case of DR and restore the rest from backup to take some IOPS load of the storage in DC2. This works fine, the only problem I have with it is:
- retention policy in backup copy job works differently compared to retention in backup jobs, ie. with Copy backup job with restore points =5, the full VBK is always from 5 days old, the VBK is rotated... while in a backup job VBK file is static.

This mucks up my backup to tape idea as being able to do incrementals daily (currently still use SBE for GFS tape rotation, defined as always incremental) as this works fine with Veeam doing backup to disk with daily inrementals and weekly full. The backup copy job works differently though, which means my full VBK file changes daily and so will my backup to tape data will be a lot higher daily (unwanted behavior) as backup to tape window is limited.

Is there a way to get around this?
- can I set the copy backup job restore points to keep differently to keep full VBK static until the actual backup job does another full backup (maybe using GFS option??)
- how does the backup to tape functionality work? I have not been able to play with this yet as library still assigned for SBE. Does it look into the VBK, VIB files and can it do an incremental compared to what it transferred before?
- I could do backup to offsite first and then copy backup job in reverse direction (means double the data goes accross the line though)
- I could use the old approach defining 2 backup jobs (but means double snapshot per VM for backup, plus the replica snapshot - every night :s)
- I could use rsync or another third party tool but prefer to use Veeam
- any other ideas?

Cheers,

Kristof

[veremin]

Can I set the copy backup job restore points to keep differently to keep full VBK static until the actual backup job does another full backup (maybe using GFS option??)

No, it's not possible. Once the specified retention policy is exceeded, the full backup starts moving forward on regular basis.

How does the backup to tape functionality work? I have not been able to play with this yet as library still assigned for SBE. Does it look into the VBK, VIB files and can it do an incremental compared to what it transferred before?

Yes, backup to tape job works exactly how you've described, copying only the restore points that have not been copied to medias previously. However, please be aware, that with current implementation it's not possible to use backup copy job as a source for backup to tape job.

Any other ideas?

How often do you need to archive backup files to tapes? If it's possible to copy backup files on less frequent basis (say, weekly or so), the combination of backup copy job and file to tape one might be a valid option.

Thanks.