In a previous question we discussed the pros and cons of relying exclusively on NetApp snapshots for backups. I now find myself in this situation as our aging legacy Symantec BackupExec tape server catastrophically failed while rebuilding its RAID-1 array. Something, something, something proactive upgrading of critical systems . . .
Our current backup strategy is limited to on-filer Snapshots through either SnapManager for SQL, NetApp Virtual Storage Console vSphere Plugin (SMVI) or hopefully manually configured Volume Snapshots. These are then SnapMirror-ed offsite to another filer far, far away in the frozen wastes of the tundra where it is guarded by snow wookies.
Due to space constraints we age these Snapshots out rather quickly and previously relied on tape for retention periods longer than a month and as discussed in my other question there are number of other significant shortcomings with reliance on Snapshots and SnapMirrors as a sole source of backup and recovery.
We have a Dell TL4000 LTO-6 tape drive in our offsite location that is currently being used by an existing BackupExec 2010 R3 server. As a stop-gap, I want to use it to write our SnapMirror Volumes to tape.
I have some basic questions:
- Can I use NDMP to write our SnapMirror Volumes to tape? Everything I have read says this is very common and simple to setup. I really wish we had NetBackup not BackupExec but "you go to war with the army you have, not the one you wish you had"
- What happens if the SnapMirror changes during the NDMP job? Do I need to coordinate the SnapMirror sync times so they do not change during the backup process?
- How can I test my restores? With agent-based flat files, it's pretty dead simple but I'm not sure what the procedure would look like for a Volume-based backup.
- Any other general advice and/or noticeable problems with my plan?