Forum Discussion

mjd42970's avatar
mjd42970
Level 3
9 years ago

Best practice offsite

We have Netbackup 7.6 running on a Windows server. The server environtment being backed up is mixed with RedHat, Windows and Solaris servers. We have seperate policy/volume pools for the onsite tape ...
  • Genericus's avatar
    9 years ago

    I separate pools by retention, and I have a separate catalog pool, but I think that is really only for clarity.

    As long as you are set not to mix retentions on media, even though they may be in the same pool, they will not mix.

    Depending on your DR plans - you may want to ensure you can easily find your catalog tape. I actually use a separate volume range so I can find it with no issues.

    The only other reason I seperated some of my tapes, is that I have limited drives in DR and I cannot afford to have restore #1 using tapes required for restore #2. You need to review your processes - if you have simultaneous restores planned you need to make sure they are on separate tapes.

    I am currently stuck on VTL using FC, or physical tape for DR, I cannot wait to get on disk based DR recovery.

  • Nicolai's avatar
    9 years ago

    No problem in using one volume pool for offsite backup.

    I have been mixing Windows, Oracle, SAP, UNIX in the same volume pool for 15 years.

    Volume pool and retention are to been seen as "costly" resources when using tape. Cut down on both and you will get better efficiency.

  • jim_dalton's avatar
    9 years ago

    Heres a problem you might encounter: tape blocking on invoking DR. Say youve got 20 servers data all on one tape, when you come to restore at DR you can only drive one restore job at a time, so if youve got one restore that takes 3 hrs, you've got yourself a bottleneck.

    This is when having multiple pools works as it breaks up the resource bottleneck at DR, hence we have volume pools for oracle, for sap, for flat file, for ndmp, for exchange and so on. This means you can restore each of these in parallel if need be assuming you have the drives and the human resources available. Add your RTOs into the mix and it will determine your requirements.

    Jim