From: Martin S. <ma...@li...> - 2006-03-15 18:13:18
|
>>>>> On Wed, 15 Mar 2006 08:48:48 -0500, Ryan Novosielski <nov...@um...> said: > > Some weeks ago I was going to ask the exact same thing, but what I > settled on was keeping job records as long as I'd need to keep the job, > file records about the same time (since my retention periods are short > -- 2 weeks of data is the SLA that I have), and volumes just long enough > so that they'll expire before I want them re-written. My assumption is > that this warning that you've pointed out is to keep you from having a > year of jobs with files saved for each one, because that could make for > a messy database. What I'm NOT sure of is why you'd want to keep jobs > any longer than the volume, or even why the functionality is there to > keep jobs and volumes for different lengths of time (for the record, > this exists in OmniBack too). I'm guessing it's so the jobs are there > for tapes that may have expired but have not yet been rewritten. I find it very useful because it allows tapes in several pools to contain different ages of backup from the same client/fileset. Weekly pool tapes are written once a week and recycled after 1 month. Monthly pool tapes are written once a month and recycled after 1 year. There is only one job, which changes pool according to the schedule. To allow restore from any of the 12 Monthly tapes, the job retention has to be 1 year. This is greater that the Weekly pool's retention. A similar thing happens if you have Differential daily jobs in a different pool that is recycled faster than the Full backup pool(s). __Martin |