AS 11.5 SP3 - Staging area housekeeping

[clowe1234]

I have been testing how to setup a D2D2T rotation and so far I have been sucessful. Here is what I have working...

A five day GFS rotation with M-Thur Incremental and Friday full with the last Friday of each month being the Monthly. We keep 4 weeks of incremental backups, 4 weeks worth of Friday full backups and 1 year of monthly backups. With the tapes and staging, this all works very well. The M-Thu backups all fit on 1 tape and the Fridays get their own tape. The media pools all rotate tapes in and out correctly. Except....

Here is the problem. The staging area does not "cleanup" sessions that are no longer valid in synch with the tapes. For instance, when the daily incremental or Friday full tapes gets overwritten after 4 weeks, it's staging sessions don't get removed. What I need is after 1 year, have the staging backups look exactly like the tape pools (last 28 days of incrementals and fulls and 12 end-of-month fulls).

I am 98% of the way there to make this work and to come up against this is really frustrating.

Any suggestions?

Chris Lowe
Datacenter Administrator
City of Boise, Idaho
(208) 395-7812

 

[Clowntrigger]

I am in the same situation. If you find a way to do this, please post it here! CA's forums are of no help and their tech support is useless.

I have a dedicated disk array setup for my staging. The threshold is set at 95%. I guess that means it will use up to 95% then stop.

 

[clowe1234]

I finally just gave up and have set the staging purge for 31 days (M-F for 4 weeks). the monthlies won't be there but only about 2% of our restores are from monthlies so it is not a big deal to get it from tape.

 

[Clowntrigger]

Did you setup Purge from the Utilities menu? I've set my Staging Policy to purge Full or Inc/Diff jobs after 3 days. It appears if a migration job fails, it won't purge the data.

 

[clowe1234]

Open up the backup job (modify), go to the Staging tab, click on the Policy button, select the Miscellaneous tab. There is a check box for "Purge Failed Sessions from disk immediately".

Although if you are talking about a successful backup from D2D but a failed D2T then by default the parent job should create a makeup migration D2T child job and put it on hold. That setting is also in the Miscellaneous tab.

 

[Clowntrigger]

I'm still having a problem with this. As best I can tell, completed/failed/incomplete jobs are not being purged. Any help is appreciated.

 

[clowe1234]

After much playing around here is what I have ended up with.

One GFS job that is staging to disk. The staging policy is set to purge failed jobs, begin the migration to tape 30 minutes after the backup is done, create a makeup job if the tape migration fails and keep all staging jobs around for 5 weeks before purging.

I have been testing this by setting the date on the server back about 3 months and then incrementing the time manually and watching how the job runs each "day". Everything appears to be working. I even tested the "failed" backup and it did purge the stage so I can't be of much help there.

One thing I can tell you is that when you create a staging policy with a purge delay (5 weeks in my case) the purge value seems to become part of every staging job after that. so if you originally set the purge for 5 weeks, then 2 weeks later set it to 3 weeks, those first 2 weeks will still not purge out for 5 weeks unless you do it manually.

Also, I did have one time when the test job would not do the staging. I ended up deleting the job and recreating it.