Offsite reclamation question

[gernthefish]

We've got a primary tape stg pool with 2 copy pools (LTO and FILE). Each week, we fully reclaim the LTO copy pool and send the tapes offsite. Is there a way to reclaim it using the primary stg pool instead of using the FILE copy? Thanks!

 

[marclant]

When doing offsite reclamation, the TSM server is using one of the online copies. Typically that's the primary pool, but the TSM Server will pick the one that it thinks makes most sense. If your primary is tape and then your online file copy pool makes more sense, because it's not using tape mounts. It's no different then trying to do a restore and both the primary and copy pools are available, it will pick what's best.

 

[marclant]

I guess you could write a script, put the file copy pool volumes unavailable, run the offsite reclamation, then put them read/write.

 

[gernthefish]

Great suggestion! Could probably update the volumes or the whole stg pool. Thanks!

 

[marclant]

Just curious, what's the problem with reading the data from one of the copy pools, as long as it's online and the process completes successfully?

 

[gernthefish]

Well, we currently have an offsite tape copy pool that gets fully reclaimed and sent offsite each week for D/R. We just added a 2nd copy pool consisting of Amazon Stg Gateway volumes (devtype=FILE) for testing, with the eventual goal of replacing the tape process. So both copy pools will be running as we test. The problem is (surprise surprise) we have been dealing with some performance issues with the Amazon pool, and the weekly reclaim process decided to use this pool to reclaim the offsite tapes instead of using the tape pool, causing the process to run forever. Your suggestion to set access=unavailable on the Amazon pool sounds like it would work.

Any further comments, suggestions on this would be greatly appreciated. Thanks!

 

[moon-buddy]

Well, we currently have an offsite tape copy pool that gets fully reclaimed and sent offsite each week for D/R. We just added a 2nd copy pool consisting of Amazon Stg Gateway volumes (devtype=FILE) for testing, with the eventual goal of replacing the tape process. So both copy pools will be running as we test. The problem is (surprise surprise) we have been dealing with some performance issues with the Amazon pool, and the weekly reclaim process decided to use this pool to reclaim the offsite tapes instead of using the tape pool, causing the process to run forever. Your suggestion to set access=unavailable on the Amazon pool sounds like it would work.

Any further comments, suggestions on this would be greatly appreciated. Thanks!

The last word I heard from IBM is that cloud options are not really ready for prime time. There are issues with reads which tend to be slow. This is a at 7.1.6 and is a show-stopper for them.

In light of this, we stopped all POC with cloud storage.