retrogeek
Newcomer
I'm working on improving our DRM tape rotation, and I have gotten storage pool backups and checkouts fully automated and working more or less trouble-free. My next step is to automate checking in, but I'm concerned about the ever-present potential for human error.
Suppose the day's checkouts are done, and a set of tapes are changed to "vault" status and transferred to the I/O slots on the library. If the human responsible for removing those tapes and replacing them with any returns for that day fails to do so, what are the consequences if TSM later blindly checks in all tapes in the I/O slots with status=private? I take it I'll wind up with a tape that is now both a LIBVOL and still in "vault" state. Is there a way to detect this occurrence, recover somewhat automatically, and perhaps re-check-out these sorts of mistakes during the next checkout operation? Is there any risk of the system either losing track of the volumes or their status as DRM tapes or otherwise jeapordizing the data on the tapes? I don't have a test environment to see what actually happens, and I'd rather not just DO this to see what the results are when I'm working with real data.
Any advice welcome. I have other questions relating to this automation that I'll handle in separate threads.
Suppose the day's checkouts are done, and a set of tapes are changed to "vault" status and transferred to the I/O slots on the library. If the human responsible for removing those tapes and replacing them with any returns for that day fails to do so, what are the consequences if TSM later blindly checks in all tapes in the I/O slots with status=private? I take it I'll wind up with a tape that is now both a LIBVOL and still in "vault" state. Is there a way to detect this occurrence, recover somewhat automatically, and perhaps re-check-out these sorts of mistakes during the next checkout operation? Is there any risk of the system either losing track of the volumes or their status as DRM tapes or otherwise jeapordizing the data on the tapes? I don't have a test environment to see what actually happens, and I'd rather not just DO this to see what the results are when I'm working with real data.
Any advice welcome. I have other questions relating to this automation that I'll handle in separate threads.