On Fri, 1 Oct 1993 14:20:33 EDT, Paul Zarnowski <VKM AT CORNELLC.CIT.CORNELL
DOT EDU>
wrote:
>So, it sounds like if you want reclaimed tapes to be re-used, then they must
>be assigned to (owned by) DSSERV. When a VMTAPE SCRATCH tape is released by
>WDSF, it cannot be re-allocated by VMTAPE until it has been expired. I see
>no way to automate this expiration, as I see no way to tell VMTAPE that WDSF
>is through with the SCRATCH tape.
We are running with ADSM and VMTAPE as well. One solution is to have a
SCIF monitor the ADSM servers console messages and when it sees
something that looks like the following:
ANR1341I Scratch volume V13459 has been deleted from storage pool
BACKUP_TAPES.
ANR5300I Verifying volume label V13459 on device 0B21.
ANR5220I Deleting CARTRIDGE V13459 from Storage Pool BACKUP_TAPES.
have it do a VMTAPE CATALOG volser RETPD 0. Appropriate authorizations
in place as well of course. This seems like a big hammer for such a
little problem. Would be nice if the mount exit were called again to
do this.
>
>We pre-allocate tapes to the WDSF server (using the DEFINE VOLUME command)
>and try not to make use of VMTAPE SCRATCH tapes at all. However, if WDSF
We only use SCRATCH tapes with a 1000 tape limit for both BACKUP and
ARCHIVE tape pools. I don't expect to reach that any time soon. ;-)
The idea is that if the server needs a new tape there is likely someone
performing a backup or archive. We want this to get done as quickly as
possible.
>ran out of pre-allocated tapes, and started using SCRATCH tapes, we would
>end up with the same problem you have identified. It would be a pain to find
>the list of tapes that WDSF used and released, and then expire them to VMTAPE.
>It would be nice if there were a WDSF "tape release exit" that would be called
>whenever WDSF deleted the tape from its repository configuration.
Yes it sure would.
Have fun,
Carl Forde phonenet: 604-389-3234
VM Systems Software Group bitnet : CFORDE@BCSC02
British Columbia Systems Corporation internet: cforde AT bcsc02.gov.bc DOT ca
|