On Thu, 22 Sep 1994 08:25:25 EDT, Hans Kilian wrote:

>We are experiencing a problem with ADSM when our database runs out of free
>space.
>
>When this happens and a migration starts from one storage pool to another
>(in our case from a storage pool on disk to a storage pool on tape), ADSM
>asks for a scratch tape on which it'll put the data from the disk storage
>pool.
>
>It then discovers that there's no room in the database for the activity
>log entries, and it stops the migration process and it dismounts the
>tape. After 60 seconds, the migration process starts again, and ADSM asks for
>a new scratch tape. The problem goes away when we add some space to the ADSM
>database. But while this process is going on, ADSM uses every scratch tape
>in the stackers on our tape units.
>
>We do not have a tape robot and we don't have operations people on work
>at night, so we've programmed our automation software to mark ADSM's
>tapes as read-only as soon as ADSM dismounts it's tapes. This means that
>ADSM will always ask for a scratch tape when it needs one for its
>migration to tape.
>
>We are running ADSM under MVS at level 0.4/1.4.
>
>Have any of you experienced this problem or do you know what to do about
>it (other than to increase our database space by 500 MB :-) )?
>
We had something similar, on CMS at level 0.8/1.8.  In our case, it flooded
the console with messages.  It was trying to do a reclamation.  It mounted
the two tapes involved, then found that it couldn't write the log and
terminated.  When it restarted the process, it forgot that the tapes were
mounted and tried to mount them again.  The operators tried cancelling the
mounts, but that didn't help either.

*------------------------+---------------------------*
| Lee Maschmeyer         | lmaschm AT cms.cc.wayne DOT edu  |
| 103 Computing Center   | (313)577-0741             |
| Wayne State University | Fax: (313)577-1338        |
| Detroit, Mi 48202      | Home: (810)545-0170       |
*------------------------+---------------------------*