A few weeks ago I reported that our ADSM server was suffering repeated 0C4
ABENDs while running 'backup stgpool' commands. IBM has established that the
ABENDs are caused by a bug first discovered on the AIX server, and documented
in APAR IY03390. The problem only occurs with 'wait=yes' specified on the
'backup stgpool' commands. IBM has suggested switching to 'wait=no' as a
circumvention. Unfortunately, the batch jobs we use to manage the ADSM server
depend heavily on the behavior produced by 'wait=yes'. The whole system will
fall apart if we change the storage pool backups to 'wait=no'. IBM does not
expect to provide a permanent fix in the next service release because that
release is already frozen, but does expect to provide a fix in the service
release after that. This latter release is not expected to ship until some
time in the year 2000. This does not strike me as a satisfactory response to
repeated server crashes.