With regard to the 4031 error (and assuming this is some version of
8i), does this only occur during the rman backups? If so, check your
init.ora for the backup_tape_io_slaves setting. If this setting is
set to "true", then your rman processes could be using large amounts
of your shared_pool during the backups (if your large_pool is not
configured). Either set the large_pool, so your backup slaves use it,
or set backup_tape_io_slaves to false (then memory comes from the PGA,
not the shared_pool).
Received on Tue Mar 09 2004 - 14:00:13 CST
