Trying to stretch my wings a bit here, playing around with the job
scheudling system and with compiled procedures. Looking for some
'best practices' advice.
I've built some rather simple procedures to take care of some dba
chores like refreshing stats for the optimizer and collecting stats
into a 'stats_history' table for trend analysis, and I have these
running under the dbms_jobs system on a test database. In getting to
this point, I worked thru the problem of procedures not having user
privs that are inherited via a role. I'm also finding it a bit
difficult to keep up with scheduled jobs running under various
user-ids. As a result, I'm wondering if I wouldn't be better off just
setting up the 'procedures' as sql files and feed them into SQL-Plus
from a cron job. What problems would I be trading for if I went this
direction? Are there any generally accepted practices or is it really
pretty much all over the map?
Received on Tue May 06 2003 - 14:30:53 CDT
