user for dbms_jobs / dbms_stats

I have been investigating creating dbms_jobs jobs , particularly for automatically running statistics using the dbms_stats package. I currently have one job representing one schema running; I was planning on have 5 separate dbms_jobs entries (one for each relevant schema), so they could run at different intervals. The database has 5 schemas, but one of the schemas is 100 times the size of the other four combined. The database / schemain question is in an 8.1.7.4 instance. Additional schemas are infrequently added to this database / instance.

Which user(s) would you recommend to:
a) create the dbms_jobs entry (to execute the dbms_stats.gather_new_schema.
b) create the dbms_stats.stats_table to store the statistics? c) is dbms_stats.gather_database_statistics preferred?

Presently, I am using the SYSTEM user to create the dbms_jobs entries.

I understand I will likely need to grant execute privs on the dbms_stats package for different users. Are there other gotchas?

We also have several 9iR2 instances, would you recommend a different answer for these database? (I've read enough to know that system stats should be gathered in 9iR2, but have not actually implemented them on our systems yet).

Future Thanks Received on Fri Jun 24 2005 - 18:07:16 CDT