Joan, > I have created a role with just select privilege on all sysadm objects. The > script run very fast at first time. Since there maybe new tables created in > the prod. So I cron the script, run it every night. Now the script ran so > slow, each grant statement at least took 2 to 3 seconds. So it never > finished to run in one day. (totally had more than 49000 tables and views). > The system is not loaded, most time is just this session is active. The > other function seems normal, users didn't complain for their activities. Do > you know any reason why the "grant select on syadm.xxxx to role" run so > slow? I had exactly this situation a while ago. It turned out to be Shared pool issue - Latching in the shared pool and library cache, and a steep spike in parsing. You might want to check STATSPACK or AWR (if 10g) during that period. In my case, the previous DBA had written a SQL script to generate GRANT statements on *ALL* objects, and I rewrote this to GRANT only on *New* objects, and this problem went away.. What was happening is this: From a DB shared pool perspective, these thousands of GRANT statements were non-bind SQLs and each needed to be parsed. This flood of non-bind SQL shredded the shared pool and the DB went crazy trying to juggle free space in the shared pool (latching, etc). At the risk of violating my BAAG agreement, I suspect this is also happening in your DB as well (Alex G: Please excuse my trespass :) -- John Kanagaraj <>< DB Soft Inc http://www.linkedin.com/in/johnkanagaraj http://jkanagaraj.wordpress.com (Sorry - not an Oracle blog!) ** The opinions and facts contained in this message are entirely mine and do not reflect those of my employer or customers ** -- //www.freelists.org/webpage/oracle-l