Re: DBMS_STATS.GATHER_SYSTEM_STATS samples only BADSTATS

As far as I know, there is only one thing that actually results in BADSTATS, and that's an absence of any single block reads in the time period. (But there may be other oddities that could be version dependent - like multiblock read times being faster than single block read times).

What do you get from

    select * from sys.aux_stats$

Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/cbo_book/ind_book.html Cost Based Oracle: Fundamentals

http://www.jlcomp.demon.co.uk/appearances.html Public Appearances - schedule updated 29th Nov 2005

• Original Message ----- From: "Radoulov, Dimitre" <cichomitiko_at_gmail.com> To: <oracle-l_at_freelists.org> Sent: Wednesday, December 07, 2005 9:24 PM Subject: DBMS_STATS.GATHER_SYSTEM_STATS samples only BADSTATS

The same 9.2.0.4 instance on Solaris 8 (do you remember the "hash join vs nested loops" thread :))
Trying to gather system statistics in a period with common workload, CPU load 30-70%, only one instance on this node. I get only BADSTATS.

Metalink says that we have BADSTATS when:

=> collection is completed but not pertinent, because there was no
workload

      that could justify any statistics

=> start the collection again when the workload is relevant.

Tried with interval 5 min,15 min, 1h, 2h, 8h, and I get only BADSTATS. Tried auto(interval)and manual(start/stop) mode, always BADSTATS.

I'll try with an interval > 8h, but, meanwhile, any suggestion is welcome.

Regards,
Dimitre Radoulov

--
http://www.freelists.org/webpage/oracle-l