Re: Statspack

not replying to syltrem , didn't see the original post

> "Mill" <mill.h_at_ntlworld.com> a écrit dans le message de
> news:9mPjc.305$fp.243_at_newsfe6-gui.server.ntli.net...
> > Have just run statspack on a oracle 8.1.7.4 DB on windows 2k with 2gb
RAM,
> > can anybody see
> > anything in the results i need to be concerned about ?
> >
> > Appreciate any comments/tips

I'll add my voice to the list that recommend oraperf.com

> > Elapsed: 63.92 (mins)

You are doing a lot of waits of indexed lookups.

> > Buffer Gets Executions Gets per Exec % Total Hash Value
> > --------------- ------------ -------------- ------- ------------
> > 8,513,149 80 106,414.4 42.7 3942090408
> > DECLARE job BINARY_INTEGER := :job; next_date DATE := :mydate;
> > broken BOOLEAN := FALSE; BEGIN Transaction_SYS.Process_All_Pendi
> > ng__(2,'Heavy Queue','en',job); :mydate := next_date; IF broken
> > THEN :b := 1; ELSE :b := 0; END IF; END;

Ok you have run this , presumably since it calls dbms_job async, process 80 times in 64 minutes or every 48 seconds on average. Does that make sense? Would every 5,10 or 15 minutes make more sense.

> > 8,501,059 1 8,501,059.0 42.6 1773115984
> > BEGIN Mrp_Process_API.Do_Mrp_Process__(:attr_); END;

This however only got called once but was responsible for 8.5m LIOs, again is this reasonable?

-- 
Niall Litchfield
Oracle DBA
Audit Commission UK
http://www.niall.litchfield.dial.pipex.com/