From: "John Morrison" Newsgroups: comp.databases.oracle.server References: <95bml0$g0j$1@nnrp1.deja.com> Subject: Re: Will setting timed_statistics=TRUE create much overhead? Lines: 31 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.3018.1300 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300 Message-ID: NNTP-Posting-Host: 1cust152.tnt2.geelong.au.da.uu.net X-Trace: ozemail.com.au 981126580 210.84.253.152 (Sat, 03 Feb 2001 02:09:40 EST) NNTP-Posting-Date: Sat, 03 Feb 2001 02:09:40 EST Organization: OzEmail Ltd, Australia Distribution: world Date: Sat, 3 Feb 2001 02:02:40 +1100 The rule of thumb with versions 6 7 and 8 was to not run production systems with timed statistics, due to the approximately 5-15% overhead, which could be very significant if tuning a system that had serious performance problems. But that was when systems had very little excess capacity (due to the cost constraints). Now, I would simply say, benchmark before and after setting the parameter dynamicaly, and if it hurts performance turn it off when you have the tuning information you need. wrote in message news:95bml0$g0j$1@nnrp1.deja.com... > > > An Oracle DB monitoring product I'm using (Spotlight on Oracle) wants > to set timed_statistics=TRUE. The product docs claim this won't create > significant DB overhead (this is for an ORCL 8.0.5 database, which will > be upgraded to 8.1.6 eventually). > > What's the real answer? I know ANALYZE TABLE COMPUTE STATISTICS > statements can be a real drag on performance. What will happen if > timed_statistics is turned on? > > > Thanks very much. > > - Dana > > > Sent via Deja.com > http://www.deja.com/