Re: db_block_lru_extended_statistics

From: anonymous <anonymous_at_anonymous.com>
Date: Sat, 26 Sep 1998 20:08:26 +0200
Message-ID: <01HW.B232FA3A000051830209A3A0@anonymous.com>

On Thu, 24 Sep 1998 15:39:19 +0200, Michael L. Mason wrote (in message <360A4B87.7957_at_osf1.gmu.edu>):
>
> Help!
>
> In an attempt to tune our database, I have enabled the above parameter and
> have looked at the x$kcbrbh table. According to my Database Tuning
> instructor the 'count' column on said table should have decreasing values.
> When I looked at the table, I saw values in the 'count' column vary greatly.
> In buffers 1-50 there were some buffers with anywhere from 1-30 'hits'.
> However, on buffer number 205 we had 1151 hits. Why would there be so many
> hits on this particular buffer. I was thinking that it might have something
> to do with the process that scans for free buffers. Because some buffers
> stay populated longer than others, these buffers may be skipped when
> searching for clean ones. Does this sound right?
>
> Thanks! Mike

I think you should have a look in the manual. The interpretation should be: when adding x blocks to the buffer, how many more hits do I get ?

Don't use this on a 7.3.4 production database; it causes a hang. Oracle removed the code from the oracle-8 because of the problems it caused.

--
Ronald

---

mailto:ronr_at_wxs.nl
http://home.wxs.nl/~ronr/RonR.html (last update: june 8th, 1998) http://www.maccentral.com/news/9804/migration.shtml (forward migration home) Received on Sat Sep 26 1998 - 13:08:26 CDT