Re: 2 Buffer Cache Questions

RogBaker_at_gmail.com wrote:
> I have a script, which ecludes SYSTEM and SYS objects. A Co-worker is
> running a tool from the Enterprise Manager performance pack or
> something that is showing graphs of what's in the Buffer Cache, and
it
> was showing a lot of SYSTEM objects initally after the database is
> started. I suppose I could have edited the script I was using to
> include SYS and SYSTEM.
>
> I analyzed the scripts, and the bases for them are dba_objects.blocks
> and the number of rows returned from v$bh.
> Just ignoring the scripts and looking at dba_objects and v$bh, I get
29
> blocks total from v$bh, and 32 blocks from dba_objects.
> That gives me the same percentage of the scripts (91% for this
example
> table). Anyway, the scripts are below for your review.
>
> Whenever I use a script that I find on the internet, I try to verify
it
> with something I found somewhere else.

That's a good thing, but of course, myths abound, in books as well as the net. That's why so many people advise testing things for yourself to learn, as you are so excellently doing. http://asktom.oracle.com/pls/ask/f?p=4950:8:::::F4950_P8_DISPLAYID:2170006092044

> Here is the script I used, it's from
> http://www.pafumi.net/multi_buffers.htm
> --------------------------------------------
[snip]
> ---------------------------------------
>
> I found another version that had just about the same results from
> "everybody's favorite DBA:"
> http://www.dba-oracle.com/art_builder_buffer.htm
> ----------------------

[snip]

The first script is pretty useful. I can guess you were trying to use the second script to get to 100%, so I guess you were concerned about not being able to get there - since that script didn't have a pool identifier, after "explaining" the pools, it, um, muddied the water. Hopefully learning about the various pools plus Richard's answer has made it more clear.

jg

--
@home.com is bogus.
No one ever got fired for choosing IBM.
http://www.signonsandiego.com/uniontrib/20050505/news_1b5ibm.html