RE: Maximum Db_cache_size?
- From: "Laimutis Nedzinskas" <Laimutis.Nedzinskas@xxxxxxxxxxxxx>
- To: <oracle-l@xxxxxxxxxxxxx>
- Date: Mon, 19 Jun 2006 10:54:39 -0000
Just follow up:
May be it is somewhat easier to think about the problem in terms of "db buffer
cache scalability".
Generally the scalability is not linear at the best.
Often the scalability graph has a turning point when adding more resources only
makes performance worse.
The question is what RAM value(or formula) is the turning point for Oracle
buffer cache, if any?
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On
Behalf Of Laimutis Nedzinskas
Sent: 19. júní 2006 10:44
To: oracle-l@xxxxxxxxxxxxx
Subject: Maximum Db_cache_size?
Is it good or bad to use all available (not used for other purposes, eg.
shared pool or pga) RAM for buffer cache?
I am talking about 32GB RAM range.
Oracle is not in-memory database but anyway - memory access is faster than disk
access.
However I understand that the way(algorithms) Oracle uses RAM may have a
practical turning point when adding more RAM will only slow down things.
Therefor the question is:
- is Oracle(9.2 version) better at utilizing RAM(say, 10-20GB) for buffer cache?
- or is it it is better to let file system to utilize this RAM for file cache?
- none of the above, stay with moderate RAM usage (few gigabytes for buffer
cache) just because buffer cache hit ration is good(98-99%)? I would just like
to point out that 1% of disk ("raw") access makes up a considerable response
time, may be 50% or so.
Thank you in advance,
Laimis N.
Fyrirvari/Disclaimer
http://www.landsbanki.is/disclaimer
--
//www.freelists.org/webpage/oracle-l
Fyrirvari/Disclaimer
http://www.landsbanki.is/disclaimer
--
//www.freelists.org/webpage/oracle-l
Other related posts:
