Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Huge Swap area being used despite having available ram (Solaris 8 , ora9.2.)

Re: Huge Swap area being used despite having available ram (Solaris 8 , ora9.2.)

From: <fitzjarrell_at_cox.net>
Date: 30 Mar 2005 08:02:49 -0800
Message-ID: <1112198569.836538.15890@z14g2000cwz.googlegroups.com>


Comments embedded.

Mike Rowland via DBMonster.com wrote:
> Hello.
> I am relatively new to Oracle as a DB, but we use Oracle 9i under the
> following conditions.
>
> A sun e6500 ten processor system is direct connected to a sunfire
V240 two
> processor system with 8 GB RAM.
>

Please define 'direct connected'.

> Here is my problem. The V240 has high IO wait times when large
numbers of
> processes are underway. The swap in use number goes through the roof
when
> oracle is running and I still have some 60+% of system memory listed
as
> free. I also cannot seem to increase the total SGA size beyond what
it
> presently is without crashing oracle
>

Apparently the V240 is running the Oracle database. Given that, please provide the adjusted kernel parameters found in the /etc/system file.

> My three questions are:
> 1) why is oracle using so much swap?

I can't tell you without seeing the kernel parameters you've set. My guess is you either haven't set the parameters or they're not set correctly.

> 2) is there a way to tell oracle on solaris not to use swap (or at
least
> curb its appetite for it)

Again, this usually reduces to incorrectly set kernel parameters.

> 3) how can I use more of the system ram (and hopefully decrease the
use of
> swap)
>

I've just told you, twice. Please provide the /etc/system contents and we can continue this discussion.

> I am attaching the sample output from startup and top commands for
context.
>
> As I have indicated, I am somewhat new to oracle, so forgive my
ingnorance
> of what might be obvious to my learned collegues.
>
> Mike Rowland
> 308-255-6694
>
> ====output from top on DB server=======
>
> last pid: 27538; load averages: 0.03, 0.05, 0.04
> 06:42:03
> 163 processes: 162 sleeping, 1 on cpu
> CPU states: 98.8% idle, 0.0% user, 1.1% kernel, 0.1% iowait, 0.0%
swap
> Memory: 8192M real, 5219M free, 3382M swap in use, 1624M swap free
>
> PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
> 1104 root 6 59 0 4464K 3680K sleep 382:02 0.09% picld
> 26596 root 1 59 0 2960K 1880K cpu/0 1:22 0.07% top
> 27376 oracle 1 59 0 3149M 1098M sleep 0:12 0.03% oracle
> 1791 root 1 59 0 28M 14M sleep 2:56 0.00% Xsun
> 1208 root 1 59 0 4928K 1944K sleep 1:01 0.00% skipd
> 27378 oracle 1 59 0 3149M 1097M sleep 0:09 0.00% oracle
> 1392 root 1 59 0 6216K 3648K sleep 0:08 0.00%
nsrexecd
> 1391 root 1 59 0 6008K 2880K sleep 0:03 0.00%
nsrexecd
> 27366 oracle 1 59 0 3094M 1081M sleep 0:02 0.00% oradism
> 1327 root 24 59 0 3616K 2920K sleep 0:02 0.00% nscd
> 1561 root 7 59 0 2672K 2232K sleep 0:02 0.00% mibiisa
> 1355 root 1 59 0 1056K 672K sleep 0:02 0.00% utmpd
> 27374 oracle 1 59 0 3150M 1103M sleep 0:01 0.00% oracle
> 1479 root 3 59 0 2744K 2088K sleep 0:01 0.00% vold
> 1249 root 1 59 0 2528K 1488K sleep 0:01 0.00% rpcbind
> =====================
>
> The startup of oracle gives the following output
> Total System Global Area 3222770856 bytes
> Fixed Size 734376 bytes
> Variable Size 2382364672 bytes
> Database Buffers 838860800 bytes
> redo buffers 811008 bytes
> database mounted
> Database opened
>
> ===============
>
> --
> Message posted via http://www.dbmonster.com

David Fitzjarrell Received on Wed Mar 30 2005 - 10:02:49 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US