What platform are you running on? The memory numbers reported by 'ps'
don't really provide a good picture of the real memory usage of a
process. This is especially true on Linux, since modern versions of
the kernel allow memory to be overcommitted by default.

--James

--- In psftdba@yahoogroups .com, SuRaj Sharma <suraj_sharma1981@ ...> wrote:
>
> Hi,
>
> We are experiencing  high memory and swap utilization on our middle
tiers. When I searched for top 20 processes consuming high Swap from
OS level all processes I found were something like this:
>
> USER       PID %CPU %MEM   VSZ  RSS TTY      STAT START   TIME COMMAND
> pspabcdv 23240  0.0  1.6 1527900 266572 ?    Sl    2008   1:35
PSAPPSRV -C dom=PPABCV12_ 253088 -g 99 -i 7 -u aupodabcd14 -U
/pabcdiv/psoft/ pt849/epm90/ appserv/PABCDIV1 2/LOGS/TUXLOG -m 0
-s@../psappsrv. lst-s@/psqcksrv. lst-sICQuery -sSqlQuery:SqlReque st --
-C psappsrv.cfg -D PABCDIV12 -S PSAPPSRV
>
>  
> When I drilled down further into this from the database level I fond
that all these processes are INACTIVE at the database level and logged
in since very long. The output from v$session view was something like
this:
>  
>        SID    SERIAL# STATUS   login Since          LAST_CALL_ET MACHINE
> ---------- ---------- -------- ------------ -------- ------------
------------ --------
>       1648      13558 INACTIVE 22-dec-2008 01:19 AM        21796
aupoduabcd14
>       1662       9004 INACTIVE 02-jan-2009 02:27 PM        36548
aupodabcd14
>       1666      40727 INACTIVE 05-jan-2009 07:33 AM       111971
aupodabcd14
>       1676      34961 INACTIVE 05-jan-2009 07:40 AM       468687
aupoduabcd14
>  
> Please let me know how to sort this issue out.
>  
>  
> Thanks & Regards,
>