Hi, I am trying to track down a supposed memory leak. All I know
about the problem is that the memory gets up to 90%, and then we have
to restart the application and database.

In glance, I view the processes, but it doesn't seem to be all of the
processes. ps -ef | wc -l shows 534 processes. When I check glance,
it only shows about 20.

B3692A GlancePlus C.03.55.00 16:23:33 lh_hlab 9000/800
Current Avg High
------------------------------------------------------------------------------------------------------------------------------------
CPU Util S SU U
| 24% 31% 90%
Disk Util FF
| 2% 4% 11%
Mem Util S SU
UB B | 76% 76% 76%
Swap Util U UR R
| 40% 40% 40%
------------------------------------------------------------------------------------------------------------------------------------
PROCESS
LIST Users= 3
User CPU Util Cum Disk
Thd
Process Name PID PPID Pri Name ( 200% max) CPU IO Rate
RSS Cnt
--------------------------------------------------------------------------------
PATGEN.EXE 7450 345 154 prod 9.4/ 0.5 62.2 15.0/ 0.7
5.7mb 1
GS.EXE 4005 4004 189 prod 9.2/ 9.2 0.5 2.3/ 2.3
na 1
perl 3996 3995 184 prod 4.7/ 2.3 0.2 0.9/ 0.4
5.7mb 1
FORMAE.EXE 4002 4001 182 prod 3.3/ 3.3 0.2 0.0/ 0.0
na 1
DOCREND.EXE 3998 407 179 prod 2.9/ 2.9 0.2 0.5/ 0.5
14.1mb 1
router 20295 1 168 pwllink 1.8/ 0.4 10753.9 2.5/ 0.6
18.0mb 1
glance 3915 26338 154 rdba 1.6/ 2.0 0.9 0.0/ 0.1
6.4mb 1
RPTCTL.EXE 812 345 154 prod 1.2/ 0.1 295.4 0.0/ 0.0
10.4mb 1
DOCAGNT.EXE 417 345 154 prod 1.2/ 0.6 1582.0 0.0/ 0.0
45.0mb 1
glance 407 345 154 prod 0.2/ 0.1 268.0 2.1/ 0.4
14.1mb 1
DOCDIST.EXE 815 345 154 prod 0.2/ 0.1 231.2 1.3/ 0.2
26.5mb 1
vxfsd 18 0 138 root 0.0/ 26k 10772.3 5.1/ 2m
1.8mb 58
ora_lgwr_PW 12978 1 156 oracle 0.0/ 0.0 759.5 1.5/ 1.1
4.5mb 1
sh 1750 26589 158 patrol 0.0/ 0.0 0.0 0.0/ 0.0
25.1mb 1
DOCAGNT.EXE 417 345 154 prod 0.0/ 0.6 1582.0 0.0/ 0.0
45.0mb 1
sh 2230 26589 158 patrol 0.0/ 0.0 0.0 0.0/ 0.0
25.8mb 1
perl 26670 471 168 prod 0.0/ 0.1 3.8 0.0/ 0.0
22.5mb 1
PWLPOST.EXE 854 345 154 prod 0.0/ 0.8 2072.9 0.0/ 0.0
92.9mb 1
ora_ckpt_PW 12980 1 156 oracle 0.0/ 0.0 519.0 1.1/ 0.9
4.5mb 1
ora_ckpt_PW 12935 1 156 oracle 0.0/ 0.0 620.7 1.1/ 1.1
3.2mb 1
rcvd 20283 1 154 pwllink 0.0/ 0.0 688.9 2.8/ 0.7
684kb 1
mib2agt 6696 1 154 root 0.0/ 0.0 914.4 0.0/ 0.0
199.2mb 1
sh 26507 26589 158 patrol 0.0/ 0.0 0.0 0.0/ 0.0
26.4mb 1
ora_dbw0_PW 12976 1 156 oracle 0.0/ 0.0 516.9 8.6/ 1.4
8.7mb 1
PWLPOST.EXE 719 345 154 prod 0.0/ 0.5 1244.3 0.0/ 0.0
20.4mb 1


Also, if I look in top, it shows the RSS size of all the processes,
but I don't know if I can list the processes by RSS size. I don't
think that top is as accurate as it could be. Also, if you have any
other tools to find which process is causing the memory leak, please
let me know.

Thanks, Margie

spy_234432@yahoo.com (Margie) wrote:

>Hi, I am trying to track down a supposed memory leak. All I know
>about the problem is that the memory gets up to 90%, and then we have
>to restart the application and database.
<SNIP>

If you have access to HP's ITRC Knowledge Base, take a look at a document
titled something close to "How to account for memory usage in HP-UX".

It will provide info on the tools and procedures you're looking for.

Eric Stahl

> In glance, I view the processes, but it doesn't seem to be all of the
> processes. ps -ef | wc -l shows 534 processes. When I check glance,
> it only shows about 20.
If you want to see all processes, you must turn off filtering in glance
menu.