vhand process

[Kennethyap]

Dear Experts,
What is the vhand deamon running on UNIX OS ver.11.11 ?
Its grep most of the CPU Utilization.

Thks

[Tom Ward_1]

Hello,

Vhand is the paging deamon. If it's high in the process list, your system is likely swapping. Check with "swapinfo -tad"

Regards,
Tom

[bhavin asokan]

hi,

vhand is the paging daemon.use glance or vmstat to findout your system is giving large no. of pageouts.if pageouts are more you should have to increase your physical memory.

#vmstat 1 10

check what is coming under pi & po.

regds,

[Fred Ruffet]

Are you using Oracle ?
vhand uses lots of CPU happens when you have an overload of caching (Oracle cache + IO Cache). It may be better decreasing valus for dbc_max_pct and dbc_min_pct HP kernel Parameters, or even better mounting FS for DB files with convosync=direct and mincache=direct options.

Even if you don't use oracle, you can consider first point.

Regards,

Fred

[Kennethyap]

It cause the system hang can't login but can ping the IP. I have to force the server to TOC. How can I check what's the process cause the vhand very high utilization.

[Fred Ruffet]

There's no "one-solution-answer".

You may watch what processes are using lot of memory, but some may have to...

Just tell us :
. Are you using Oracle ?
. What is your dbc_min_pct value ?
. What is your dbc_max_pct value ?
. How many RAM you have ?
. How many swap ?
. If using Oracle, output of ipcs -mb

Regards,

Fred

[Kennethyap]

Are you using Oracle ?
Yes
. What is your dbc_min_pct value ?
5
. What is your dbc_max_pct value ?
15
. How many RAM you have ?
2048MB
. How many swap ?
4096MB
. If using Oracle, output of ipcs -mb

IPC status from /dev/kmem as of Tue Oct 26 17:55:51 2004
T ID KEY MODE OWNER GROUP SEGSZ
Shared Memory:
m 0 0x411c8de2 --rw-rw-rw- root root 348
m 1 0x4e0c0002 --rw-rw-rw- root root 61760
m 2 0x41204f7a --rw-rw-rw- root root 8192
m 4611 0x0c6629c9 --rw-r----- root sys 35121576
m 4 0x06347849 --rw-rw-rw- root root 77384
m 517 0xffffffff --rw-r--rw- root root 22908
m 25606 0x5e1000fc --rw------- root root 512
m 17927 0x18117500 --rw-rw---- oracle dba 129105920
m 8 0x0003235c --rw-rw-rw- rcebmp tcs 291376
m 9 0x0000c062 --rw------- rcebmp tcs 16924
m 10 0x00000000 --rw------- rcebmp tcs 628

[Kennethyap]

Correction...


. How many RAM you have ?
4096MB
. How many swap ?
10240MB

[Fred Ruffet]

Do you have Online JFS installed ? If yes, I would suggest you to re-mount all FS used to store DB files with this options :
convosync=direct,mincache=direct,nodatainlog

Depending on other IOs the server is making, with so much memory, you could consider decreasing dbs_max_pct and dbc_min_pct.
On an Oracle server with 3GB of RAM I have dbc_min_pct/dbc_max_pct are set to respectively 2 and 4. FS for DB files are mounted with options I gave. Server is running fine. Before reducing parameters and setting options, I ran into the same problems you have.

Regards,

Fred

[Kennethyap]

Thks Fred... I will try it....