Can you monitor actual 'size in use' for various kernel parameters ?

[Simon R Wootton]

I'm concerned that some of my kernel parameters are getting close to the limit.

Is there an easy way to monitor 'actual' usage for these chaps instead of just seeing the settings in SAM ?

All help appreciated & rewarded !!

Simon

[Pete Randall · Accepted solution]

If you have Glance, the "System Tables Report" will show you real time values.


Pete

[Duncan Edmonstone]

Simon,

If you are on 11iv2 or 11iv3 then take a look at kcusage(1m)

e.g. show usage of nproc kernel parm for last month:

kcusage -m nproc

For 11iv1, I think you'll only be able to look at data from sar -v (and only hsitorical data if you have configured accounting)

HTH

Duncan

[Paul McCleary]

Hi Simon,

What about looking at kcusage - I think this should give you what you want.

sysdef might also be worth looking at. I don't have a HP-UX system at the moment so I can't check the output.

Hope this helps,
Paul

[Jeeshan]

sysdef, you can use every OS

but kctune usage is limited in 11iv2 and 11iv3

[Ivan Krastev]

Set the alarm with kcalarm.
If necessary edit the dictionary file - /etc/opt/resmon/dictionary/kcmond_01.dict


And after that enable monitoring.


regards,
ivan

[Steven E. Protter]

Shalom Simon,

The glance system has several useful features.

There is a graphical high water mark for important measures in the kernel and swap.

This lets you see how close you came to running out.

SEP

[Dennis Handly]

>ahsan: sysdef, you can use every OS

You shouldn't recommend using this "TO BE OBSOLETED" command. Instead use kctune(1m) or kmtune(1m) (11.11).

[Michal Tuma]

Hi Simon,
log on as root and use '/usr/sbin/kcweb'.
There is everythink about kernel parameters.

Michal

[Simon R Wootton]

Thanks to all for replies