topic Re: hpasm compaq agents priority/nice problem in Operating System - Linux

hpasm compaq agents priority/nice problem

spanky mcfoo — Mon, 12 May 2003 18:38:05 GMT

Anyone else running RHAS 2.1 with compaq agents (hpasm 6.30) seeing their priority at "-20"? At this priority, init, getty, etc. are all pre-empted. I'm wondering if there's a good reason that they run at the highest possible priority.

The machine is almost unusable as-is--very unresponsive on the terminal & it takes a long time to boot up (after cpq agents start).

I have a bug open with compaq on this issue--I'll post any result here as well.

Thanks for your insights,

Rob

[root@beta-db0 root]# uname -a
Linux beta-db0 2.4.9-e.16enterprise #1 SMP Mon Mar 17 16:37:14 EST 2003 i686 unknown

[root@beta-db0 root]# rpm -qa|grep hp
hpasm-6.30.0-12

[root@beta-db0 root]# ps -eo pid,user,nice,cmd --sort -nice | tail -20
10259 root 0 cmawebd
10261 root 0 cmawebd
10329 root 0 cmastdeqd -p 30
10353 root 0 cmahealthd -p 30 -s OK -t OK -i
10465 root 0 cmawebd
10466 root 0 cmawebd
10467 root 0 cmawebd
10775 root 0 ps -eo pid,user,nice,cmd --sort -nice
10776 root 0 tail -20
10137 root -20 /opt/compaq/cpqhealth/cpqevt/cevtd
10142 root -20 /opt/compaq/cpqhealth/cpqasm/casmd
10185 root -20 cmafdtnpeerd
10187 root -20 cmafdtnpeerd
10188 root -20 cmafdtnpeerd
10189 root -20 cmafdtnpeerd
10303 root -20 cmasvrpeerd
10324 root -20 cmasvrpeerd
10325 root -20 cmasvrpeerd
10326 root -20 cmasvrpeerd
10432 root -20 /opt/compaq/cpqhealth/cpqasm/casmd

Re: hpasm compaq agents priority/nice problem

Steven E. Protter — Mon, 12 May 2003 18:59:48 GMT

Does the system become more responsive when you shut down (hpasm 6.30) ?

If the process is not critical to server operation, shut it down.

You have identified what the server really needs to run.

If you need (hpasm 6.30) running then explore the renice command to lower this processes priority.

SEP

Re: hpasm compaq agents priority/nice problem

spanky mcfoo — Tue, 13 May 2003 21:20:41 GMT

thanks for the feedback.

Yeah, we considered renicing, but we don't want to start doing that until we've confirmed that it's a bug with HP. Also, the renicing only seems to last for a little while in my tests before the priority is back--we'd have to run a periodic checker to make sure everyone's "in line".

The agents are critical for us--we'll have no idea if a disk is blown without them!