Re: accessing a system with high load

From: Robert Miller (rmiller@SMUD.ORG)
Date: Thu Feb 27 2003 - 10:23:30 EST


In general, I believe that if you have physical access to the console you
have a much better chance of being able to deal with the problem gracefully.
I *believe* that there is code in the kernel that allows the root user a
higher priority for their processes when they're logged into the console.
This is one reason people leave locked root sessions on the console. I'm
not sure if it'd help much in the case of being Really Low on memory/swap,
but it may... especially if you left a 'top' session running on the console
so you could see some PID's that would be candidates for 'kill'-ing.

I've had something similar happen, and thankfully I was able to run around
to a couple machines and close some X sessions, and had a window open on the
console as root, so I was able to take some stuff out and get the system
running again without having to ungracefully power-cycle it. (would this
have been an option on a Mickey$oft box? not on its best day...)

Anyhow, there's my $.02 worth, hope it helps some...

--rm

-----Original Message-----
From: IBM AIX Discussion List [mailto:aix-l@Princeton.EDU]On Behalf Of
Holger.VanKoll@SWISSCOM.COM
Sent: Thursday, February 27, 2003 6:15 AM
To: aix-l@Princeton.EDU
Subject: accessing a system with high load

Hello,

I am thinking about what to do to ensure access to a system where some
application uses that much paging-space that connections (telnet/ssh/getty)
cant be made anymore (fork fails).

Aix5.1 has the ability (shconf) to do certain things if certain-priority
apps dont get cpu anymore.
Also, one could start a high-priority ssh-daemon on bootup.

Thats fine, but I solves the problem when applications consume too much cpu.
That doesnt help if they consume too much paging-space.

As far as I see even ulimit/wlm has no way to solve this problem.

I could try to start sshd with plock(); but that would only get sshd up
running... any command started from there still fails (fork - not enough
memory available now).

So far, I see no other possibility than to increase paging-space and set
high values for npswarn and npskill (vmtune).
The only disadvantage I currently see is more disk-usage for paging-space.

What do you think / what do you do to ensure access to a high-paging system?

Regards,

Holger



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:37 EDT