You should also look to see if the machine is running out of memory. I've had this problem before and it turned out the machine was running out of RAM and also running out of SWAP...
If you keep a console open before the problem starts to appear you might be able to see the issue - maybe even running top from an ssh session so you can see cpu and memory usage without spawning new processes when it starts happening On Mon, Aug 25, 2008 at 2:48 PM, Martin Barry <[EMAIL PROTECTED]> wrote: > $quoted_author = "Grant Street" ; > > > > I have a problem where I am running thirdparty software sporadically my > > machine(s) get into a state where they are responsive to pings but not > > ssh or a local console. > > Re-nicing the troublesome software doesn't help? > > > > 1 enable Magic SysRq to hopefully get somthing out of it. > > You could give that a try and see if it gets you a console back at least. > > > > 2 Is there a way to tune the way/algorithm the kernel uses so that under > > load it will kill off processes earlier or more aggresively under > > extreme load. > > I'm not aware of anything like a CPU version of OOM killer but I have come > across a project aimed at process level CPU limiting. > > http://cpulimit.sourceforge.net/ > > cheers > marty > > -- > "It'd go along with the 'Caution - hidden dip' signs on some of then more > undulating local roads, which always have me looking out for concealed > bowls of > guacamole." --Tanuki > > alt.sysadmin.recovery - <[EMAIL PROTECTED]> > -- > SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ > Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html > -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
