From: Rus F. <rg...@fs...> - 2004-03-30 11:17:52
|
Hi All, We've just started using 2.6 hosts and in production are seeing a problem testing didn't chuck up Config Host: Dual Opteron 240 1.4Ghz 2.6.3 + blaisorblade patches Kernel: 2.4.23 + mmremap from 2.4.25 + uml-patch-2.4.23-2 We had a UML hang, non responding to SSH so I tried to bring it back to life using sysrq s sysrq u sysrq b cad (no luck) halt At this point things were still running so I just manually tried to kill then kill -9 the "linux" process. Still didn't die. Now I'm not much of a kernel hacker but strace -p <pid> showed nothing. "gdb -p <pid> kill" didn't kill it Little bit lost to do what next rather than reboot. Also this isn't the first box I've seen this on running 2.6 Guidance would be appreicated Rus -- e: su...@vp... t: 1-888-327-6330 www.jvds.com - Root on your own box www.vpscolo.com - Your next hosting company |
From: Jeff D. <jd...@ad...> - 2004-03-30 16:32:59
|
rg...@fs... said: > At this point things were still running so I just manually tried to > kill then kill -9 the "linux" process. Still didn't die. What did ps show as its state? If it is 'T', then you need to send it a SIGCONT before it will see any other signals. If not, and SIGKILL doesn't make it go away, then that's a host kernel bug. As for the hang itself, if you could get a stack trace from it, that would be useful. Jeff |
From: Rus F. <rg...@fs...> - 2004-03-30 16:42:07
|
On Tue, 30 Mar 2004, Jeff Dike wrote: > rg...@fs... said: > > At this point things were still running so I just manually tried to > > kill then kill -9 the "linux" process. Still didn't die. > > What did ps show as its state? > > If it is 'T', then you need to send it a SIGCONT before it will see any > other signals. If not, and SIGKILL doesn't make it go away, then that's a > host kernel bug. That will be the SIGNOT to clear then. I will try to reproduce with a debug kernel Rus -- e: su...@vp... t: 1-888-327-6330 www.jvds.com - Root on your own box www.vpscolo.com - Your next hosting company |