Bug#496917: BUG: soft lockup - CPU#1 stuck for 4096s! [java:3174] and crash (kvm guest)

[Lukas Kolbe <lukas@einfachkaffee.de>]

Am Sonntag, den 14.12.2008, 23:50 +0100 schrieb Moritz Muehlenhoff:
> On Mon, Sep 01, 2008 at 03:56:49PM +0200, Lukas Kolbe wrote:
> > Hi!
> > 
> > > >seeing if it is fixed in 2.6.27-rc5 might be more interesting.
> > > >thanks
> > > 
> > > 2.6.27-rc5 has now been running fine in the guest for more than four
> > > hours (and me restarting jboss every now and then). I'll report back tomorrow evening, that would be the timeframe
> > > the bug should've triggered. 
> > 
> > So far the 2.6.27-rc5 seems to be stable, at least it hasn't crashed on
> > me. What can I do to help to get the needed fix to testing? (I know that
> > this kernel won't make it and that's a good thing, but I don't really
> > know how to identify what's needed to fix this).
> 
> Does this bug still persist with the current Lenny kernel?

I am sorry to report that I cannot reproduce this bug anymore, but
that's because the system locks up hard (I have no serial console to it
and it is remotely accessible only).

When I do an rsync of ca. 100GB of data on a guest from a remote
location (both guest and host with 2.6.26-1-amd64 2.6.26-12, but it also
happened with -9), after around 10-25GB the host (!) locks up. Ping does
work, but the shells are dead and no further access (via ssh) is
possible at all. After a forced reboot, the logs show up completely
clueless (as in: syslog marks, but nothing else until the new bootup
messages).

When I do the rsync in the host system (chroot'ed to the guest system
for easyness), I can transfer all 100GB without a hitch.

I am using virtio by the way, as in:

#!/bin/bash

KERNEL=2.6.26-1-amd64
NAME=myname

kvm -smp 2 \
 -drive if=virtio,file=/dev/vg0/${NAME}-root,cache=on,boot=on \
 -drive if=virtio,file=/dev/vg0/${NAME}-log,cache=on,boot=off \
 -m 512 \
 -nographic \
 -daemonize \
 -name ${NAME} \
 -kernel /boot/kvm/${NAME}/vmlinuz-${KERNEL} \
 -initrd /boot/kvm/${NAME}/initrd.img-${KERNEL} \
 -append "root=/dev/vda ro console=ttyS0,115200" \
 -serial mon:unix:/etc/kvm/consoles/${NAME}.sock,server,nowait \
 -net nic,macaddr=DE:AD:BE:EF:21:75,model=virtio \
 -net tap,ifname=tap04,script=/etc/kvm/kvm-ifup \
 -net nic,macaddr=DE:AD:BE:EF:21:76,model=virtio \
 -net tap,ifname=tap14,script=/etc/kvm/kvm-ifup-${NAME} \

I'm sorry as I can't at the moment reproduce the problem that lead to
open this bug. Maybe I should open a new bug for this?

> Cheers,
>         Moritz

All the best and a happy new year, 
Lukas