[CentOS-virt] Cannot start xen domUs anymore, domUs hang on kernel startup, happens after a long dom0 uptime

[Pasi Kärkkäinen]

On Thu, Aug 14, 2008 at 10:22:29AM +0300, Pasi Kärkkäinen wrote:
> On Wed, Aug 13, 2008 at 08:51:34PM -0500, Christopher G. Stach II wrote:
> > Pasi Kärkkäinen wrote:
> > >Hello!
> > >
> > >I've noticed this problem two times now.. last time I fixed it by rebooting
> > >the (centos 5.1 x86 32b) xen host/dom0.
> > >
> > >Symptoms:
> > >
> > >- Already running domUs (debian 2.6.18-6-xen-686 32b PAE) continue running
> > >  and working OK 
> > >
> > >- Cannot start new domUs (debian 2.6.18-6-xen-686).. kernel bootup just
> > >hangs before running initrd. Same domU with the exact same xen domU 
> > >cfgfile worked
> > >earlier. 
> > >
> > >- This problem starts to happen after a "longer" uptime.. atm uptime for 
> > >the
> > >dom0-host is 174 days. Cannot say what the actual "limit" is.. last time I
> > >used the dom0 everything worked fine, now it doesn't.. 
> > >
> > >
> > >"xm list" shows the domain as "-b----" and the Time(s) for the domain does
> > >not increase.. it stays at 1.9 or so. It might increase to 2.0 after 15 
> > >mins
> > >or so..  it looks like nothing happens in domU kernel.
> > >
> > >"xm console" shows the bootup has stalled.. destroying the domU and
> > >restarting it might show the hang in a bit different place 
> > >(+- a couple of lines). It always hangs before actually executing initrd
> > >image. 
> > >
> > >Example console output:
> > >
> > >checking if image is initramfs... it is
> > >Freeing initrd memory: 12028k freed
> > >Grant table initialized
> > >NET: Registered protocol family 16
> > >SMP alternatives: switching to SMP code
> > ><hangs here, nothing happens anymore>
> > >
> > >Like said, it could be a couple of lines later or earlier where it hangs.. 
> > >
> > >I tried changing the domU to use just a single vcpu.. doesn't help. I tried
> > >changing (lowering and increasing) the amount of memory.. didn't help 
> > >either.
> > >
> > >Any ideas how to debug this? 
> > >
> > >-- Pasi
> > >_______________________________________________
> > >CentOS-virt mailing list
> > >CentOS-virt at centos.org
> > >http://lists.centos.org/mailman/listinfo/centos-virt
> > 
> > Sometimes xenconsoled dies after a while.  Is it running when you 
> > attempt this?
> > 
> 
> Yep, it was running.. maybe I'll try restarting it. 
> 

After killing xenconsoled and restarting it I'm now able to "xm create"
domUs again.. 

Any idea if this is a known bug? 

-- Pasi