Bob Scheifler wrote:
> I've had several instances recently of a zone in the DOWN state,
> for which zoneadm halt just hangs, and the only recourse seems
> to be to reboot the machine.  The zone may have gotten into the
> down state due to bug 6504424, but I'm not sure.  Looking
> at a forced crash dump, the zone's zoneadmd has a thread stuck in
> cv_wait_sig on zone_destroy_cv:
> 
> [ fffffe8004f13d10 _resume_from_idle+0xf8() ]
>    fffffe8004f13d40 swtch+0x10b()
>    fffffe8004f13d90 cv_wait_sig+0x15c()
>    fffffe8004f13dd0 zone_destroy+0x124()
>    fffffe8004f13ec0 zone+0x149()
>    fffffe8004f13f10 sys_syscall32+0x101()
> 
> Is this related to bug 6516822, or something else?  Any clues appreciated.
> I'm happy to provide the crash dump or info from it.

There is really no way to tell if this is a dup of another bug without
more information.  We do have these bugs open:

6540312 Unable to shutdown zone.
6533188 killall() doesn't seem to be killing all

In order to diagnose a problem like this we need a system dump to
try to figure out where in the kernel things are hung.  It could
be a bug in zones or in some other subsystem which is causing
a thread to hang in the kernel.  Even though the symptom is
the same (zone won't halt) it could be because of different bugs,
so we always need a system dump to analyze this.  The best thing
is to always file a bug with an attached crash dump so we have the
data we need to work on these problems.

Thanks,
Jerry


_______________________________________________
zones-discuss mailing list
zones-discuss@opensolaris.org

Reply via email to