Non responsive state for systemd

Lennart Poettering mzerqung at 0pointer.de
Tue Mar 5 00:05:23 UTC 2013


On Mon, 04.03.13 10:24, David Highley (dhighley at highley-recommended.com) wrote:

> "Lennart Poettering wrote:"
> > 
> > On Mon, 04.03.13 07:56, David Highley (dhighley at highley-recommended.com) wrote:
> > 
> > > Twice now we have one Fedora 18 system where systemd seems to get into a
> > > non responsive state. We are not able to get the status of any service
> > > and we're not able to do an init 6 to restart the system.
> > > 
> > > Did notice today that a full process list showed a message about abrt
> > > and something to the effect "nobody cared". We also see a number of
> > > defunct processes that seem to never clear. So far the only remedy we
> > > have found is a hard power cycle.
> > 
> > Can you get a stack trace of PID1? "sudo pstack 1" should already give a
> > hint, but even better would be a a "bt full" via gdb.
> 
> We are offsite right now so will dig deeper later. We had checked the
> log files and noticed that it complains about rsyncd not being able to
> connect to a port and there was another complaint about Gnome. The
> rsync one repeats as there are back ups that are not being serviced
> which is is what alerted to something being wrong. We are sending and
> receiving email from this system. It also has an internal web, mysql,
> and other subsystems which seem to work fine. So when this state occurs
> it sometimes takes a while to notice.

This is a bug in libselinux:

https://bugzilla.redhat.com/show_bug.cgi?id=901812

Lennart

-- 
Lennart Poettering - Red Hat, Inc.


More information about the devel mailing list