[arch-general] Shutdown and reboot not working after last weekend update

pants pants at cs.hmc.edu
Tue Jun 12 16:42:01 EDT 2012


Well, in general shutdown itself will *not* be the halted process; that
would be rather strange.  But hey, it could have happened.  Usually,
shutdown and reboot hang because they are coded to wait until all other
processes reported having quit, and the process in question hasn't quit
yet (because it never will).  Are there any other similar messages for
other processes?

Also, now that you've rebooted it's unlikely that you still have a
process in uninterruptible sleep.  I would run fsck (the one for the
filesystem you're using) to make sure that forcing a shutdown didn't
cause any problems, and perhaps check the SMART status of the drives,
but I wouldn't worry too much.

pants.

On Tue, Jun 12, 2012 at 05:36:26PM -0300, Victor Silva wrote:
> I do have this exact error messega as I've posted before.  I will check
> htop and post the output here. So now some questions which come out are: is
> it somehow related to bad sectors? Shutdown is the locked process with the
> message:
> 
> > INFO: task [shutdown] blocked for more than 120 seconds.
> 
> Still I've made a forced reboot via sysctl and on the next session I've
> also hit this problem.
> 
> About xfs_fsr (which I assume is some kind of fsck right?) I need to run
> fsck from a live cd, I just got one and will try it tonight.
> 
> So basically we are talking about the samething, still I'm not sure of what
> is going wrong and sincerely I'm  bit afriad of having a system just messed
> up because of a failing disk. How can I cover this scenario?
> 
> Regards,
> Victor
> 
> 2012/6/12 pants <pants at cs.hmc.edu>
> 
> > It sounds like you might have had a process in uninterpretable sleep.
> > This happens sometimes when a process is in a system call and stops
> > working.  You can check for this by launching htop and seeing if any
> > processes are permanently in the 'D' state (in the column labeled 'S').
> > Alternatively, look in /var/log/errors.log and see if you can find
> > messages which look like:
> >
> > > INFO: task [processinfo] blocked for more than 120 seconds.
> >
> > If you do have a process in this state, you won't be able to kill it or
> > unmount the filesystem that it's stuck modifying.  Your best bet is to
> > manually kill everything you can, manually umount everything you can,
> > and then force the system down.  It sucks, I've had to do this with
> > processes stuck writing to a RAID and then have to wait for a few days
> > while the whole thing is resynced.
> >
> > It also probably isn't the fault of an upgrade, just some piece of buggy
> > software that finally hit its error.  I know that xfs_fsr will quickly
> > and deterministically go into it when run on a filesystem which is
> > almost full.
> >
> > Good luck!
> >
> > pants.
> >
> > On Tue, Jun 12, 2012 at 05:22:43PM -0300, Victor Silva wrote:
> > > I got a ubuntu livecd with gparted lets see what happens. About using
> > > thesystem as root. I only run pacman related commands as root have  a
> > user
> > > I use for mostly day tasks. This user is *NOT* on sudo, basically cause I
> > > like to enforce my self to login as root and think about what I'm doing.
> > > Also sudo is cumbersome to use in some instances as buil-in commands are
> > > not avaliable.
> > >
> > > So the only thing I've run as root was the pacman -Syu and followed the
> > > arch announces regarding the updates. Nothing bad there I'mho. And the
> > > problem I'm facing now has nothing to do with my system usage. Imho
> > either
> > > I have a bad sector or something went broken for some instances of the
> > > system like mine or the other guy who also reported on the forums.
> > >
> > > Regards,
> > > Victor
> > >
> > > 2012/6/12 Attila Vangel <vangel.attila at gmail.com>
> > >
> > > > Hi, I am not familiar with the problem, but I think the easiest way
> > > > (if you are not against graphical tools) is to grab a live cd (or
> > > > "live usb dongle") containing gparted (it's not a bad thing to have,
> > > > anyway), but I am not an expert at this.
> > > >
> > > > What really surprises me is that you use your system as root !?! I
> > > > would not dare to do that. I tend to add my user to various groops
> > > > according to the arch wiki documentation where neeeded, and for the
> > > > other commands I think I can safely run as root I alias them to be
> > > > 'sudo <command>', and I maintain this list of safe commands in
> > > > /etc/sudoers (edited by visudo (you can change the editor of it, just
> > > > google it)), so that these commands can be executed without entering
> > > > the password all the time... Maybe not the best thing still, but I
> > > > guess it's OK with me.
> > > >
> > > > Regards,
> > > > Attila
> > > >
> > > > 2012/6/12 Victor Silva <vfbsilva at gmail.com>:
> > > > <snip>
> > > > > Folks after the last upgrade I can no longer shutdown nor reboot my
> > > > machine
> > > > > (I'm using it as root). The command simply hangs and nothing happens.
> > > > <snip>
> > > >
> > > > > I was asked to perform a fsck which failed.It reported /dev/sda5 was
> > > > > mounted. Is there any proper way I should use to call fsck? I did
> > create
> > > > a
> > > > > /fsck file on / is there other more appropriate command to do it?
> > > > Problably
> > > > > you asked for my fstab expecting an error like this right? Would it
> > be
> > > > > better to run fsck from a livecd?
> > > > >
> > > > > regards,
> > > > >
> > > > > Victor
> > > >
> > >
> >
> 


More information about the arch-general mailing list