Announcement

**oshunluvr** · Mar 15, 2016, 01:45 PM

Could be a process still has an open file and is preventing umountroot from remounting the root filesystem as read-only (this is part of the shutdown process). I see a couple Network Manager messages there on your screen shot. It could be the culprit. Just for fun, boot up that system and kill network manager completely then try shutting down. If that doesn't work, maybe adding this to the umountroot script will give some useful output:

ps ajx or maybe lsof

As always, have a liveUSB bootable handy before doing any system edits.

The other screenie shows a kernel panic (not KDM panic?). Is that what you meant? We haven't used KDM by default in quite awhile.

**Frank616** · Mar 15, 2016, 04:11 PM

oshunluvr,

I thought I had edited that KDM comment. Sorry. It is indeed a kernel panic, and I knew that when I posted. Just missed editing it out, I guess.

OK, how can I kill network manager before trying a shutdown?

Where will I find the unmountroot script?

Thanks for the suggestions.

Frank.

**vinnywright** · Mar 15, 2016, 04:31 PM

next time it hangs ,,,,,,try ,,,,,press and hold "ctrl+alt+SysRq" wile holding them down press one at a time S then U then B ,,,,,,,,,,the system should reboot.

VINNY

**Frank616** · Mar 15, 2016, 04:37 PM

Vinny:

I'll try that.

I booted from a Kubuntu USB stick, and ran fsck on the offending sdb1 partition. It found a block number error. I asked it to fix it, and now it comes back 'clean', but I still get the kernel panic. I'll try gparted tomorrow or Thursday when I have some time off again.

Frank.

**vinnywright** · Mar 15, 2016, 06:29 PM

Originally posted by Frank616 View Post

Vinny:

I'll try that.

I booted from a Kubuntu USB stick, and ran fsck on the offending sdb1 partition. It found a block number error. I asked it to fix it, and now it comes back 'clean', but I still get the kernel panic. I'll try gparted tomorrow or Thursday when I have some time off again.

Frank.

if this is the same one you did this to ,,,,,,,

I changed this line in file /etc/default/grub to look like this: GRUB_CMDLINE_LINUX_DEFAULT="quiet splash init=/lib/systemd/systemd"
After that, I ran sudo update-grub in a Terminal

,,,,,,,then undo it ,,,,their is no systemd in 14.04

OOPS wrong about that ,,,,,,,,,,but I would still try undoing it.

VINNY

**oshunluvr** · Mar 16, 2016, 12:20 PM

You should be able to kill networkmanager from ksysguard.

My system shows umountroot at:

:~$ locate umountroot
/etc/init.d/umountroot
/etc/rc0.d/K11umountroot
/etc/rc6.d/K11umountroot
/lib/systemd/system/umountroot.service

The last one is a link to null, so must be a place holder for future use. The 2nd and 3rd both link to the first, so edit /etc/init.d/umountroot. try adding the lsof and/or ps ajx commands below the three lines that contain the remount command.

**Frank616** · Mar 17, 2016, 01:26 PM

oshunluvr and vinny:

next time it hangs ,,,,,,try ,,,,,press and hold "ctrl+alt+SysRq" wile holding them down press one at a time S then U then B ,,,,,,,,,,the system should reboot.

Did nothing. System is shut down, right? Should the keyboard do anything at this point?

Just for fun, boot up that system and kill network manager completely then try shutting down

As soon as I kill it, it restarts.

I will try yet another update today and see if that changes anything. If not, then I'll try as you suggest. I just looked up those commands and now understand that they should give more information so we know where to go from here. However, I cannot find an 'ajx' command Is there a spelling error there?

Correction: I see that ajx are flags. sorry.

try adding the lsof and/or ps ajx commands below the three lines that contain the remount command.

Finally, do I understand correctly that the commands should go after the last UNMOUNT command? I see no 'remount' commands in there. The file should look like:

Code:

:~$ locate umountroot
/etc/init.d/umountroot
/etc/rc0.d/K11umountroot
/etc/rc6.d/K11umountroot
lsof ps ajx
/lib/systemd/system/umountroot.service

I thank you both for taking the time to help out.

Correction: I totally misread the above LISTING. sorry. Another post to follow this one....

Frank.

**Frank616** · Mar 17, 2016, 02:07 PM

oshunluvr:

OK, I finally stopped and READ you example of your SEARCH for umount. I have opened /etc/init.d/umountroot

The script is as follows:

Code:

#! /bin/sh
### BEGIN INIT INFO
# Provides:          umountroot
# Required-Start:
# Required-Stop:
# Should-Stop:       halt reboot kexec
# Default-Start:
# Default-Stop:      0 6
# Short-Description: Mount the root filesystem read-only.
### END INIT INFO

PATH=/sbin:/bin
. /lib/init/vars.sh

. /lib/lsb/init-functions

do_stop () {
    [ "$VERBOSE" = no ] || log_action_begin_msg "Mounting root filesystem read-only"
    # Ask init to re-exec itself before we go down if it has been
    # upgraded this cycle.  It'll lose all its state, but at least
    # it won't hold open files on the root filesystem (lp:#672177).
    if [ -f /var/run/init.upgraded ]
    then
        old_map=$( cat /proc/1/maps )
        map=$old_map
        telinit u || :
        i=0
        timeout=5
        while [ "$map" = "$old_map" ]
        do
            sleep 1
            map=$( cat /proc/1/maps )
            /usr/bin/logger "waiting for init to respawn"
            i=$((i+1))
            if [ $i -eq $timeout ] ; then
                break
            fi
        done

        if [ "$map" = "$old_map" ] ; then
            /usr/bin/logger "FAIL: init failed to respawn in $timeout seconds - unmounting anyway"
        else
            /usr/bin/logger "SUCCESS: init respawned after $i seconds (within $timeout seconds timeout)"
        fi
    fi

    MOUNT_FORCE_OPT=
    [ "$(uname -s)" = "GNU/kFreeBSD" ] && MOUNT_FORCE_OPT=-f
    # This:
    #     mount -n -o remount,ro /
    # will act on a bind mount of / if there is one.
    # See #339023 and the comment in checkroot.sh
    mount    $MOUNT_FORCE_OPT -n -o remount,ro -t dummytype dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro              dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro                       /
    ES=$?
    [ "$VERBOSE" = no ] || log_action_end_msg $ES
}

case "$1" in
  start)
    # No-op
    ;;
  restart|reload|force-reload)
    echo "Error: argument '$1' not supported" >&2
    exit 3
    ;;
  stop)
    do_stop
    ;;
  *)
    echo "Usage: $0 start|stop" >&2
    exit 3
    ;;
esac

:

So, If I have finally understood correctly, I should end up with the following snippet, right?

Code:

MOUNT_FORCE_OPT=
    [ "$(uname -s)" = "GNU/kFreeBSD" ] && MOUNT_FORCE_OPT=-f
    # This:
    #     mount -n -o remount,ro /
    # will act on a bind mount of / if there is one.
    # See #339023 and the comment in checkroot.sh
    mount    $MOUNT_FORCE_OPT -n -o remount,ro -t dummytype dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro              dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro                       /
    lsof ps ajx
    ES=$?
    [ "$VERBOSE" = no ] || log_action_end_msg $ES
}

Sorry, its been a long day already....

Thanks for your patience.

Frank.

**oshunluvr** · Mar 17, 2016, 04:20 PM

Its two commands, so try

lsof
ps ajx

**Frank616** · Mar 24, 2016, 11:41 AM

OK, just got back to sleuthing this issue.

I edited /etc/init.d/umountroot to look as follows:

Code:

MOUNT_FORCE_OPT=
    [ "$(uname -s)" = "GNU/kFreeBSD" ] && MOUNT_FORCE_OPT=-f
    # This:
    #     mount -n -o remount,ro /
    # will act on a bind mount of / if there is one.
    # See #339023 and the comment in checkroot.sh
    mount    $MOUNT_FORCE_OPT -n -o remount,ro -t dummytype dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro              dummydev / 2>/dev/null \
    || mount $MOUNT_FORCE_OPT -n -o remount,ro                       /
    lsof       >  /home/irma/lsof.txt
    ps ajx    >  /home/irma/ps.txt
    ES=$?
    [ "$VERBOSE" = no ] || log_action_end_msg $ES
}

As you can see, I played with the commands in an effort to pipe the output to a text file that I could read later, as otherwise the output goes by on the screen too quickly to read. Anyway, the system did not like my little effort, calling my command a 'garbage option'! (snif). The output I got is on the attached screenshot.

Suggestions? Corrections?

Thanks.

Frank.

Attached Files

**Frank616** · Mar 25, 2016, 03:48 PM

Could still use some help with this, thanks.

Did the pipe fail because the machine is mostly shut down? It worked on my own machine when I tried it from a console. Should I try

lsof | more

?

Frank.

**oshunluvr** · Mar 25, 2016, 05:05 PM

Can't hurt. Likely cant redirect to a file this late in the shutdown.

You could remove the redirect, record the scrolling text with your phone, play it back in slo-mo

**Frank616** · Mar 25, 2016, 09:32 PM

Hadn't thought of taking a video.

Though I think I'll try lsof | more first, and take pictures one screen at a time.

Once I have the output, what am I looking for?

Frank.

**oshunluvr** · Mar 26, 2016, 04:59 AM

Some process has a file open. You want to know what processes are running that have open files. I'm still wagering it's network manager. You should really try and kill it and all network processes then shutting down manually to see if it breezes past the remount.

Announcement

mount: / is busy and system hangs without shutting down

mount: / is busy and system hangs without shutting down

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment