Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot

[PC]

It looks like the symlinks for S31umountnfs and S40umountfs in rc0.d and rc6.d are wrong because if you access an nfs mount over a VPN, then the machine fails to shut down unless you manually umount the nfs before going to init 0 or init 6.

I changed the S31umountnfs and S40umountfs to K31umountnfs and K40umountfs and the problem seems to have gone.

[devnull (sscott-westnet)]

This is true, either we need to change rc0/6.d S??umountnfs to K??umountnfs, or as I did, change /etc/init.d/umountnfs.sh to do the same thing on both start and stop.

Ive just done this now, so I havent tested it, but someone please CONFIRM this bug.

Thanks

[Scott James Remnant (Canonical) (canonical-scott)]

Though we'll fix this through upstart, and the init scripts will go away

[tweedledee (terrywatt-deactivatedaccount)]

I have the problem described in the duplicate of this bug (CIFS only, not SMBFS or other types) and above. However, the workarounds posted here do not solve the problem in my case - the computer still waits 30-60 seconds for each network connection to fail to shut down. Is there any other way to get around this problem, maybe some script I can tell to automatically execute when I log off/shut down to unmount these before switching run levels? And is this fixed in Fiesty, or will this continue to be a problem?

[Bret Towe (magnade)]

I've hit this bug also I had thought it was just a general computer doesn't want to shutdown issue
till I had a nfs mount on my laptop and it did the same

no vpn is being used here and moving the files from S to K did make it shutdown right again

[Alistair Phipps (alistairphipps)]

Changing the S to K makes the nfs unmount happen earlier during the shutdown process, which seems to solve this issue for some, but not for me. However, this bug was fixed a better way in Debian, which should solve the issue for all. See bug report and patch here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=367944

[Helge Willum Thingvad (helgesdk)]

I can confirm this bug is still present in Ubuntu 7.10 Gutsy.
However, a minute or so after the message the computer usually shuts down anyway.

[Max-Ulrich Farber (maxulrichfarber)]

I am eperiencing the same issue, but only if the network has been started in runlevel 5 by Network Manager or Wicd. If the network has been configured manualy and therefore is started in runlevel 2, there is no shutdown problem even with cifs mounts.

[Reinhard Tartler (siretart)]

bug #181541 contains a patch which could give the base for fixing this bug.

[Robert Di Gioia (digioiar)]

Did this fix make hardy? Putting a failure message up after usplash stops doesn't seem too user friendly.

BTW, my main network connection is wireless, so I fixed it on my system by unmounting cifs and smbfs filesystems just before S15wpa-ifupdown runs by creating a script based on S40umountfs that only unmounts cifs and smbfs filesystems. I run it as S14umountcifs :)

[Matt Austin (mattaustin)]

Confirming that this is still an issue in Ubuntu Hardy Alpha 5 (i386).

[Stefan Carslöv (odur)]

Confirming the bug in Kubuntu Hardy Heron Beta.

[psypher (psypher246)]

Confirming this is still also a bug in Ubuntu Hardy Beta and related to: https://bugs.launchpad.net/ubuntu/+source/samba/+bug/153444

[tweedledee (terrywatt-deactivatedaccount)]

I have also observed this bug a few times when trying to hibernate with a share that has dropped without being unmounted first. Because of the long delay during whatever process is happening to generate the error messages, hibernate fails due to processes not shutting down. (Hibernate will work if tried again if the dead share is unmounted, so it does not have any lasting effect.) I also occasionally see this message in my log files, possibly because I run scripts to dynamically load and unload shares on my network. Any solution to this bug requires significantly more than a revision to the shutdown scripts to be a true resolution.

[toobuntu (toobuntu)]

@Reinhard : How are we to apply the patch/fix you provided to sysvinit in Hardy if it conflicts with upstart?

[EagleDM (eagle-maximopc)]

I did not experienced any boot problems after applying the Patch.

The problem is more serious "without" the Patch than with it, because
basically, if you do not apply this patch, shutdown hangs waiting and you
have to manually shut down the pc or reset which is dangerous.

It should be an alternative, maybe you could try to umount cifs mounts
BEFORE killing the SMBFS service, that is the root of the problem.

Daniel

On Fri, Apr 4, 2008 at 1:16 PM, toobuntu <email address hidden> wrote:

> @Reinhard : How are we to apply the patch/fix you provided to sysvinit
> in Hardy if it conflicts with upstart?
>
> --
> Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot
> https://bugs.launchpad.net/bugs/42121
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[EagleDM (eagle-maximopc)]

Sorry, got confused with another bug number

I was referring to THIS patch

https://bugs.launchpad.net/bugs/153444

I did not try the sysvinit

the aparent problem with closing network CIFS shares is present here: https://bugs.launchpad.net/bugs/153444

The problem seems related to the way ubuntu do the shutdown sequence, killing the smb service before umounting cifs shares.

[Teoh Han Hui (teohhanhui)]

This looks like a more promising fix:

http://blog.avirtualhome.com/2008/03/10/ubuntu-shutdown-problem-cifs-related/

[Laurent Dinclaux (dreadlox)]

problem still present in hardy at time of writing

[EagleDM (eagle-maximopc)]

The fix provided Teih Han Hui quicky fix it for me (hardy heron final),
it's a shame it wans't included being so easy to implement, just a change in
the RC order.

On Wed, May 7, 2008 at 1:29 AM, Laurent Dinclaux <email address hidden>
wrote:

> problem still present in hardy at time of writing
>
> --
> Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot
> https://bugs.launchpad.net/bugs/42121
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[Peter Schneider (ps-euler)]

I can confirm this was still an issue for me on my laptop with Hardy final. I fixed this by renaming S31umountnfs.sh to S18umountnfs.sh in /etc/rc6.d and /etc/rc0.d. It would have been nice if this could have been included in the final release, since a quick google search shows this is still annoying people, and the original bug is from 2006, and the fix is easy.

[EagleDM (eagle-maximopc)]

Yes, totally agree, also, I reported this bug a while ago, when hardy was in
beta stage, this could had being easily patched, there are a lot of us that
are using cifs shares, this is not a minor issue.

On Thu, May 8, 2008 at 5:39 PM, Peter Schneider <email address hidden>
wrote:

> I can confirm this was still an issue for me on my laptop with Hardy
> final. I fixed this by renaming S31umountnfs.sh to S18umountnfs.sh in
> /etc/rc6.d and /etc/rc0.d. It would have been nice if this could have
> been included in the final release, since a quick google search shows
> this is still annoying people, and the original bug is from 2006, and
> the fix is easy.
>
> --
> Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot
> https://bugs.launchpad.net/bugs/42121
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[Martin Pitt (pitti)]

Currently merging sysvinit, Debian's version fixes this.

[Launchpad Janitor (janitor)]

This bug was fixed in the package sysvinit - 2.86.ds1-56ubuntu1

---------------
sysvinit (2.86.ds1-56ubuntu1) intrepid; urgency=low

  * Merge with Debian unstable (first time since Dapper!). Debian adopted most
    of our changes, sometimes with slight modifications. Adopt to the Debian
    structure as far as possible to minimize the delta. Remaining changes:
    - Support Cell processor:
      + debian/initscripts/postinst: Create spu system group and /spu mount
        point if we are running on a Cell processor.
      + debian/initscripts/etc/init.d/mountkernfs.sh: Mount spufs if Cell
        processor is detected.
      + debian/control: Add initscripts dependency 'passwd' for groupadd.
      (Forwarded to Debian #483399)
    - Use tmpfs mounts for /var/lock and /var/run:
      + debian/initscripts/share/default.rcS: Enable RAMRUN and RAMLOCK by
        default.
      + debian/initscripts/postinst: Enable RAMRUN and RAMLOCK in
        /etc/default/rcS on upgrades. This needs to be kept until the next
        LTS.
      + debian/initscripts/etc/init.d/mountkernfs.sh: Propagate files from the
        initramfs to our new /var/run, so that we can populate
        /var/run/sendsigs.omit from initramfs.
    - Different boot order: (to be checked if still required):
      + debian/initscripts/postinst: Change the rcS priorities of a bunch of
        init scripts.
      + Disable bootlogd, since it is handled by upstart in Ubuntu.
    - Usplash support:
      + debian/initscripts/etc/init.d/sendsigs: Call usplash_down on stop.
        (This should go away and get integrated into the existing splash API.)
    - Usplash fsck integration:
      + debian/initscripts/lib/init/usplash-fsck-functions.sh: Functions for
        reporting fsck progress in usplash.
      + debian/initscripts/etc/init.d/check{root,fs}.sh: Include
        usplash-fsck-functions.sh and use it if usplash is running.
    - debian/initscripts/etc/init.d/sendsigs: Always mount devpts, and do not
      touch /dev/ptmx (which is already managed by udev).
    - debian/initscripts/etc/init.d/checkroot.sh: If ACPI is available, load
      the ac module before checking the root filesystem, so that fsck can
      skip the check when running on battery. (LP #89752, forwarded to
      Debian #483394)
    - debian/initscripts/etc/init.d/mountkernfs.sh: Drop mounting of usbdevfs
      (/proc/bus/usb), it was deprecated long ago. (Forwarded to Debian #483392)
    - debian/patches/91_sulogin_lockedpw.dpatch: Disable "root account is
      locked" warning, since this is the default in Ubuntu. Document this in
      sulogin.8.
    - debian/initscripts/etc/init.d/mountall.sh: Set $LANG from
      /etc/default/locale, so that ntfs-3g and friends can get correct file
      name encodings. (LP #132357, forwarded to Debian #483396)
    - debian/initscripts/etc/init.d/umountroot: mkdir a few essential
      directories (/proc, /sys, /var/{run,lock}), right before mounting root
      r/o. It is a convenient (and one of the very few possible) place to
      ensure that the next boot will succeed. (Forwarded to Debian #483393)
    - debian/initscripts/preinst: Fix typo in eliminate_conffile() which b...

[Daniel Ellis (danellisuk)]

I am not sure how this will fix the issue on a default Ubuntu 8.04 installation. Both my desktop and my laptop suffer from not shutting down after adding a cifs mount. I have been using the following script to patch my 7.10 and 8.04 installations and the machines shutdown fine:-

------------------------------------
  sudo mv /etc/rc0.d/S31umountnfs.sh /etc/rc0.d/K31umountnfs.sh
  sudo mv /etc/rc0.d/S40umountfs /etc/rc0.d/K40umountfs
  sudo mv /etc/rc6.d/S31umountnfs.sh /etc/rc6.d/K31umountnfs.sh
  sudo mv /etc/rc6.d/S40umountfs /etc/rc6.d/K40umountfs
-------------------------------------

I see this bug is regarding the sysvinit package, however that package is not installed on my system (Ubuntu 8.04). I have a feeling the packages upstart and upstart-compat-sysv may require this fix applied also.

Do you think we should create another bug for this issue in the upstart package? Can someone confirm if I am on the right lines here?

[Ralph Wabel (rwabel)]

I've now tried all possibilities I've found on forums and by googling, still I always get this error message and my pc is not shutting/restarting. I need to poweroff my pc.

When I start the script manually, it unmounts my cifs and I can successfully restart or shutdown my pc

Quote:
rwabel@ubuntu:/etc/rc0.d$ ls -ll *umoun*
lrwxrwxrwx 1 root root 22 2008-07-26 14:19 K19umountcifs -> /etc/init.d/umountcifs
lrwxrwxrwx 1 root root 22 2007-01-19 03:20 S14umountnfs.sh -> ../init.d/umountnfs.sh
lrwxrwxrwx 1 root root 18 2007-01-19 03:20 S40umountfs -> ../init.d/umountfs
lrwxrwxrwx 1 root root 20 2007-01-19 03:20 S60umountroot -> ../init.d/umountroot
rwabel@ubuntu:/etc/rc0.d$

Is there any fundamental mistake I did in the ordering?

[EagleDM (eagle-maximopc)]

To eliminate CIFS mounts you only have to do this:

cd /etc/rc6.d
ls -la

Type:
sudo mv S31umountnfs.sh S14umountnfs.sh

then cd to rc6.0

type the same line as before.

That should fix the problem.

That 2 steps is the ONLY thing you have to do after installing ubuntu.

Daniel

On Sat, Jul 26, 2008 at 9:25 AM, Ralph Wabel <email address hidden> wrote:

> I've now tried all possibilities I've found on forums and by googling,
> still I always get this error message and my pc is not
> shutting/restarting. I need to poweroff my pc.
>
> When I start the script manually, it unmounts my cifs and I can
> successfully restart or shutdown my pc
>
> Quote:
> rwabel@ubuntu:/etc/rc0.d$ ls -ll *umoun*
> lrwxrwxrwx 1 root root 22 2008-07-26 14:19 K19umountcifs ->
> /etc/init.d/umountcifs
> lrwxrwxrwx 1 root root 22 2007-01-19 03:20 S14umountnfs.sh ->
> ../init.d/umountnfs.sh
> lrwxrwxrwx 1 root root 18 2007-01-19 03:20 S40umountfs ->
> ../init.d/umountfs
> lrwxrwxrwx 1 root root 20 2007-01-19 03:20 S60umountroot ->
> ../init.d/umountroot
> rwabel@ubuntu:/etc/rc0.d$
>
> Is there any fundamental mistake I did in the ordering?
>
> --
> Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot
> https://bugs.launchpad.net/bugs/42121
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[Ralph Wabel (rwabel)]

I've done that on both rc0.d and rc6.d, it doesn't work for me

[Ralph Wabel (rwabel)]

oo I just see you mentioned S14 and not K14, what's the difference btw?

[Ralph Wabel (rwabel)]

I'm confused :-) I do have S14 and not S14...still it's not working. I also don't konw the difference to S and K

[kritek (kritek)]

On Sun, Jul 27, 2008 at 1:39 PM, Ralph Wabel <email address hidden> wrote:
> I'm confused :-) I do have S14 and not S14...still it's not working. I
> also don't konw the difference to S and K

I moved from S31 or 32, whichever it was, to K13 in rc0.d/ and rc6.d/
Works fine for me, Ubuntu Feisty and Debian testing

-
aRDy Music and Rick Dicaire present:
http://www.ardynet.com
http://www.ardynet.com:9000/ardymusic.ogg.m3u

[Ralph Wabel (rwabel)]

Apparently K is coming first then S. I've it with K12 now, but still not working

lrwxrwxrwx 1 root root 13 2008-05-21 18:18 K01gdm -> ../init.d/gdm
lrwxrwxrwx 1 root root 17 2007-01-19 03:20 K01usplash -> ../init.d/usplash
lrwxrwxrwx 1 root root 22 2008-07-26 14:19 K12umountcifs -> /etc/init.d/umountcifs

When I execute it manually it successfully unmount the cifs.

[Thom Pischke (thom-pischke)]

I can confirm both this bug and the workaround in intrepid beta:

https://bugs.launchpad.net/ubuntu/+source/sysvinit/+bug/42121/comments/26

[Thom Pischke (thom-pischke)]

Correction, the workaround is not working after all. Still hangs on shutdown with the CIFS Server not responding error....

[markba (mark-baaijens)]

Bug is stilll present in Intrepid Beta dated 2008-10-24.

[Scott Severance (scott.severance)]

Re-opening due to reports that the bug is still present.

[Martin Pitt (pitti)]

This is still a bug, but it's not in sysvinit. Renaming S31umountnfs to K31umountnfs might help not because S31umountnfs is called with "start" (everything is called with "stop" for rc0.d and rc6.d), but because it simply happens earlier in the shutdown sequence.

This is most probably a problem with the NetworkManager shutdown timing. This is discussed in bug 211631.

[psypher (psypher246)]

SERIOUSLY?!?!?!?!

more than 2 years and 3 different editions of ubuntu and this bug is STILL!!! present.

I can confirm that daniel ellis' script fixes it.

this makes me VERY worried about the direction Ubuntu is going. Every new version all these fancy new things are added (gvfs pulseaudio... blerg) which cause more issues and have more bugs but none of the old CRITICAL bugs are fixed. Why could there not be just one version of ubuntu where NO new features are added, just for once, and the devs spend their time on squashing bugs. call it the: Incenerated Insect, I dunno just don't add more unstable thats not ready.

So frustrating.

[Scott James Remnant (Canonical) (canonical-scott)]

On Mon, 2008-11-10 at 10:06 +0000, psypher wrote:

> more than 2 years and 3 different editions of ubuntu and this bug is
> STILL!!! present.
>
This bug is marked Fix Released.

Scott
--
Scott James Remnant
<email address hidden>

[toobuntu (toobuntu)]

This is the same issue as reported in Bug #211631. Reinhard Tartler says there that he may not have fully understood the problem, so his Fix Committed here may not be addressing the issue reported here.

See Mathias Gug's comments in Bug #211631:

"The issue discussed here boils down to the fact that network interfaces
can be brought down *before* network filesystems are unmounted thus
leading to a long timeout.

One option proposed was to move the umountnfs script earlier in the
shutdown sequence. Doing leads to the possibility that running processes
still have files opened on the network share. This is the reason why
S31umountnfs.sh is run *after* S20sendsigs. Some packages have their
shutdown scripts set too early in the boot sequence. These should be fix
them rather then moving the umountnfs script earlier in the shutdown
sequence.

Another option suggested was to use the ifdown.d infrastructure. That
means writing a script that is able to unmount network filesystems
according to the interface been brought down. However the script should
not unmount the remote filesystems when *a* network interface goes down
but rather unmount them when the *corresponding* network interface goes
down.

What *should* be happening in all cases is that the network route is
gone. Trying to send to it will return a no-route-to-host which can be
detected and handled by the kernel. So either the route isn't being torn
out when it should be, and we should fix that; or the cifs driver
doesn't handle no-route-to-host, and we should fix that. Adjusting the
timeouts or moving/adding init scripts shouldn't matter at all."

From my experience, a reliable *workaround* is to use Sander Marechal's umountcifs script (attached), and symlink as K15umountcifs (i.e. the key is that it is called prior to K16dhcdbd).

$ sudo ln -s /etc/init.d/umountcifs /etc/rc0.d/K15umountcifs
$ sudo ln -s /etc/init.d/umountcifs /etc/rc6.d/K15umountcifs

[psypher (psypher246)]

Scott, what does fix released mean when this bug is still present in the latest version of intrepid?

[Scott James Remnant (Canonical) (canonical-scott)]

On Tue, 2008-11-11 at 05:37 +0000, psypher wrote:

> Scott, what does fix released mean when this bug is still present in the
> latest version of intrepid?
>
It means that while you might be seeing similar symptoms, it is not the
same as this bug - and you should file a new one.

Scott
--
Scott James Remnant
<email address hidden>

[psypher (psypher246)]

ok but what is the exact problem then, since the symptoms are exactly the same and have not changed since gutsy. how do i figure out where the problem is when the same solution that has worked for 3 different editions of ubuntu is still the solution?

[Felix (apoapo)]

I've got the same problem as psypher here on a Intrepid machine. What shall we do? The time out isn't fixed by the umountcifs script called at K01, K12,K19 or K20. (At least not *always*).

@psypher Did you already file a new report?

[fermulator (fermulator)]

Confirmed. This bug still exists in the final release of Intrepid Ibex (8.10).

Followed the helpful tips from: http://www.jejik.com/articles/2007/07/automatically_mounting_and_unmounting_samba_windows_shares_with_cifs/ to resolve.

However, of course, this needs to be fixed in the main so that users need to apply this fix manually.

[Jackflap (deriziotis)]

Is this definitely the same bug as #153444 ?

This claims to have a fix released, but I've also just reproduced it in Intrepid. Perhaps they are separate bugs, and #153444 is really the one I'm experiencing as it contains the message that I'm seeing?

[Patola (patola)]

Bug still present in Jaunty Jackalope. It doesn't unmount my CIFS shares when I am shutting down my laptop.

[keppy (josh-joshandmeg)]

Confirming still present in UNR 9.04. If it helps any, none of the proposed fixes work for me. Moving S31umountnfs.sh to K31umountnfs.sh or S14umountnfs.sh does not work. Please let me know if any info more info is needed.

[fermulator (fermulator)]

I'm running Jaunty 9.04 now, and I just followed what I followed for intrepid (https://bugs.launchpad.net/ubuntu/+source/sysvinit/+bug/42121/comments/44), and the "manual hack/fix" works fine still for me.

[fermulator (fermulator)]

NOTE: As keppy has mentioned, I can confirm that the "bug" is still present for 9.04.

[koshari (holto2go)]

i get "cifs vfs server not responding" on shutdown of my system with cifs lines in fstab, in 9.04

[vnevoa (vasco-nevoa)]

Problem still exists in Jaunty (9.04) with CIFS mounts.
Renaming the links from "S" to "K" solved it for me (maintaining the same order number).

[emil.s (emil.s)]

Confirming here to.
Release: 9.04

[Hilikus (hilikus)]

i'm using 9.04 and i get the same error
however the bug is still marked as Fix released and the original bug brought VPN into account. i don't use VPN. just a samba shared mounted on /etc/fstab from another ubuntu box
should we open a new bug report?

[Hilikus (hilikus)]

This has not been fixed

[Sven-Hendrik Haase (svenstaro)]

I'd just like to confirm that this is still an issue.

[tekstr1der (tekstr1der)]

This bug is still affecting Karmic Alpha 4 with network-manager 0.8~a~git.20090804t185522.4bab334-0ubuntu1

[Alexis de Lattre (alexis-via)]

I don't get the message "CIFS VFS server not responding" on shutdown any more on the Ubuntu PC that I have upgraded to Karmic (and which have cifs lines in /etc/fstab). So the bug seems to be fixed...

[markba (mark-baaijens)]

I did two clean installs with 9.10 karmic. In one of them, the problem is solved, in the other one, the problem is still present. Since they are both clean installs (no upgrades), it's ver odd. Maybe it has something to do with the sequence in which fstab is modified and the updates are downloaded and installed.

So the bug seems to be solved on some karmic systems, but not on all.

[neuromancer (neuromancer)]

In my fresh installation of Ubuntu Karmic the message isn't show, so i think that was fixed.
But comment #58 need more investigation...

[blabbermouth (luka-batistic)]

Fresh install of Karmic and the issue is still here.

Running on a wireless network.

[Igor A Tarasov (dicr)]

This bug affect me too since I using wireless connections and permanent cifs mounts in fstab.
But I think we need to fix it in upstart, not init.d.

[EagleDM (eagle-maximopc)]

Strange

Upgraded to Karmic and the issue is Gone for me.

Now I have a new bug never present before Karmic, From time to time the cifs
responds something about memory error and I have to restart the Windows 7
server :(

On Fri, Jan 8, 2010 at 7:51 AM, Dicr <email address hidden> wrote:

> This bug affect me too since I using wireless connections and permanent
> cifs mounts in fstab.
> But I think we need to fix it in upstart, not init.d.
>
> --
> Symlinks for umountnfs / sendsigs wrong: hang on shutdown / reboot
> https://bugs.launchpad.net/bugs/42121
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[David Tombs (dgtombs)]

This particular issue has been fixed. If you still experience the issue, it's probably bug 211631. Please continue discussion there.