fsck doesn't update system info on login

Bug #692355 reported by geeknik on 2010-12-19
220
This bug affects 45 people
Affects Status Importance Assigned to Milestone
update-notifier (Ubuntu)
Medium
Canonical Foundations Team

Bug Description

Every time I login to my Ubuntu 11.04 Alpha 1 (server), I see this message:

*** /dev/sda5 will be checked for errors at next reboot ***

I have done both the 'touch /forcefsck' (file is gone after reboot) and I have done 'shutdown -rF now' and yet the message is still displayed. I checked my boot.log and fsck did run:

fsck from util-linux-ng 2.17.2
fsck from util-linux-ng 2.17.2
/dev/mapper/sh*tbox1-root: clean, 103845/4669440 files, 1179707/18669568 blocks
/dev/sda5: clean, 245/124496 files, 83515/248976 blocks

Since fsck did run successfully, the '*** /dev/sda5 will be checked for errors at next reboot ***' message should not appear again unless the system has determined the filesystem is dirty. I seriously doubt that this not very old drive is having this many problems, especially since the OS runs 24/7 with no problems besides this.

lsb_release -rd:
Description: Ubuntu natty (development branch)
Release: 11.04

uname -a output:
Linux node1 2.6.37-10-server #24-Ubuntu SMP Thu Dec 16 18:10:40 UTC 2010 x86_64 GNU/Linux

Related branches

tags: added: alpha error fsck login natty
Fabio Marconi (fabiomarconi) wrote :

Assign to e2fsprogs as it seems a fsck issue

affects: ubuntu → e2fsprogs (Ubuntu)
Theodore Ts'o (tytso) wrote :

No, I don't think it's an e2fsprogs issue. It looks like it's an init scripts issue. Fsck should always run on boot, regardless of whether or not /forcefsck is set. What forcefsck should do is set the -f flag, to request a forced check even though the file system appears clean.

jgreenso (james-green-mjog) wrote :

I'm seeing this on an upgraded Natty vmware instance.

Relevant fstab entry:
UUID=cbdda93e-2d89-47be-9bee-ce24bf1376ce /boot ext3 relatime 0 2

touched /boot/forcefsck and rebooted - file still there.

boot.log says:
/sbin/fsck.xfs: XFS file system.
/sbin/fsck.xfs: XFS file system.
/dev/sda1: clean, 236/36144 files, 52572/144552 blocks

jamesg@ubuntu:~$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/myvg-system
                       10G 7.7G 2.4G 77% /
none 242M 180K 242M 1% /dev
none 249M 0 249M 0% /dev/shm
none 249M 356K 248M 1% /var/run
none 249M 0 249M 0% /var/lock
/dev/sda1 137M 47M 83M 37% /boot
/dev/mapper/myvg-mysqllv
                       30G 1.5G 29G 5% /var/lib/mysql

Paul McEnery (pmcenery) wrote :

I was affected by this and bug and discovered that it appears to be caused when the file /var/lib/update-notifier/fsck-at-reboot has a datestamp in the future. This will never rectify itself unless the actual date passes the datestamp on the file.

The attached patch fixes the problem.

Paul McEnery (pmcenery) wrote :

After applying the above patch; either reboot, or regenerate /var/lib/update-notifier/fsck-at-boot as follows:

$ sudo -s
# run-parts --lsbsysinit /etc/update-motd.d > /var/run/motd

tags: added: patch
Changed in e2fsprogs (Ubuntu):
status: New → Confirmed
Brian Murray (brian-murray) wrote :

It's actually update-notifier that provides update-motd-fsck-at-reboot.

affects: e2fsprogs (Ubuntu) → update-notifier (Ubuntu)
Changed in update-notifier (Ubuntu):
importance: Undecided → Medium
status: Confirmed → Triaged
Changed in update-notifier (Ubuntu):
assignee: nobody → Canonical Foundations Team (canonical-foundations)
Michael Vogt (mvo) on 2012-01-09
Changed in update-notifier (Ubuntu):
status: Triaged → In Progress
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package update-notifier - 0.118ubuntu1

---------------
update-notifier (0.118ubuntu1) precise; urgency=low

  * fix "fsck doesn't update system info on login" LP: #692355
    Thanks to Paul McEnery
 -- Michael Vogt <email address hidden> Mon, 09 Jan 2012 12:58:55 +0100

Changed in update-notifier (Ubuntu):
status: In Progress → Fix Released
Nicholas Lee (nicholas-lee-n) wrote :

This bug still seems to exist in Ubuntu 12.04 LTS....

Welcome to Ubuntu 12.04 LTS (GNU/Linux 3.2.0-24-generic-pae i686)

 * Documentation: https://help.ubuntu.com/

*** /dev/sdb1 will be checked for errors at next reboot ***
*** /dev/sdc1 will be checked for errors at next reboot ***

Last login: Thu May 3 09:14:20 2012

Steve Langasek (vorlon) wrote :

Looks like the bug is here in /usr/lib/update-notifier/update-motd-fsck-at-reboot:

# check time when we did the last check
stamp="/var/lib/update-notifier/fsck-at-reboot"
if [ -e "$stamp" ]; then
    stampt=$(stat -c %Y $stamp)
else
    stampt=0
fi

now=$(date +%s)
if [ $(($stampt + 3600)) -lt $now ] || [ $stampt -gt $now ]; then
    #echo $stampt $now need update
        NEEDS_FSCK_CHECK=yes
fi

If you log in on console triggering a refresh of the file, then reboot and get the fsck, then log in again once the system finishes booting, there will be a current stamp file and the check will not be re-run.

We should probably compare the file's timestamp with the uptime, and set NEEDS_FSCK_CHECK=yes if it's older than the last reboot.

Changed in update-notifier (Ubuntu):
status: Fix Released → Triaged
Berend Dekens (cyberwizzard) wrote :

On my media computer, I am running the latest updates for Ubuntu 11.10 and this fix is not released. Can we expect this to be back-ported to 11.10?

Chris Cote (1fpl) wrote :

I performed a remote server package upgrade (over SSH) to 12.04 from 10.04. I am effected by this bug too. I have no direct console access to the machine since it's 50+ miles away, only SSH.

Please take this into consideration when you troubleshoot this problem.

Thanks.

Server 12.04. It is an upgrade from 11.10, which was installed via CD. I have tried setting FSCKFIX to "yes", no change. The /var/lib/update-notifier/fsck-at-reboot file has a date/time stamp set to May 22.
Touched /boot/forcefsck and rebooted - file still there.

will be checked for errors at next reboot *** still comes up on motd.

Curtis Bangert (codecae) wrote :

I had this issue with my logical filesystems after upgrading to 12.04 as well... fsck was mysteriously skipping filesystems at boot.

I fixed it by making sure each entry in /etc/fstab had the option 'errors=remount-ro', executed 'touch /forcefsck', and rebooted... To my delight, fsck ran and updated the mount count and last check dates.

Unfortunately, however, after logging in I still had the notification that the filesystems would be checked at reboot. I manually rm'd /var/lib/update-notifier/fsck-at-reboot and logged out and back in again. The file was regenerated by the script, only this time, without any notifications.

Given the nature of this oversight, I'm sure I would've likely encountered it whether or not I had upgraded to 12.04... likely a coincidence.

hello,
(Ubuntu 11.10 "server")
it is not a big issue but it is a bit scary at first but the filesystem is not compromised!

tags: added: precise
Thorsten Rhau (thorre) wrote :

Upgraded from 10.04.4 to 12.04.1 yesterday and now i am getting these error messages as well. I have done an manual fsck but that does not help.

System seems fine.

thorre@ida:~$ uname -a
Linux ida 3.2.0-30-generic #48-Ubuntu SMP Fri Aug 24 16:52:48 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux
thorre@ida:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 12.04.1 LTS
Release: 12.04
Codename: precise
thorre@ida:~$

GouZi (jalliot) wrote :

Still there with:

$ uname -a
Linux merlin 3.2.0-32-generic #51-Ubuntu SMP Wed Sep 26 21:33:09 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 12.04.1 LTS
Release: 12.04
Codename: precise

Christoph Lühr (cluehr) wrote :

I have the same issue on Ubuntu 12.10.

[cluehr@xxxx: /] $ uname -a
Linux D-TXL-00429084 3.5.0-18-generic #29-Ubuntu SMP Fri Oct 19 10:26:51 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux
[cluehr@xxxx: /] $ lsb_release -a
LSB Version: core-2.0-amd64:core-2.0-noarch:core-3.0-amd64:core-3.0-noarch:core-3.1-amd64:core-3.1-noarch:core-3.2-amd64:core-3.2-noarch:core-4.0-amd64:core-4.0-noarch
Distributor ID: Ubuntu
Description: Ubuntu 12.10
Release: 12.10
Codename: quantal

Jeremy LaCroix (j-jlacroix) wrote :

Same here with Ubuntu Server 12.04.

I tried rm'ing the file, but it did not fix it. I didn't run 'touch /forcefsck' because I am using LVM with a RAID1 config and not sure if that would mess everything up.

Andreas Ntaflos (daff) wrote :

Still a problem on 12.04.1. Removing /var/lib/update-notifier/fsck-at-reboot and logging out, then back in works. The /etc/motd is updated and no longer contains this misleading and wrong message about pending fscks.

Mark Wormgoor (markwo) wrote :

Having the same problem here on 12.10 with all updates installed. I fixed my fstab (one filesystem didn't have the check set) and did a touch /forcefsck. After the reboot, the message is still there, stating that my filesystems will be checked on reboot. However, I checked my filesystems, and they were both checked when I rebooted just now:
Last checked: Tue Jan 22 08:26:15 2013

Stephen (1012web) wrote :

If you want to remove the message from motd without rebooting here is the command to delete the file that has the problem message cached and fully regenerate the motd:

sudo bash -c 'rm /var/lib/update-notifier/fsck-at-reboot && for file in /etc/update-motd.d/*; do $file; done > /var/run/motd' && cat /etc/motd

tags: added: rls-s-tracking
maletin (maletin) wrote :

# dumpe2fs -h /dev/sdb1 | grep 'ount count'
dumpe2fs 1.42 (29-Nov-2011)
Mount count: 23
Maximum mount count: 20

~$ sudo sh -x /usr/lib/update-notifier/update-motd-fsck-at-reboot
[...]
Mount count: 23
Maximum mount count: 20
[...]
+ next_check_date= Tue Jul 23 09:21:20 2013
+ [ -z 0 ]
+ date -d Tue Jul 23 09:21:20 2013 +%s
+ next_check_tstamp=1374564080
+ check_occur=
+ [ 20 -gt 0 -a 23 -ge 20 ]
+ check_occur=yes
+ [ 15552000 -gt 0 -a 1374564080 -lt 1368184556 ]
+ [ -n yes ]
+ check_occur_any=yes
+ echo *** /dev/sdb1 will be checked for errors at next reboot ***
[...]

maletin (maletin) wrote :

$ grep 'home' /etc/fstab
# /home was on /dev/sda4 during installation
#UUID=cf9c1b11-5604-4b67-80bc-e9a2a082b220 /home ext4 defaults 0 2
/dev/sdb1 /home auto defaults 0 2

Steve Langasek (vorlon) wrote :

> # dumpe2fs -h /dev/sdb1 | grep 'ount count'
> dumpe2fs 1.42 (29-Nov-2011)
> Mount count: 23
> Maximum mount count: 20

Which means this is working exactly as designed.

maletin (maletin) wrote :

Steve Langasek (vorlon) wrote on 2013-05-11: #24

> Which means this is working exactly as designed.

correct. but why does my vmware not start a fsck after reboot (even with /forcefsck)?

On Mon, May 13, 2013 at 09:27:06AM -0000, maletin wrote:
> Steve Langasek (vorlon) wrote on 2013-05-11: #24

> > Which means this is working exactly as designed.

> correct. but why does my vmware not start a fsck after reboot (even with
> /forcefsck)?

I don't know, but that seems to be describing a different issue than this
bug.

Changed in update-notifier (Ubuntu):
status: Triaged → In Progress
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package update-notifier - 0.136

---------------
update-notifier (0.136) saucy; urgency=low

  [ Steve Langasek ]
  * If the /var/lib/update-notifier/fsck-at-reboot stamp file is older than
    the current uptime, any information about fscking at "next boot" is
    definitely stale - so force a refresh. LP: #692355.

  [ Brian Murray ]
  * Remove support for choosing not to auto-launch applications (LP: #947008)
  * Remove reboot required notification as update-manager itself indicates
    when a reboot is required and it only appeared when auto-launch was set to
    false.
  * Move update-notifier checking for CDs with packages into a new binary
    distro-cd-updater
  * Add an upstart user session job that runs upon CD insertion
  * Remove sections of uevent related to hp firmware and replace them with an
    upstart user session job
  * Modified avahi disabled notification to run as an upstart user session
  * Modified new release check to run as an upstart user session
  * Added in a weekly cron job to check for a new release of ubuntu
  * Added an upstart user session job to check for crashes
  * Ensure we ask before launching a pkexec dialog for an apport crash
 -- Brian Murray <email address hidden> Tue, 04 Jun 2013 10:08:21 -0700

Changed in update-notifier (Ubuntu):
status: In Progress → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Duplicates of this bug

Other bug subscribers