5

My MOTD is showing:

*** /dev/md2 will be checked for errors at next reboot ***
*** /dev/md1 will be checked for errors at next reboot ***
*** /dev/md3 will be checked for errors at next reboot ***

However, a standard sudo shutdown -r now does not appear to check the disks and the message remains.

My guess is that the partitions 'should' be checked and that they are not being... so how can I get Ubuntu to check the partitions at reboot and keep things nice and safe?

Update - this is the output of cat /etc/fstab

proc /proc proc defaults 0 0
none /dev/pts devpts gid=5,mode=620 0 0
/dev/md/0 none swap sw 0 0
/dev/md/1 /boot ext3 defaults 0 0
/dev/md/2 / ext4 defaults 0 0
/dev/md/3 /home ext4 defaults 0 0

Update 2 - One message gone...

Using @christianwolff's suggestions...

sudo rm /var/lib/update-notifier/fsck-at-reboot 
sudo touch /forcefsck
sudo shutdown -r now

And now the motd is down to

*** /dev/md1 will be checked for errors at next reboot ***
*** /dev/md3 will be checked for errors at next reboot ***

So 'md2(ext4)' has been checked and the message updated.

jamesc
  • 181
  • 1
  • 2
  • 11

2 Answers2

5

That's a known bug. Your partitions seem to be fine. It's just the MOTD which is "out of date". Just delete the file /var/lib/update-notifier/fsck-at-reboot and it's gone. Had the same issue and this fixed my problem.

Here are some details about it: https://bugs.launchpad.net/ubuntu/+source/update-notifier/+bug/692355/comments/9

chrw
  • 1,071
  • 5
  • 11
  • The `fsck-at-reboot` file was recreated for me at reboot - so a bit more prodding required. But thanks for the pointer to the bug. – jamesc Oct 14 '13 at 17:05
  • And I assume the file is still not empty? The file itself exists permanently but should be empty. For ext3/ext4 filesystems you can force a check using `touch /forcefsck`. Which filesystem type are you using? – chrw Oct 15 '13 at 05:53
  • In the list above they file systems are a mix of ext3 and ext4 - I've updated the question. – jamesc Oct 17 '13 at 10:49
  • And yes, the file is still not empty - so it was recreated with the messages. I'll see if I can force a check now. – jamesc Oct 17 '13 at 10:52
  • At least one filesystem is down! ;-) Have you tried to do a fsck on the other two? I think you can umount /boot and /home while your system is running. You may want to prevent other users to connect while you're checking. The command should be `fsck` oder `fsck.ext3`/`fsck.ext4`. – chrw Oct 18 '13 at 08:41
5

Thanks to @cwo's suggestions I've been able to clear out the problem. This is a server on Hetzner's system - it's configured in their default Ubuntu way.

Unmount /boot and /home and check them manually...

sudo umount /boot
sudo fsck /boot

It does appear that the file systems weren't being checked at all at reboot time...

/dev/md1 has gone 320 days without being checked, check forced.

Then mount that back again when done.

sudo mount /boot

I repeated the same process for /home, however had to ensure that that no services were running on the filesystem which would prevent the umount.

Once completed and remounted, I cleaned out the message file, marked the system for a file check and rebooted.

sudo rm /var/lib/update-notifier/fsck-at-reboot
sudo touch /forcefsck
sudo shutdown -r now

This brought the system back up with all file systems checked and the MOTD cleaned.

jamesc
  • 181
  • 1
  • 2
  • 11