Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #157525 > unrolled thread

Fsck errors at boot, but boots anyway

Back to article view | Back to linux.debian.user

Contents

  Fsck errors at boot, but boots anyway Sam Smith <debian@net153.net> - 2015-09-07 20:40 +0200
    Re: Fsck errors at boot, but boots anyway Linux 4 Bene <linux4bene@telenet.be> - 2015-09-09 15:20 +0200
      Re: Fsck errors at boot, but boots anyway Sam Smith <debian@net153.net> - 2015-09-12 04:00 +0200
    Re: Fsck errors at boot, but boots anyway David Wright <deblis@lionunicorn.co.uk> - 2015-09-09 20:20 +0200
      Re: Fsck errors at boot, but boots anyway Sam Smith <debian@net153.net> - 2015-09-12 04:10 +0200
    Re: Fsck errors at boot, but boots anyway Sam Smith <debian@net153.net> - 2015-09-12 21:00 +0200

#157525 — Fsck errors at boot, but boots anyway

Hello,

I've been getting some strange fsck errors for over a year now on my 
Debian 8 workstation. See photo: http://tinyurl.com/otzxebh Ignore the 
softreset errors. I can't really remember when the issue started, but it 
was sometime after I upgraded to using systemd. I kept thinking it was a 
bug and would fix itself, but apparently not.

As the photo shows /dev/md_d1 (a partitionable raid5 array) is brought 
online, hangs for about 10 seconds, displays fsck: error 2 while 
executing fsck.ext3 for /dev/md_d1p1
followed by error code 8 (operational error)
Then the systemd-fsck runs next and completes fine and the system boots. 
So is this some conflict between util-linux and systemd? All my 
filesystems are ext4 by the way.

Another issue is when a fsck actually does run, systemd doesn't display 
the output, the screen just hangs for 20 minutes or so which to the 
unknown user, they may reboot the machine and screw stuff up. Is that 
just the nature of systemd or is there a fix?

I tried googling all these issues but really couldn't get anywhere.

Thanks,
Sam

[toc] | [next] | [standalone]

#157627

Op Mon, 07 Sep 2015 13:23:03 -0500, schreef Sam Smith:

> Hello,
> 
> I've been getting some strange fsck errors for over a year now on my
> Debian 8 workstation. See photo: http://tinyurl.com/otzxebh Ignore the
> softreset errors. I can't really remember when the issue started, but it
> was sometime after I upgraded to using systemd. I kept thinking it was a
> bug and would fix itself, but apparently not.
> 
> As the photo shows /dev/md_d1 (a partitionable raid5 array) is brought
> online, hangs for about 10 seconds, displays fsck: error 2 while
> executing fsck.ext3 for /dev/md_d1p1 followed by error code 8
> (operational error)
> Then the systemd-fsck runs next and completes fine and the system boots.
> So is this some conflict between util-linux and systemd? All my
> filesystems are ext4 by the way.
> 
> Another issue is when a fsck actually does run, systemd doesn't display
> the output, the screen just hangs for 20 minutes or so which to the
> unknown user, they may reboot the machine and screw stuff up. Is that
> just the nature of systemd or is there a fix?
> 
> I tried googling all these issues but really couldn't get anywhere.
> 
> Thanks,
> Sam

Are you sure your disk(s) are ok?
Have you run a manual check? Maybe boot from a live cd and do the check 
from there. 1 of the disks may be getting faulty. I would first check 
that before chasing the possibility of systemd causing this.

Cheers,
Bene

[toc] | [prev] | [next] | [standalone]

#157781

On 09/09/2015 08:10 AM, Linux 4 Bene wrote:
> Op Mon, 07 Sep 2015 13:23:03 -0500, schreef Sam Smith:
>
>> Hello,
>>
>> I've been getting some strange fsck errors for over a year now on my
>> Debian 8 workstation. See photo: http://tinyurl.com/otzxebh Ignore the
>> softreset errors. I can't really remember when the issue started, but it
>> was sometime after I upgraded to using systemd. I kept thinking it was a
>> bug and would fix itself, but apparently not.
>>
>> As the photo shows /dev/md_d1 (a partitionable raid5 array) is brought
>> online, hangs for about 10 seconds, displays fsck: error 2 while
>> executing fsck.ext3 for /dev/md_d1p1 followed by error code 8
>> (operational error)
>> Then the systemd-fsck runs next and completes fine and the system boots.
>> So is this some conflict between util-linux and systemd? All my
>> filesystems are ext4 by the way.
>>
>> Another issue is when a fsck actually does run, systemd doesn't display
>> the output, the screen just hangs for 20 minutes or so which to the
>> unknown user, they may reboot the machine and screw stuff up. Is that
>> just the nature of systemd or is there a fix?
>>
>> I tried googling all these issues but really couldn't get anywhere.
>>
>> Thanks,
>> Sam
>
> Are you sure your disk(s) are ok?
> Have you run a manual check? Maybe boot from a live cd and do the check
> from there. 1 of the disks may be getting faulty. I would first check
> that before chasing the possibility of systemd causing this.
>
> Cheers,
> Bene
>


Disks are fine, I have scripts that monitor the smart stats. I am not 
trying to blame systemd, just seems like something tries to run 
'fsck.ext3' on an ext4 fs and bombs out, followed clearly by systemd 
running fsck.ext4 successfully.

Thanks,
Sam

[toc] | [prev] | [next] | [standalone]

#157643

Quoting Sam Smith (debian@net153.net):

> Another issue is when a fsck actually does run, systemd doesn't
> display the output, the screen just hangs for 20 minutes or so which
> to the unknown user, they may reboot the machine and screw stuff up.
> Is that just the nature of systemd or is there a fix?

My systems typically have three partitions mounted, and there's no
problem seeing fsck in progress on the non-root ones. Getting the
root partition to fsck at all (on demand) was more of a challenge
because it has to be done from the initramfs stage.

Looking at scripts/functions there, the kernel parameter needed to
force a full check is forcefsck (not fsck.mode=force), but the
parameter to give you a progress spinner seems to be disabled:

        # spinner="-C" -- only if on an interactive terminal
        spinner=""

I presume I see progress for non-root partitions because
/etc/init.d/checkroot.sh (or whatever systemd derives from it) gets
this right:

                spinner="-C"
                case "$TERM" in
                  dumb|network|unknown|"")
                        spinner="" ;;
                esac

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#157782

On 09/09/2015 01:11 PM, David Wright wrote:
> Quoting Sam Smith (debian@net153.net):
>
>> Another issue is when a fsck actually does run, systemd doesn't
>> display the output, the screen just hangs for 20 minutes or so which
>> to the unknown user, they may reboot the machine and screw stuff up.
>> Is that just the nature of systemd or is there a fix?
>
> My systems typically have three partitions mounted, and there's no
> problem seeing fsck in progress on the non-root ones. Getting the
> root partition to fsck at all (on demand) was more of a challenge
> because it has to be done from the initramfs stage.
>
> Looking at scripts/functions there, the kernel parameter needed to
> force a full check is forcefsck (not fsck.mode=force), but the
> parameter to give you a progress spinner seems to be disabled:
>
>          # spinner="-C" -- only if on an interactive terminal
>          spinner=""
>
> I presume I see progress for non-root partitions because
> /etc/init.d/checkroot.sh (or whatever systemd derives from it) gets
> this right:
>
>                  spinner="-C"
>                  case "$TERM" in
>                    dumb|network|unknown|"")
>                          spinner="" ;;
>                  esac
>
> Cheers,
> David.
>

I was thinking that checkroot.sh was obsolete with systemd?
At least /var/log/fsck/checkroot has a last mod time of Jun 1, 2014..

Thanks,
Sam

[toc] | [prev] | [next] | [standalone]

#157804

On 09/07/2015 01:23 PM, Sam Smith wrote:
> Hello,
>
> I've been getting some strange fsck errors for over a year now on my
> Debian 8 workstation. See photo: http://tinyurl.com/otzxebh Ignore the
> softreset errors. I can't really remember when the issue started, but it
> was sometime after I upgraded to using systemd. I kept thinking it was a
> bug and would fix itself, but apparently not.
>
> As the photo shows /dev/md_d1 (a partitionable raid5 array) is brought
> online, hangs for about 10 seconds, displays fsck: error 2 while
> executing fsck.ext3 for /dev/md_d1p1
> followed by error code 8 (operational error)
> Then the systemd-fsck runs next and completes fine and the system boots.
> So is this some conflict between util-linux and systemd? All my
> filesystems are ext4 by the way.
>
> Another issue is when a fsck actually does run, systemd doesn't display
> the output, the screen just hangs for 20 minutes or so which to the
> unknown user, they may reboot the machine and screw stuff up. Is that
> just the nature of systemd or is there a fix?
>
> I tried googling all these issues but really couldn't get anywhere.
>
> Thanks,
> Sam
>


Ok, found my issue. I had rootfstype=ext4 as a kernel parameter, yet the 
FS was not converted fully to ext4. Must have been like that for awhile, 
because I converted all my FS's to ext4 back in 2009.

--Sam

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web