Mint9 Boot Stalls on fsck

[altair4]

Description of Problem
---------------------------------
Whenever the system requires an fsck at boot the boot process stalls. Recovery requires one to press "c" to cancel fsck and continue the booting process.

Steps to Reproduce
----------------------------
Wait until the next automatic fsck
Or: sudo touch /forcefsck

It was reported elsewhere that removing 'quiet splash' from the boot options was a workaround but that does not work and all attachments and descriptions in this bug report reflect both those options removed.

My Results
---------------
* I let this run on one occasion for 28 minutes without resulting in a boot.
* I have Mint9 installed on two different boxes and it happens to both.
* If I immediately reboot after a forcefsck all partitions are declared clean
* If I force an fsck the stall occurs again

My attachments should make this more clear.

Ubuntu Bug Report:
-----------------------------
https://bugs.launchpad.net/Ubuntu/+source/mountall/+bug/571707?comments=all
This bug report is apparently closed (?) because it was fixed by version 2.15 of package: mountall.
I have version 2.15 installed. It's implied that package: Plymouth is also involved in this problem so I have included the versions of both packages that I am currently running.

Relevant version levels:
----------------------------------
apt version mountall: 2.15
apt version plymouth: 0.8.2-2-1mint3

Mint Forum Post:
-----------------------
http://forums.linuxmint.com/viewtopic.php?f=46&t=48011

[Borja López Soilán (NeoPolus) (borjals)]

There is more people reporting similar problems: https://bugs.launchpad.net/linuxmint/+bug/580030

[Borja López Soilán (NeoPolus) (borjals)]

I just reproduced it on my Mint 9.

Being a virtual machine (VirtualBox) the fsck took less than 10 seconds on the first too forced fscks (sudo touch /forcefsck; sudo reboot).

The third time it got stuck on the "Press C to cancel all checks currently in progress" screen after displaying the 98% (I think it got stuck just after completing the fsck), I pressed ESC to see the text mode messages and this was displayed:

-----------------
"(process: 216) GLib-WARNING **: getpwuid_r(): falied due to unknown user id (0)

fsck from util-linux-ng 2.17.2
/dev/sda2: 134297/2498560 files (0.2% non-contiguous), 911478/9985280 blocks
-------------------

Pressing ESC twice again (to go to the graphic screen and back to text mode) would cause another fsck message to be printed.
It remained stuck for several minutes until I pressed "C".

I tried doing the forced fsck a fourth and fifth time: it did got stuck again on the 5th. So the overall is 2 of each 5 times.

apt version mountall: 2.15
apt version plymouth: 0.8.2-2-1mint3

[bwjpg (esthiaque)]

Have this exact problem.

I'm running Mint 64 if that makes a difference.

[martin welsh (martin-mhwelsh)]

Displayed message;
"Your disc drives are being checked for errors, this may take some time
Checking disc 1 of 6 (24% complete)
Press C to cancel all checks currently in progress."

There is no drive activity light shining or flickering.
There are only two hard disks.
Raid is not fitted.

Hope this helps.

martin welsh

[tlu (thomas-ludwig-gmx)]

I can confirm that my system (Mint 9 32bit) is also affected. I'm somewhat astonished that the importance is set to "undecided". Mint is meant to be a user-friendly distro, and I'm afraid that Linux newbies in particular will be highly irritated by this bug. That's why this bug should be fixed as soon as possible to prevent the reputation of Mint from suffering damage.

Editing fstab to disable the checks is not really a solution.

[Peter Hohensasser (peter-hohensasser)]

same here - also after cancelling it just sits there doing nothing instead of booting

[tux-sven (sven.b)]

It happened on a 6 years old HP desktop I yesterday installed for a neighbor who never had seen Linux. I simply changed to something that works well, LM8 Helena.. :)

[freedomrun (freedomrun)]

After every 30th boot there is fsck (this is normal) .. I do let it finish but after check is complete it stalls (this isn`t normal)
---
mint is installed on externall USB HD

[Crom (asfmaili)]

Same problem here. I have to press 'c' or ctrl-alt-del to continue. There are no errors. For someone who doesn't know these things (like my girlfriend) this effectively prevents you from using the computer.

I'm using Linux Mint 9 64bit on Asus X70AC Laptop. Didn't have this problem on Ubuntu 10.

[tux-sven (sven.b)]

It still happens on my Acer Aspire One. It looks like "arrow right" brakes the stop and continues the login. (My machine is updated to newest packages) LM9 32bit Gnome. Installed to HD.
I could not see this bug clearly before, as a result of my other very serious start problems. (see bug/580726)

[Chris Morley (c-morley)]

Confirmed same problem on Mint 9 x64 Edition.

Quite often a reboot will force another check and I get caught in a "loop" of disk checks.

Chris.

[lannatwin (lannatwin)]

Same problem on Mint 9 x64 Edition. SATA drive.

[steve w (watzo99)]

I have same problem on my Asus eee netbook (1000HE) running Linux Mint 9 (32bit). It hangs during boot, but "C" (or anything else for that matter) will not interrupt it. Had to power off twice to get machine booted.

[Alexey Loukianov (lexa2)]

Same problem here. Fresh install of Mint 8 x86 (32bit) on ASUStek F7Z notebook.
Partitions layout is separate partition for boot, root and home.
I was forced to disable autofscks in the fstab for now, as this notebook will be used by unexperienced people who wouldn't be able to troubleshoot bootup problems with stalled fsck.

[rlindau (rlindau)]

I am having the same problem - Clean install of Mint 9, four partitions (root, swap, home, backup)

[Richard Postlewait (sandman6471)]

I have same problem, on my desktop and on my Acer Aspire One Netbook. I have also posted a bug report on this matter. Bug Report #606588.

[Mr. Aljoriz Dublin (aljoriz)]

Mint 9 LXDE.. hangs after sudo touch /forcefsck. It checks the drive but after 100% it just freezes.

[Mr. Aljoriz Dublin (aljoriz)]

Test Machine Asus 1005ha Mint9 XFCE rc. fsck scan is very slow and still hangs w/o reaching 100%

[David Gregg (davegregg)]

HP DV2 laptop, Isadora 64-bit, Mountall v. 2.15, more bug details and hardware specs on my forum post related to this bug: http://forums.linuxmint.com/viewtopic.php?f=46&t=48011&sid=91cb67e36129fed59699979e639b7558&start=20#p299667

[Mr. Aljoriz Dublin (aljoriz)]

merwiz79 at http://forums.linuxmint.com/viewtopic.php?f=110&t=48740&p=299997#p299997 says he haven't installed the Level3 updates and fsck runs well on his system. Currently re-installing to check his theory

[Mr. Aljoriz Dublin (aljoriz)]

Re-installed. It would seem that the Lv3 updates are whats causing the problems.

Before updating:
After the fresh install goto mintupdate, select Edit>Preference uncheck all lv3 updates, apply. I rebooted and did a sudo touch /fsck... now it worked NO MORE PROBLEMS

[Clement Lefebvre (clementlefebvre)]

Ok, it seems one of the lvl3 updates is causing a regression here. We need to identify which one so we can move its level to 5. This is also happening upstream obviously, so I'll be trying to mark the bugs related to this issue as duplicates.

[Mr. Aljoriz Dublin (aljoriz)]

Kindly mark bug #607507 as its a duplicate of this bug.

[Mr. Aljoriz Dublin (aljoriz)]

Weird did a clean install of LXDE after which I did the "sudo touch /force fsck" to check fsck still hangs even w/o the application of the Lv3 updates.

[Epson (epson-deactivatedaccount)]

I have this problem too. LM KDE 10.04.
Asus motherboard
Nvidia video card

[Epson (epson-deactivatedaccount)]

I'm using an EXT4 partition by the way.
LM KDE 10.04 32 bit.

[Christopher Kosciuk (cjkosciuk)]

Just to confirm, I'm having the same issue on two separate laptops, a Dell Latitude 2100n and an HP dv6t SE. I've edited the /etc/fstab file to stop the fsck from running, but I'm hoping that will only have to be a temporary workaround until a permanent fix is available.

[Christopher Kosciuk (cjkosciuk)]

Oh, I'm running Linux Mint 9 in both the 32- and 64-bit versions, and with kernels 2.6.32 and 2.6.34.

[Orang_Gila (bug-thug)]

I am new to Linux: ASUS P4R800-VM, Intel Celeron CPU 2.4GHz, 1GB RAM, Linux Mint 9 Isadora, Kernel 2.6.32-21-generic (i686), mountall 2.15, 40GB PATA, 1TB external USB. Every 25 starts the computer comes up with this message and I can not get the keyboard (pressing "c") to abort this frustrating time-waster which seems last 10 minutes, unless I hit escape, then I only have to wait about 3 minutes until it final allows me to hit escape again which shows the mint startup animation/counter.

<ESC>
"(process:245):GLib-WARNING **: getpwuid_r(): failed due to unknown user id (0)
fsck from util-linux-ng 2.17.2
/dev/sda1 has been mounted 25 times without being checked, check forced.
udevd[341]: can not read '/etc/udev/rules.d/z80_user.rules'
---- third time ---------
udevd[344]: can not read '/etc/udev/rules.d/z80_user.rules'
/dev/sda1: 186035/1313280 files (1.3% non-contiguous), 2187669/5242880 blocks"

This is very tiring, I want to be able to bypass this permanently in a setting somewhere!?

[Orographic (temp850)]

I have this issue on my desktop system: Core i3, Gigabyte motherboard H55-USB3. I have Mint 9 installed with kernel 2.6.35 installed from here:

http://kernel.ubuntu.com/~kernel-ppa/mainline/

by following these instructions:

http://forums.linuxmint.com/viewtopic.php?f=42&t=40185&start=0

I can replicate the problem by typing:

sudo touch /forcefsck

In my case, the file system check finishes and disk activity stops and I can press C to continue with no problems.

[axel (axel334)]

I noticed the same bug. Mint Isadora Main, kernel 2.6.34-020634. Actually, I can't see any progress bar indicating how many % left. Pressing C continues to boot.

[tux-sven (sven.b)]

Yes axel, the error is still there, it's a LIE to tell it's fixed. It's remarkable that this error has been without a solution for such a long time.. about 20 weeks I think... I recommend people to use LM8 until it's fixed.

[sacha@ubuntu (sacha-b77)]

Hello!

Sorry for my bad english... ;-)

I'm a newbies with Linux Mint 9 KDE. But, i'm "ubuntu's fan" since 2006...

I've installed LM9 KDE on two machines, the first one : acer P4, 1.5 gig ram and Intel GMA. The second : Dell Inspiron 6000, 512 mo ram and Intel chipset GMA.

Both have the problem. The package mountall 2.15 is up to date like ubuntu lucid. On ubuntu lucid, no problem (see this bug report : https://lists.ubuntu.com/archives/ubuntu-sponsors/2010-May/000927.html). When plymouth is blocked, and, when i press ESC key, i see the same massage with "GLib-WARNING **: getpwuid_r(): falied due to unknown user id (0)".

If i press ESC to go back under plymouth and i press "C" key, the the boot finished with no problem or the boot stalls with a blank screen and the message "GLib-WARNING **: getpwuid_r(): falied due to unknown user id (0)". If i press right arrow key instead of C under plymouth, the boot finished without problems.

I think that there are two problems in this bug (two questions in fact) :

1) Fsck is really finished when plymouth stalls ? If yes, is not a big problem, "press right arrow key" continues the boot process. "Press C" abort the scan, is not a solution to prevent disk errors.

2) What process causes the problem? Mountall ? Plymouth? Ureahead ? I've read that one user have uninstalled Ureahead package and he shall have no problems after. I've not tested. I don't find this forum...

Anyone can confirm that the fsck process is good and finished BEFORE the problem occurs and if the disk is safe or not? It's important for the security and the stability of the system. (disk crashs, errors no repared, ...) If the fsck prosess is good and finished, press "right arrow key" continues the boot without problem. Modify fstab files is not a good idea for me.

Thank's!

[sacha@ubuntu (sacha-b77)]

I have tested with a live usb session.

sudo fsck /dev/sda1 (my boot partition).

Result : clean.

If i look in var/log/boot.log after a forcefsck, the system have a description of hard disk like : /dev/sda1 : 1627009/3571712 fichiers (0.3% non-contigus), 6708217/14284032 blocs. It seems to be ok for the check disk during boot, no?

So, when plymouth loops without end, press "esc" and then press "right arrow", i have seen a brief message "skip apparmor" before the "re"start of the boot.

Apparmor should have link with the problem?

[sacha@ubuntu (sacha-b77)]

Look at this :

http://forums.linuxmint.com/viewtopic.php?f=90&t=49520&hilit=boot+check+disk&start=40#p307318

The user called mick55 explain that, after the upgrade of the kernel to the kernel 2.6.35-020635-generic, he has not had problem on 2 machines with LM9.

Anyone has tried?

The fact that updates level 3 might be in cause, an update to new kernel 2.6.35 can fix it? This is one theory...

[axel (axel334)]

I can confirm. After kernel update to 2.6.32-25-generic #45-Ubuntu (level 5 from mintupdate) it is ok now. I see progress in % , and the whole process is quick and without problem.

[tux-sven (sven.b)]

I have tried to update to kernel 2.6.32-25-generic, but it didn't solve the problem.
I used Mint Update (In Mint 9 Gnome) and included the level 5 updates. During update process I got some questions about substitute rows in scripts, but I ignored this because they where changing rows with "Linux Mint 9" to "Ubuntu 10.04".
But i wonder, I also saw something about not finished rows or something like that...? Is that something that could have an impact???

(Linux Mint 9 Gnome, Acer Aspire One D250)

[sacha@ubuntu (sacha-b77)]

@axel :

Did you have the same problem described by tux-sven? Or, the upgrade is totally fonctionnal with no specific message or error?

Did you always upgrade your system with levels 4 and 5 or once for the kernel upgrade with mint-update?

@tux-sven :

same second question.

Could you post your comments on bug #571707. This page is a duplicate of ubuntu bug.

[axel (axel334)]

@sacha@ubuntu
Now, I have installed Linux Mint 10 Julia 64-bit, so I changed the kernel too.
Linux 2.6.35-22-generic #35-Ubuntu SMP Sat Oct 16 20:45:36 UTC 2010 x86_64 GNU/Linux
I checked with command:
sudo touch /forcefsck && sudo reboot
Unfortunately the problem still exist. I don't have progress bar showing in %. The process was very quick, didn't stalled and system started but I didn't see what was actually happening. Only information that it will be checked.
I also posted the same comment for bug 571707.

"Did you always upgrade your system with levels 4 and 5 or once for the kernel upgrade with mint-update?"
I don't do level for update because in Mint they are described as untested and dangerous. I do level 5 update because they are described as dangerous but tested. And I suppose kernel update may have some serious security updates, that's why I use level 5. But of course I do have a backup. This is topic about kernel update
kernel update and security question
http://forums.linuxmint.com/viewtopic.php?f=90&t=58476
That's why I do it. I prefer to risk reinstatement than keep system potentially vulnerable. But I have to say that after recent update I deleted old kernel and my system broke. So, I think it is better not to remove old kernel. But I'm not experienced enough to comment on kernels anyway. So get advice on forum. :)

[Orang_Gila (bug-thug)]

SOLUTION TO OVER-RIDE FDSK

Under Microsoft's "chkdsk" command Windows users are prompted that they can not run any automatic file system error corrections until the next restart. Linux needs to add this same message and scheduling functionality to "fdsk" (e2fsk). Instead we are stuck with this very frustrating annoyance of being lockout until fdsk completes it error checking after a predetermined number of restarts. Fdisk needs this very basic functionality. Most people moving from Microsoft to Linux expect this.

After days of frustration I found the following solution by Peter Gordon:
My system uses "e2fsprogs (1.41.11)" which includes "tune2fs" which manages "e2fsck".

1. In a terminal window, check your fstab to make sure the sixth column item has a positive integer for your Linux boot device (in my case this partition is an ext3 format with a positive integer of "1").
 code: $ sudo gedit /etc/fstab

2. Restart using a liveCD (because he states that this is a low level procedure which requires all devices to be unmounted).

3. Start a terminal window and unmount all devices.
 code: $ sudo umount -a

3. Disable the <max-mount-counts> & <interval-between-checks> automatic checking described in "man tune2fs".
 code: $ tune2fs -c 0 -i 0 /dev/sda1

4. Restart/Shutdown-restart back to internal hard drive.

Note: Now the onus is on you to remember to manually 'e2fsck' from time to time using a LiveCD or going back and resetting some <max-mount-counts> with tune2fs that is except able to you!
 code: $ sudo e2fsck -C0 -p -f -v /dev/sda2

links:
http://forums.gentoo.org/viewtopic-t-305871-start-0-postdays-0-postorder-asc-highlight-.html
http://linux.die.net/man/8/tune2fs

[Orang_Gila (bug-thug)]

correction typos:
SOLUTION TO OVER-RIDE FSCK

Under Microsoft's "chkdsk" command Windows users are prompted that they can not run any automatic file system error corrections until the next restart. Linux needs to add this same message and scheduling functionality to "fsck" (e2fsck). Instead we are stuck with this very frustrating annoyance of being lockout until fsck completes it error checking after a predetermined number of restarts. FSCK needs this very basic functionality. Most people moving from Microsoft to Linux expect this.

After days of frustration I found the following solution by Peter Gordon:
My system uses "e2fsprogs (1.41.11)" which includes "tune2fs" which manages "e2fsck".

1. In a terminal window, check your fstab to make sure the sixth column item has a positive integer for your Linux boot device (in my case this partition is an ext3 format with a positive integer of "1").
 code: $ sudo gedit /etc/fstab

2. Restart using a liveCD (because he states that this is a low level procedure which requires all devices to be unmounted).

3. Start a terminal window and unmount all devices.
 code: $ sudo umount -a

3. Disable the <max-mount-counts> & <interval-between-checks> automatic checking described in "man tune2fs".
 code: $ tune2fs -c 0 -i 0 /dev/sda1

4. Restart/Shutdown-restart back to internal hard drive.

Note: Now the onus is on you to remember to manually 'e2fsck' from time to time using a LiveCD or going back and resetting some <max-mount-counts> with tune2fs that is except able to you!
 code: $ sudo e2fsck -C0 -p -f -v /dev/sda2

links:
http://forums.gentoo.org/viewtopic-t-305871-start-0-postdays-0-postorder-asc-highlight-.html
http://linux.die.net/man/8/tune2fs

[Alexey Loukianov (lexa2)]

01.12.2010 01:19, Orang_Gila wrote:
> SOLUTION TO OVER-RIDE FSCK
>

Orang, this is the workaround, not the solution. And this information is a
well-known thing every experienced unix user knows about (while any other user
may read about it in man pages if he or she wants).

Adequate end-user of the workstation would prefer filesystems to be checked on
the regular basis. Turning this auto-checks off is a bad idea in general and one
may only afford it as a short-term solution while waiting for bug to be fixed.
It's a shame on Mint that this bug is still unfixed after several months forcing
users to use short-term solution far longer then one may want to.

--
Best regards,
Alexey Loukianov mailto:<email address hidden>
System Engineer, Mob.:[phone]
*nix Specialist