fsck runs on every boot (clean install, single ext4 filesystem)

[jimav]

fsck runs early on *every* boot (and always says "/dev/sda1: clean ...")
The system was always shut down using the regular Unity mechanism (Gear->Shut Down...)

Strangely, the "Last checked" time shown by "sudo dumpe2fs /dev/sda1" shows the date of installation and never changes.

This system is a brand new, clean install from the 15.10 beta2 iso, with subsequent updates.
I took all the installer defaults, which created a single ext4 file system.

Please see attached:
  bootscreen.png - screen shot of fsck output at start of bootup (from VM window)
  dumpe2fs_out.txt - output from "sudo dumpe2fs /dev/sda1"

ProblemType: Bug
DistroRelease: Ubuntu 15.10
Package: upstart 1.13.2-0ubuntu16
ProcVersionSignature: Ubuntu 4.2.0-14.16-generic 4.2.2
Uname: Linux 4.2.0-14-generic x86_64
ApportVersion: 2.19.1-0ubuntu2
Architecture: amd64
CurrentDesktop: Unity
Date: Fri Oct 9 15:10:05 2015
InstallationDate: Installed on 2015-10-07 (2 days ago)
InstallationMedia: Ubuntu 15.10 "Wily Werewolf" - Alpha amd64 (20150924)
ProcKernelCmdline: BOOT_IMAGE=/boot/vmlinuz-4.2.0-14-generic root=UUID=5944d29d-f058-4836-a160-4da2d77554e3 ro quiet splash
SourcePackage: upstart
UpgradeStatus: No upgrade log present (probably fresh install)
UpstartBugCategory: System
UpstartRunningSessionVersion: upstart 1.13.2
UpstartRunningSystemJobs:
 Error: command ['initctl', '--system', 'list'] failed with exit code 1: initctl: Name "com.ubuntu.Upstart" does not exist
UpstartRunningSystemVersion: Error: command ['initctl', '--system', 'version'] failed with exit code 1: initctl: Name "com.ubuntu.Upstart" does not exist

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[A. (a-ahhyes)]

same issue here..

Distributor ID: Ubuntu
Description: Ubuntu 15.10
Release: 15.10
Codename: wily

from boot.log...

[ESC[32m OK ESC[0m] Activated swap /dev/disk/by-uuid/237df0a8-e41a-4377-8902-92c3ea931804.
[ESC[32m OK ESC[0m] Reached target Swap.
[ESC[32m OK ESC[0m] Started File System Check on /dev/d...fbe-61dd-49e1-9ace-950d1bb2aaf0.
^MChecking in progress on 1 disk (0.0% complete)^M Mounting /home...
[ESC[32m OK ESC[0m] Found device ST3000DM001-1CH166 2.
         Starting File System Check on /dev/...d-76d1-4db8-b1b4-d290c8584eaa...
[ESC[32m OK ESC[0m] Mounted /home.
[ESC[32m OK ESC[0m] Started File System Check on /dev/d...c6d-76d1-4db8-b1b4-d290c8584eaa.

Every single boot without fail... Only an issue after upgrading from 15.04 to 15.10.

[asmodey (pavel-finkelshtein)]

Same for me afetr ubuntu upgrade

[A. (a-ahhyes)]

Any ideas why this happening?

[jimav (james-avera)]

Hello "A" (for Anonymous...?),

I did not debug the problem, only filed the bug report. Are you a
developer?

On 10/31/2015 11:27 PM, A. wrote:
> Any ideas why this happening?
>

[A. (a-ahhyes)]

@James - no. I am just an ordinary end user affected by this annoying bug. The silence from the developers is deafening.

[Dread Knight (dread.knight)]

One of the most annoying bugs ever. I've manually edited some files to disable this crap.
It's simply not practical to have to wait 20+ minutes each time I have to boot my PC.

[Cihat Ertem (cihatertem)]

Same for me. I use ext4 / and /home without swap and fsck runs for /home in every boot.

[korakios (korakios667)]

kubuntu 15.10 clean install same problem. I though it was ssd and trim behaviour. I also don't have swap enabled...

[korakios (korakios667)]

Workaround: disable fsck in /etc/fstab (not recommended).
Edit fstab and replace the last digit ("1" for the root partition, "2" for the others) with zero "0".
Waiting for a fix.

[lepetit (lacsebastien)]

any fix ?

[Dominik Smatana (dominiks)]

"Same for me. I use ext4 / and /home without swap and fsck runs for /home in every boot."

Exactly the same for me (no swap, ext4, /home being checked every single time, making each boot very slow). Ubuntu 15.10 here, 64 bit.

[Dominik Smatana (dominiks)]

I'm not sure if it's related, but it looks like I fixed my problem by removing an obsolete swap line from my /etc/fstab

[Mark Pointing (mp035)]

Same problem here, on dell XPS15 9530 (512GB SSD, only seems to be happening to the home partition.) 15.10 64 bit

[Jason Benjamin (red-hax0r)]

The problem is real, but there's there's a simple solution. Edit '/etc/default/grub' and find the line: 'GRUB_CMDLINE_LINUX_DEFAULT'. Add 'fastboot' option to the list. After that update grub ('sudo update-grub'). This is not necessarily a workaround as the problem may have been an omission in the original configuration of grub.

On another note, however, the splash screen still just flashes on reboot occasionally or just doesn't show. Also changing the /etc/fstab entries is unnecessary and useless on its own from my attempts.

[Tomáš Karger (tomak)]

Clean install of 16.04 here. The solution posted in comment #17 seems to work for me. The message indicating fsck run does not appear during boot and is also not present in boot logs.

[Anthony DeRobertis (derobert)]

The "device: clean, 11/65536 files, 12955/262144 blocks" message is what e2fsck prints out when it *skips* an fsck of a filesystem because it's clean. Had it done an fsck, you'd have a message like "device: 11/65536 files (0.0% non-contiguous), 12955/262144 blocks"

Here is an example from a Debian machine:

Jun 06 17:35:43 Zia systemd-fsck[726]: SRV has gone 212 days without being checked, check forced.
Jun 06 17:35:43 Zia systemd-fsck[715]: HOME has gone 212 days without being checked, check forced.
Jun 06 17:35:43 Zia systemd-fsck[717]: /dev/md0: clean, 342/32128 files, 48967/128384 blocks
Jun 06 17:35:49 Zia systemd-fsck[726]: SRV: 161336/22892800 files (2.1% non-contiguous), 56809158/91750400 blocks
Jun 06 17:35:51 Zia systemd-fsck[715]: HOME: Inode 1828390 extent tree (at level 1) could be shorter. IGNORED.
Jun 06 17:35:51 Zia systemd-fsck[715]: HOME: Inode 1828823 extent tree (at level 1) could be shorter. IGNORED.
Jun 06 17:35:51 Zia systemd-fsck[715]: HOME: Inode 2103120 extent tree (at level 1) could be narrower. IGNORED.
Jun 06 17:36:01 Zia systemd-fsck[715]: HOME: 696295/13056000 files (0.5% non-contiguous), 34370145/52428800 blocks

/srv and /home were fsck'd. /dev/md0 was not.

So... if it's not causing a boot slowdown: this is how it's supposed to work.

If it appears to be causing a boot slowdown: It's probably not actually. Very likely, it's something else.

[Braiam Peguero (braiampe)]

That line is just a summary that is printed when not doing a "full check"[1]. This is expected behavior.

[1]: https://github.com/tytso/e2fsprogs/blob/master/e2fsck/unix.c#L433

[mike (taffod)]

Me too. Exact same symptoms. Over 20 sec to boot.

Reverted to CSM boot installation but made no difference.

Comment #17 helps, but boot still v slow and buggy.

Now reverting to 14.04.

I may consider 16 again if I hear an update that this bug is rectified.

[el_gallo_azul (el-gallo-azul)]

This started after I upgraded from 14.04.05 to 16.04 a couple of days ago.

I've just edited /etc/default/grub and changed the line

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"

to

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash fastboot"

I rebooted and don't see the message

(eg. "/dev/sda1: clean, 367968/13541376 file, 7923316/54146048 blocks")

any more during boot.

[Matthias Seidel (matthias-seidel-u)]

This is clearly a regression.

The boot process has to be quit if no errors are found. Otherwise it confuses the user!

[owise1 (o-wise-1)]

Hi, I also appear to have this bug. Does not materially impact boot time as is only involves the EFI system partition but it is a pain as when I switch to any virtual terminal I see that boot message but do not get a cmd prompt.

Clean install of 16.04 on an Intel NUC with ssd drive.

[Kfme (k4fen)]

Hello!
Same problem here. Ubuntu 16.04 LTS. Updated from 14.04 a few months ago. No problem back then but, since a few weeks, I had the same message. I haven't tried all these options but I just wanted to mention the fact that this bug still exists.

[ajgreeny (ajg-charlbury)]

Same problem here on my Xubuntu 16.04 64bit system using UEFI.
I have added the "fastboot" boot option to /etc/default/grub to remove the notification I saw at every boot, but the boot time is still quite a bit slower than 14.04, about 20 secs from end of grub to the login screen, where 14.04, still on the computer, takes about 14 secs.
Perhaps this is the difference between upstart (14.04) and systemd (16.04)?

[WinEunuchs2Unix (ricklee518)]

Brand new install 16.04.3 on Skylake w/UEFI. Originally /EFI partition was being checked so set it to 0 in /etc/fstab. But /dev/nvme0n1p5 kept being checked no matter how long interval set using tune2fs. Finally used grub's `fastboot` kernel command line to make "clean" message go away.