Bug 988231

Summary: GDM fails to start because of journal
Product: [Fedora] Fedora Reporter: Till Maas <opensource>
Component: systemdAssignee: systemd-maint
Status: CLOSED DUPLICATE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: frankly3d, harald, i.grok, johannbg, junk, lnykryn, mschmidt, msekleta, notting, opensource, plautrba, selinux, systemd-maint, vpavlin, zbyszek
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-09-06 12:45:51 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
output of journalctl -b
none
Output of "systemd-analyze plot" of system with /var/log/journal
none
Output of "systemd-analyze plot" of system without /var/log/journal none

Description Till Maas 2013-07-25 07:16:43 UTC
Description of problem:

On Fedora 19, sometimes GDM refuses to start on bootup but it works after removing /var/log/journal. Other people have similar problems:
http://www.linuxquestions.org/questions/fedora-35/fedora-boot-problem-4175468468/

It is not clear what lead to this error, but the X log file for GDM mentions an IO Error 11, but does not specify for which resource.

Comment 1 Michal Schmidt 2013-07-25 11:18:55 UTC
Do you use the btrfs filesystem?

Comment 2 Till Maas 2013-07-25 11:39:11 UTC
(In reply to Michal Schmidt from comment #1)
> Do you use the btrfs filesystem?

No.
Additional information: It is a system upgraded from F18 to F19 and the bug appeared recently.

Comment 3 Till Maas 2013-07-25 11:39:52 UTC
Also after logging in as root and running "systemctl restart gdm", gdm started without problems.

Comment 4 Till Maas 2013-07-26 07:50:04 UTC
this might be a duplicate of bug 987242

Comment 5 Lennart Poettering 2013-08-06 14:56:23 UTC
Please provide the output of "journalctl -b" when this happens and attach it to this bug!

Comment 6 Till Maas 2013-08-06 15:44:47 UTC
(In reply to Lennart Poettering from comment #5)
> Please provide the output of "journalctl -b" when this happens and attach it
> to this bug!

I believe the broken journal was deleted, therefore the error cannot reproduced on the affected system.

Comment 7 L.L.Robinson 2013-08-15 21:54:04 UTC
Created attachment 787071 [details]
output of journalctl -b

I had this issue, I ran journalctl -b and save the output. removing the /var/log/journal directory solved the boot timeout issue for me as well. Hope it helps

Comment 8 Scott Schmit 2013-08-16 11:43:21 UTC
FWIW, bug 985038, bug 967521, bug 970801, and bug 987242 are all duplicates.

In my case, it started after a Fedora 18 -> 19 upgrade.

Thanks for your attention to this issue.

Comment 9 Tom London 2013-09-06 00:10:40 UTC
I'm noticing "boot slow down" when journaling is enabled and when /var/log/journal exists.

I attach below plots from "systemd-analyze plot" for 2 consecutive boots (same kernel, same packages, same ...), the only difference being that I moved /var/log/journal to /var/log/journal.old.

The difference was striking: the first boot (with journal) took 46 seconds; the second boot (with journal moved to journal.old) took 22 seconds.

Although the first plot shows the "missing 24 seconds", I can't figure out where the time went.

Besides this (graphical) output, what info can I provide that would be helpful?

Comment 10 Tom London 2013-09-06 00:11:54 UTC
Created attachment 794548 [details]
Output of "systemd-analyze plot" of system with /var/log/journal

Comment 11 Tom London 2013-09-06 00:12:28 UTC
Created attachment 794549 [details]
Output of "systemd-analyze plot" of system without /var/log/journal

Comment 12 Harald Hoyer 2013-09-06 04:53:44 UTC
(In reply to Tom London from comment #10)
> Created attachment 794548 [details]
> Output of "systemd-analyze plot" of system with /var/log/journal

Wow.. please boot with "systemd.log_level=debug" on the kernel command line.
Then, when you are booted attach the output of:

# journalctl -o short-monotonic -ab

Comment 13 Harald Hoyer 2013-09-06 04:56:36 UTC
oh, btw, are you running on a 32bit or 64bit kernel?

Comment 14 Harald Hoyer 2013-09-06 06:10:26 UTC
pushed some journal fixes in systemd-206-11

Comment 15 Zbigniew Jędrzejewski-Szmek 2013-09-06 12:45:51 UTC
Similar issue with plymouth-quit-wait.service failing happens here too. Probably removing the journal directory un-triggers the race condition between plymouth and gdm, and is not the real issue.

*** This bug has been marked as a duplicate of bug 967521 ***

Comment 16 Tom London 2013-09-06 13:13:58 UTC
(In reply to Harald Hoyer from comment #13)
> oh, btw, are you running on a 32bit or 64bit kernel?

64 bit: kernel-3.12.0-0.rc0.git11.2.fc21.x86_64

Believe this has been occurring for quite some time however (months...).

Comment 17 Frank Murphy 2013-10-15 08:01:07 UTC
I am having this on F20, 
and plymouth is not installed
No btrfs
Xfce fully updated, lightdm.

Comment 18 Frank Murphy 2013-10-15 13:15:31 UTC
(In reply to Frank Murphy from comment #17)
> I am having this on F20, 
> and plymouth is not installed
> No btrfs
> Xfce fully updated, lightdm.

Apologies, ignore above. added to wrong bz.