Mir

lightdm with type=mir causes Ubuntu to freeze on boot. Plymouth progress bar stuck at end.

Bug #1109957 reported by Daniel van Vugt on 2013-01-30
38
This bug affects 8 people
Affects Status Importance Assigned to Milestone
Mir
Fix Released
Critical
Robert Ancell

Bug Description

ppa:~rocket-scientists/staging causes raring to freeze on reboot. That's after I enable it with: /etc/lightdm/lightdm.conf: type=mir

Plymouth is stuck at 100% progress (all orange dots). However lightdm seemingly is not started. It fails to generate any fresh log files.

This is probably nothing to do with the Mir project. However I can't log it in the public projects of lightdm or plymouth right now. Could just be because the PPA lacks plymouth packages for raring... ?

description: updated
Robert Ancell (robert-ancell) wrote :

Is there anything useful in /var/log/upstart? It may be that the lightdm starting conditions are not being met for some reason:

start on ((filesystem
           and runlevel [!06]
           and started dbus
           and (drm-device-added card0 PRIMARY_DEVICE_FOR_DISPLAY=1
                or stopped udev-fallback-graphics))
          or runlevel PREVLEVEL=S)

No. I cleaned out /var/log/upstart/ and then re-enabled booting into mir.

The freeze happens again and upon recovering I see that /var/log/upstart/ is completely empty still. It remains empty until I boot without mir (/etc/lightdm/lightdm.conf: remove "type=mir").

summary: - ppa:~rocket-scientists/staging causes raring to freeze on reboot.
- plymouth stuck at 100% progress
+ /etc/lightdm/lightdm.conf: type=mir causes raring to freeze on reboot.
+ plymouth stuck at 100% progress.
description: updated
summary: - /etc/lightdm/lightdm.conf: type=mir causes raring to freeze on reboot.
- plymouth stuck at 100% progress.
+ lightdm with type=mir causes raring to freeze on boot. plymouth stuck at
+ 100% progress.
information type: Proprietary → Public

I have a similar bug. I have been able to get past plymouth to t blank screen.

Here are contents from one of lightdm.logs

lightdm.log.5

rm: cannot remove '/run/user/ventrical/gvfs': Is a directory
Error writing to daemon: Broken pipe
Error writing to daemon: Broken pipe
Error writing to daemon: Broken pipe
Error writing to daemon: Broken pipe
Error writing to daemon: Broken pipe
Error writing to daemon: Broken pipe
*** Error in `lightdm': double free or corruption (out): 0x09d36128 ***
rm: cannot remove '/run/user/ventrical/gvfs': Is a directory

** (lightdm:1000): WARNING **: Failed to connect to Mir: connect: No such file or directory

** (lightdm:994): WARNING **: Failed to connect to Mir: connect: No such file or directory

** (lightdm:23809): WARNING **: Failed to connect to Mir: connect: No such file or directory

Dale Beaudoin (twocamels) wrote :
Download full text (16.5 KiB)

Here is previous attempt from lightdm.log.4

** (lightdm:2809): WARNING **: Failed to connect to Mir: connect: No such file or directory

** (lightdm:3649): WARNING **: Failed to connect to Mir: connect: No such file or directory
*** Error in `lightdm': double free or corruption (!prev): 0x09b745c8 ***
*** Error in `lightdm': free(): corrupted unsorted chunks: 0x09b72168 ***
======= Backtrace: =========
/lib/i386-linux-gnu/libc.so.6(+0x767e2)[0xb71e37e2]
/lib/i386-linux-gnu/libc.so.6(+0x77530)[0xb71e4530]
/usr/lib/i386-linux-gnu/libstdc++.so.6(_ZdlPv+0x1f)[0xb6f00b4f]
/usr/lib/libmirclient.so.0(_ZNSt13unordered_setIPN11mir_toolkit13MirConnectionESt4hashIS2_ESt8equal_toIS2_ESaIS2_EED1Ev+0x2d)[0xb73f5e4d]
/lib/i386-linux-gnu/libc.so.6(__cxa_finalize+0xb3)[0xb71a0b63]
/usr/lib/libmirclient.so.0(+0x1eab3)[0xb73edab3]
/lib/ld-linux.so.2(+0xf7dc)[0xb77257dc]
/lib/i386-linux-gnu/libc.so.6(+0x337c1)[0xb71a07c1]
======= Memory map: ========
08048000-08071000 r-xp 00000000 08:06 135002273 /usr/sbin/lightdm
08071000-08072000 r--p 00029000 08:06 135002273 /usr/sbin/lightdm
08072000-08073000 rw-p 0002a000 08:06 135002273 /usr/sbin/lightdm
09b70000-09bb2000 rw-p 00000000 00:00 0 [heap]
b6300000-b6321000 rw-p 00000000 00:00 0
b6321000-b6400000 ---p 00000000 00:00 0
b64ff000-b6500000 ---p 00000000 00:00 0
b6500000-b6d00000 rw-p 00000000 00:00 0 [stack:1650]
b6d00000-b6d0b000 r-xp 00000000 08:06 134223643 /lib/i386-linux-gnu/libnss_files-2.17.so
b6d0b000-b6d0c000 r--p 0000a000 08:06 134223643 /lib/i386-linux-gnu/libnss_files-2.17.so
b6d0c000-b6d0d000 rw-p 0000b000 08:06 134223643 /lib/i386-linux-gnu/libnss_files-2.17.so
b6d0d000-b6d17000 r-xp 00000000 08:06 134223647 /lib/i386-linux-gnu/libnss_nis-2.17.so
b6d17000-b6d18000 r--p 00009000 08:06 134223647 /lib/i386-linux-gnu/libnss_nis-2.17.so
b6d18000-b6d19000 rw-p 0000a000 08:06 134223647 /lib/i386-linux-gnu/libnss_nis-2.17.so
b6d26000-b6d2d000 r--s 00000000 08:06 404450423 /usr/lib/i386-linux-gnu/gconv/gconv-modules.cache
b6d2d000-b6d34000 r-xp 00000000 08:06 134279818 /lib/i386-linux-gnu/librt-2.17.so
b6d34000-b6d35000 r--p 00006000 08:06 134279818 /lib/i386-linux-gnu/librt-2.17.so
b6d35000-b6d36000 rw-p 00007000 08:06 134279818 /lib/i386-linux-gnu/librt-2.17.so
b6d36000-b6d38000 r-xp 00000000 08:06 404538542 /usr/lib/i386-linux-gnu/libck-connector.so.0.0.0
b6d38000-b6d39000 r--p 00002000 08:06 404538542 /usr/lib/i386-linux-gnu/libck-connector.so.0.0.0
b6d39000-b6d3a000 rw-p 00003000 08:06 404538542 /usr/lib/i386-linux-gnu/libck-connector.so.0.0.0
b6d3a000-b6d82000 r-xp 00000000 08:06 134414783 /lib/i386-linux-gnu/libdbus-1.so.3.7.2
b6d82000-b6d83000 r--p 00047000 08:06 134414783 /lib/i386-linux-gnu/libdbus-1.so.3.7.2
b6d83000-b6d84000 rw-p 00048000 08:06 134414783 /lib/i386-linux-gnu/libdbus-1.so.3.7.2
b6d8a000-b6d91000 r-xp 00000000 08:06 134223639 /lib/i386-linux-gnu/libnss_compat-2.17.so
b6d91000-b6d92000 r--p 00006000 08:06 134223639 /lib/i386-linux-gnu/libnss_compat-2.17.so
b6d92000-b6d93000 rw-p 00007000 08:06 134223639 /lib/i386-linux-gnu/libnss_compat-2.17.so
b6d93000-b6d96000 r-xp 00000000 08:06 1546101 /lib/i386-linux-gnu/security/pam_env.so
b6d96000-...

Daniel van Vugt (vanvugt) wrote :

Confirmed, the bug is still present with the latest PPA: https://launchpad.net/~mir-team/+archive/staging

In fact, it's now worse. I can't get lightdm to work with type=mir even when started manually.

Robert Ancell (robert-ancell) wrote :

Dale, you appear to have a bug in your lightdm.conf:
"Failed to load configuration from /etc/lightdm/lightdm.conf: Key file contains line 'type-mir' which is not a key-value pair, group, or comment"

Dale Beaudoin (twocamels) wrote :

Hi Robert,

How do I fix this bug? Any ideas ?

Thanks.

Dale Beaudoin (twocamels) wrote :

from robert ancell
Dale, you appear to have a bug in your lightdm.conf:
"Failed to load configuration from /etc/lightdm/lightdm.conf: Key file contains line 'type-mir' which is not a key-value pair, group, or comment"

 That was a typo on my part so that is an invalid report.

Daniel van Vugt (vanvugt) wrote :

Good to know. Though this bug report remains valid. I verified it's still present with the latest PPA a few days ago.

summary: - lightdm with type=mir causes raring to freeze on boot. plymouth stuck at
+ lightdm with type=mir causes Ubuntu to freeze on boot. Plymouth stuck at
100% progress.
Changed in mir:
status: New → Triaged

[SeatDefaults]
#type=unity
user-session=ubuntu
greeter-session=unity-greeter

if uncomments " type=unity", PPA lightdm can't start.
Intel GPU

kevin gunn (kgunn72) on 2013-06-17
Changed in mir:
assignee: nobody → Robert Ancell (robert-ancell)
kevin gunn (kgunn72) on 2013-06-17
Changed in mir:
assignee: Robert Ancell (robert-ancell) → Chris Halse Rogers (raof)
Changed in mir:
assignee: Chris Halse Rogers (raof) → Robert Ancell (robert-ancell)
Robert Ancell (robert-ancell) wrote :

Daniel - I'll need the logs from /var/log/lightdm to work out what is happening here. There could be a number of bugs causing this - if you are not Daniel please open new bugs and attach logs.

I am seeing a lockup on boot, where plymouthd is spinning at 100% CPU and unity-system-compositor is not responding. This causes LightDM to stop (fixed in [1]). I got the logs by ssh'ing in. If I kill plymouthd and restart LightDM it works.

https://code.launchpad.net/~robert-ancell/lightdm/sys-comp-timeout/+merge/169978

Changed in mir:
status: Triaged → Incomplete
Daniel van Vugt (vanvugt) wrote :

Still incomplete.

I am now using saucy with: https://launchpad.net/~mir-team/+archive/system-compositor-testing
and can't reproduce this bug.

I get the plymouthd CPU issue but that's a different bug.

summary: - lightdm with type=mir causes Ubuntu to freeze on boot. Plymouth stuck at
- 100% progress.
+ lightdm with type=mir causes Ubuntu to freeze on boot. Plymouth progress
+ bar stuck at end.
Daniel van Vugt (vanvugt) wrote :

The plymouthd CPU issue is now bug 1192051.

Robert Ancell (robert-ancell) wrote :

With the latest versions of mir and unity-system-compositor and lightdm with https://code.launchpad.net/~robert-ancell/lightdm/unity-assign-usc-vt/+merge/170729 I no longer get any issues on boot.

Changed in mir:
status: Incomplete → Fix Committed
Changed in mir:
milestone: none → 0.0.5
Changed in mir:
status: Fix Committed → Fix Released
Daniel van Vugt (vanvugt) wrote :

Mark,

Are you using the staging or system-compositor-testing PPA? I only ever had this bug with staging, but have not retested staging since moving to saucy. system-compositor-testing works on saucy.

Also, I believe "type=mir" is out of date (Robert?). We use lightdm "type=unity" now, meaning run on unity-system-compositor.

Mark Shuttleworth (sabdfl) wrote :

I'm using mir-team/system-compositor-testing on saucy, with type=unity
(from the 10-unity-system-compositor.conf). Can't boot to the greeter
cleanly, it ends up in a blank screen. If I use a recovery boot, then
'resume normal boot', I get to software rendering X.

Mark

Robert Ancell (robert-ancell) wrote :

Damn. Mark - have you managed to get /var/log/lightdm/lightdm.log and /var/log/lightdm/unity-system-compositor.log after this occurs? Either ssh in and copy them or copy them from recovery before lightdm starts again (they get overwritten - bug 1195185).

Changed in mir:
status: Fix Released → In Progress
Mark Shuttleworth (sabdfl) wrote :

Hi Robert

Here are the two logfiles from a blank-screen-boot. Thanks!

Mark

Micheal Hsu (mkhu) wrote :

someone has also filed this bug
https://bugs.launchpad.net/mir/+bug/1195509

Changed in mir:
milestone: 0.0.5 → 0.0.6
Robert Ancell (robert-ancell) wrote :

Mark, your logs show you have the same symptoms as bug 1195509 - can you please subscribe to that one. I'll close this bug again as this bug doesn't have detailed causes and we made some changes that did improve this for some people.

Changed in mir:
status: In Progress → Fix Released
milestone: 0.0.6 → 0.0.5
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers