| Summary: | SLiM display manager produces failed boot and desktop session problems, particularly with LXDE and udisks2 | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Richard Walker <richard.j.walker> |
| Component: | RPM Packages | Assignee: | Mageia Bug Squad <bugsquad> |
| Status: | RESOLVED OLD | QA Contact: | |
| Severity: | major | ||
| Priority: | Normal | CC: | fundawang, mageia, mageia, n54, nic, ruben33en-mandriva |
| Version: | Cauldron | ||
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | slim-1.3.3-1.mga2.src.rpm | CVE: | |
| Status comment: | |||
| Attachments: | log history of SLiM sessions | ||
|
Description
Richard Walker
2012-03-22 02:41:18 CET
we use lxdm as default so decrease the severity CC:
(none) =>
fundawang, n54 From the changelog;
* Fri Mar 30 2012 kamil <kamil> 1.3.3-2.mga2
+ Revision: 227556
- use consolekit's ck-launch-session for login_cmd (to fix possible problems with PolicyKit)
I installed this update on Friday, then I was isolated by the server outage. Sorry to report that SLiM is still not launching cleanly and it is still opening on vt2 leaving vt1 with the unfinished remnants of the boot. The desktop session is still identified by consolekit as having no active user and no local user, so nothing else depending on these characteristics being correct will work - including shutdown.
Richard
Thank you, I will have a look. Created attachment 1909 [details]
log history of SLiM sessions
I found a slim.log from last Friday. It reports things like /usr/bin/xauth: file /home/rich/.Xauthority does not exist Unfortunately there are no timestamps in the log so all I can say for certain is that the last 2 or 3 SLiM sessions are from Friday. bz2'ed and attached <<Unfortunately there are no timestamps in the log>> Of course there is the timestamp given for the xorg.0 log, which I have just spotted, but you knew that anyway :~) I cannot reproduce it. Mounting is working for me. The log is not legible.
Session1:
unix-user = '500'
realname = 'Kamil Rytarowski'
seat = 'Seat1'
session-type = ''
active = FALSE
x11-display = ''
x11-display-device = ''
display-device = '/dev/tty3'
remote-host-name = ''
is-local = TRUE
on-since = '2012-04-03T15:24:52.932723Z'
login-session-id = '1'
idle-since-hint = '2012-04-03T15:31:25.246171Z'
Session2:
unix-user = '500'
realname = 'Kamil Rytarowski'
seat = 'Seat1'
session-type = ''
active = TRUE
x11-display = ':0.0'
x11-display-device = '/dev/tty1'
display-device = '/dev/tty3'
remote-host-name = ''
is-local = TRUE
on-since = '2012-04-03T15:25:29.421515Z'
login-session-id = '1
Sorry about the log. It was big so I attached a bz2 compressed file. I can uncompress it and re-attach if that's what you need. I note you get "Seat1" as I do when using lxdm. The "Seat10" I got with SLiM was not a one-off. It is typical. Naturally your mounting/unmounting works with udisks as you are correctly identified as active and local. Presumably you also get your X session on vt1 rather than vt2 which I always get with SLiM. So something about your Beta2 dodges the bug which makes mine fail. It is starting to look less like a SLiM fault, or perhaps it is a SLiM fault which is only exposed in my Beta2 configuration. Is there a way I can usefully get information from my system about what aspect of the initialisation process is failing? Is there a list of the services needed by SLiM for a successful start? It seems to be something to do with authentication. It is perhaps relevant that I have a long history (back to Alpha2 in January) of hassles with mount and unmount authorisation for various internal and external devices. I have resolved almost all of them by removing HAL, udisks-1 and the HAL policykit stuff. Do SLiM or consolekit rely on anything like that? Richard
Kamil Rytarowski
2012-04-03 20:59:53 CEST
CC:
(none) =>
mageia Colin, any ideas? No change, but this is from a fully up-to-date Cauldron;
[rich@Tureen ~]$ ck-list-sessions
Session1:
unix-user = '501'
realname = 'Rich'
seat = 'Seat2'
session-type = ''
active = FALSE
x11-display = ':0.0'
x11-display-device = '/dev/tty2'
display-device = ''
remote-host-name = ''
is-local = FALSE
on-since = '2012-04-03T19:09:43.148031Z'
login-session-id = '4294967295'
#####################################################################################
This is the relevant slim.log entry for this session;
slim: waiting for X server to begin accepting connections
X.Org X Server 1.11.4
Release Date: 2012-01-27
X Protocol Version 11, Revision 0
Build Operating System: Linux_2.6.38.8-server-10.mga Mageia
Current Operating System: Linux Tureen.Studio 3.3.1-desktop-1.mga2 #1 SMP Mon Apr 2 23:57:52 UTC 2012 x86_64
Kernel command line: BOOT_IMAGE=linux root=UUID=2e61eac0-56cf-482a-a9ba-c8a7a8cad4f5 resume=UUID=737ee8be-ae2f-4eee-bf05-a4c6ccc9f5f8 vga=788
Build Date: 02 March 2012 06:40:51PM
Current version of pixman: 0.24.4
Before reporting problems, check http://bugs.mageia.org
to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Tue Apr 3 20:09:25 2012
(==) Using config file: "/etc/X11/xorg.conf"
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
.resize called 1024 768
/usr/bin/xauth: file /home/rich/.Xauthority does not exist
##### except it does exist and the timestamp on it is 03 Apr 20:09
####################################################################################
And this is from /var/log/messages from the point where I logged on using SLiM:
Apr 3 20:09:42 Tureen systemd-logind[772]: New session c1 of user rich.
Apr 3 20:09:42 Tureen systemd-logind[772]: Linked /tmp/.X11-unix/X0 to /run/user/rich/X11-display.
Apr 3 20:09:43 Tureen dbus-daemon[841]: dbus[841]: [system] Activating via systemd: service name='org.freedesktop.ConsoleKit' unit='console-kit-daemon.service'
Apr 3 20:09:43 Tureen dbus[841]: [system] Activating via systemd: service name='org.freedesktop.ConsoleKit' unit='console-kit-daemon.service'
Apr 3 20:09:43 Tureen console-kit-daemon[1397]: missing action
Apr 3 20:09:43 Tureen dbus-daemon[841]: dbus[841]: [system] Successfully activated service 'org.freedesktop.ConsoleKit'
Apr 3 20:09:43 Tureen dbus[841]: [system] Successfully activated service 'org.freedesktop.ConsoleKit'
Apr 3 20:09:43 Tureen console-kit-daemon[1397]: missing action
Apr 3 20:09:46 Tureen net_applet[1574]: ### Program is starting ###
Apr 3 20:09:47 Tureen mgaapplet[1582]: ### Program is starting ###
Apr 3 20:09:48 Tureen kernel: [ 46.420173] [drm] nouveau 0000:02:00.0: no native mode, forcing panel scaling
Apr 3 20:09:51 Tureen dbus-daemon[841]: dbus[841]: [system] Activating via systemd: service name='org.freedesktop.UDisks2' unit='udisks2.service'
Apr 3 20:09:51 Tureen dbus[841]: [system] Activating via systemd: service name='org.freedesktop.UDisks2' unit='udisks2.service'
Apr 3 20:09:52 Tureen udisksd[1605]: udisks daemon version 1.93.0 starting
Apr 3 20:09:52 Tureen dbus-daemon[841]: dbus[841]: [system] Successfully activated service 'org.freedesktop.UDisks2'
Apr 3 20:09:52 Tureen dbus[841]: [system] Successfully activated service 'org.freedesktop.UDisks2'
Apr 3 20:09:52 Tureen udisksd[1605]: Acquired the name org.freedesktop.UDisks2 on the system message bus
Apr 3 20:09:55 Tureen mgaapplet[1582]: running: ionice -p 1582 -n7
##### console-kit-daemon[1397]: missing action ?????
Finally the consolekit "history" log with those ridiculous huge numbers replaced by local time timestamps. 2012-Apr-03 19:05:07.013 type=SEAT_ADDED : seat-id='Seat1' seat-kind=0 2012-Apr-03 19:05:07.028 type=SEAT_SESSION_ADDED : seat-id='Seat1' session-id='Session1' session-type='' session-x11-display=':0' session-x11-display-device='/dev/tty1' session-display-device='' session-remote-host-name='' session-is-local=TRUE session-unix-user=501 session-creation-time='2012-04-03T19:05:07.025730Z' 2012-Apr-03 19:05:07.031 type=SEAT_ACTIVE_SESSION_CHANGED : seat-id='Seat1' session-id='Session1' 2012-Apr-03 19:08:24.929 type=SYSTEM_RESTART : 2012-Apr-03 19:08:30.522 type=SYSTEM_RESTART : 2012-Apr-03 19:09:22.042 type=SYSTEM_START : kernel-release='3.3.1-desktop-1.mga2' boot-arguments='BOOT_IMAGE=linux root=UUID=2e61eac0-56cf-482a-a9ba-c8a7a8cad4f5 resume=UUID=737ee8be-ae2f-4eee-bf05-a4c6ccc9f5f8 vga=788' 2012-Apr-03 19:09:43.119 type=SEAT_ADDED : seat-id='Seat1' seat-kind=0 2012-Apr-03 19:09:43.152 type=SEAT_ADDED : seat-id='Seat2' seat-kind=1 2012-Apr-03 19:09:43.176 type=SEAT_SESSION_ADDED : seat-id='Seat2' session-id='Session1' session-type='' session-x11-display=':0.0' session-x11-display-device='/dev/tty2' session-display-device='' session-remote-host-name='' session-is-local=FALSE session-unix-user=501 session-creation-time='2012-04-03T19:09:43.148031Z' ##### Note the second seat added at 19:09:43.152. That doesn't happen when using lxdm. Did I say local time? Sorry, I meant zulu time. I am on BST (UTC+1) Yeah, this is pretty messed up. logind seems to get things right, but consolekit is broke (ck-list-sessions does indeed backup what your log output shows). I tried editing /etc/pam.d/slim and added the session pam_ck_connector.so in there and it kinda helped. I'd be sorely tempted to remove the ck-launch-session from the session script for lxde and just rely on pam_ck_connector instead. I think there was a mail about this related to gnome+gdm too, but ck still seems to work properly here... not sure what magic makes it so as there is no specific pam config here.. actually looking further gdm still has console-kit support compiled in. I thought I'd read this was removed... hey ho, it's still there for now it seems. If this wasn't this late in the cycle I would certainly suggest something radical (i.e. nuke built in console-kit support from all DM's, nuke all ck-xinit*/ck-launch* usage, and ensure pam_ck_connector.so was added to all relevant pam configs. In this case, what I would suggest is a combo: Kill off ck-launch-session usage in LXDE xsession scripts and configure the slim pam.d file to use ck_connector. I *think* this will make slim work generally, and gdm, kdm etc. will continue to work with their built in ck support. I'm not sure I'll personally have time to work on this sadly, but here is the slim pam.d file I used: #%PAM-1.0 auth include system-auth auth required pam_nologin.so account include system-auth password include system-auth session required pam_loginuid.so session include system-auth -session optional pam_ck_connector.so session optional pam_gnome_keyring.so auto_start session optional pam_console.so session required pam_namespace.so It ended up creating two ck sessions, one from the connection (which was active and seemed to work fine) and one from ck-launch-sessions which was busted. Hope this helps. I have evaluated the effects of the changes to /etc/pam.d/slim. The first thing I noticed was how similar it is to the lxdm pam.d file, but subtly different. That probably explains why it works :-) I now have a working lxde session started from slim, albeit with vt1 still blocked making shutdown a bit of an adventure. At least udisks-2 is getting the message that I really am here and active ! Hi, This bug was filed against cauldron, but we do not have cauldron at the moment. Please report whether this bug is still valid for Mageia 2. Thanks :) Cheers, marja Keywords:
(none) =>
NEEDINFO My last comment still represents the current situation on Mageia 2.
Sander Lepik
2012-05-26 18:51:42 CEST
Keywords:
NEEDINFO =>
(none) Hi, I can confirm this is an issue on Mageia 2. Using LXDM and LXDE together, LXDE doesn't offer me the option to restart or shutdown the machine, and I'm asked by "org.freedesktop.udisks2.filesystem-mount" for the root password every time I try to mount an USB or DVD-RAM drive on PCManFM. Using GDM instead of LXDM solves both problems. By the way, this seems to be the same bug as https://bugs.mageia.org/show_bug.cgi?id=3936 CC:
(none) =>
ruben33en-mandriva Is Slim supported in Mageia any more? Close? CC:
(none) =>
nic was indeed removed Status:
NEW =>
RESOLVED |