You are not logged in.

#1 2008-10-08 08:19:14

intellitech
Member
From: Chicago, IL
Registered: 2008-09-18
Posts: 34
Website

[PARTIALLY SOLVED] HAL fails to start

** UPDATE: PARTIALLY SOLVED - SEE BOTTOM OF THE FIRST POST

HAL fails to start. I'm not exactly sure why this is, but I believe it has something to do with my attempts to get automounting working under gnome.

[zen@intellitech ~]$ sudo /etc/rc.d/hal start
:: Starting Hardware Abstraction Layer                                   [FAIL]
[zen@intellitech ~]$ vim /etc/rc.conf

MODULES=(acpi-cpufreq cpufreq_ondemand e1000 iwl4965 snd-mixer-oss snd-pcm-oss snd-hwdep snd-page-alloc snd-pcm snd-timer snd snd-hda-intel soundcore thinkpad_acpi uinput vboxdrv)

DAEMONS=(syslog-ng @network !net-profiles @net-auto @netfs !samba !crond @acpid @dbus @hal @alsa !daemoot !avahi-daemon @cups @ddclient @bluetooth)
[zen@intellitech ~]$ pacman -Q | grep hal
hal 0.5.11-1
hal-info 0.20080508-1
[zen@intellitech ~]$ /usr/sbin/hald --daemon=no --verbose=yes
03:13:37.451 [i] hald.c:669: hal 0.5.11
03:13:37.451 [i] hald.c:734: Will not daemonize
03:13:37.451 [i] hald_dbus.c:5381: local server is listening at unix:abstract=/var/run/hald/dbus-pdtZvbqaVq,guid=312b47921cba4c5cdbee099a48ec87d1
03:13:37.452 [E] ck-tracker.c:367: Error doing GetSeats on ConsoleKit: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.ConsoleKit was not provided by any .service files
03:13:37.452 [E] ck-tracker.c:792: Could not get seats and sessions
03:13:37.452 [W] hald_dbus.c:5806: Could not initialize seats and sessions from ConsoleKit
Runner started - allowed paths are '/usr/lib/hal:/usr/lib/hal/scripts:/usr/bin'
03:13:37.453 [i] hald_runner.c:301: Runner has pid 3675
03:13:37.453 [i] hald_runner.c:182: runner connection is 0x15bdec0
03:13:37.456 [W] osspec.c:373: Unable to open /proc/mdstat: No such file or directory
03:13:37.457 [D] util_helper.c:124: drop_privileges: could not set group id
[zen@intellitech ~]$ vim /etc/group

root::0:root
bin::1:root,bin,daemon
daemon::2:root,bin,daemon
sys::3:root,bin
adm::4:root,daemon
tty::5:root,zen
disk::6:root,zen
lp::7:daemon
mem::8:
kmem::9:
wheel::10:root,zen
ftp::11:
mail::12:
log::19:root
smmsp::25:
games::50:
network:x:90:zen
video:x:91:zen
audio::92:root,zen
optical::93:root,zen
floppy:x:94:root,zen
storage:x:95:root,zen
power:x:98:root,zen
nobody::99:
users::100:zen
locate:x:21:
uucp:x:14:
http:x:33:
scanner:x:96:
dbus:x:81:root,zen
hal:x:82:root,zen
avahi:x:84:root,zen
camera:x:97:
gdm:x:101:
vboxusers:x:108:zen
samba:x:109:zen,root

* Yes, I know it is bad practice to have a normal user listed under the disk group. It is necessary, however, in order to have raw disk access for virtual box.

[zen@intellitech ~]$ sudo /etc/rc.d/dbus stop
:: Stopping D-BUS system messagebus                                      [DONE] 
[zen@intellitech ~]$ sudo /etc/rc.d/hal start
:: Starting D-BUS system messagebus                                      [DONE] 
:: Starting Hardware Abstraction Layer                                   [FAIL]

It's odd, because D-BUS will start but hal will not... Anybody have any suggestions?

* UPDATE (10/08/08): Did a full system upgrade and removed dbus from daemons array in rc.conf. UDEV EVENTS Processing speed seems to be back to normal.

** SOLUTION UPDATE: After removing custom ntfs-3g fdi policies from /etc/hal/fdi/policy/, hal seems to load without any problems. This seems to be a bug, since my policies were working at one point, and I believe a hal update caused the problem. I'll see if I can track down the exact hal version that it breaks on, and submit a respective bug report for it.

** OFF-TOPIC UPDATE: Using hwdetect --modules to get a modules list, and using these modules in rc.conf with MOD_AUTOLOAD turned off (and also in mkinitcpio.conf and running mkinitcpio -g <path>), seems to have dramatically increased my "Load udev events" processing time. After doing this, my udev speed went from almost 4000ms to literally < 500ms.

Last edited by intellitech (2008-10-08 14:58:53)


With regards,
Adam Eberlin

My AUR Packages

Offline

#2 2008-10-08 09:46:16

koch
Member
From: Germany
Registered: 2008-01-26
Posts: 369

Re: [PARTIALLY SOLVED] HAL fails to start

intellitech wrote:
DAEMONS=(syslog-ng @network !net-profiles @net-auto @netfs !samba !crond @acpid @dbus @hal @alsa !daemoot !avahi-daemon @cups @ddclient @bluetooth)

you have dbus starting before hal but hal starts debus by default so please try this...

DAEMONS=(...@acpid @hal @alsa...)

and delete debus from that line.

Offline

#3 2008-10-08 11:11:06

intellitech
Member
From: Chicago, IL
Registered: 2008-09-18
Posts: 34
Website

Re: [PARTIALLY SOLVED] HAL fails to start

koch wrote:
intellitech wrote:
DAEMONS=(syslog-ng @network !net-profiles @net-auto @netfs !samba !crond @acpid @dbus @hal @alsa !daemoot !avahi-daemon @cups @ddclient @bluetooth)

you have dbus starting before hal but hal starts debus by default so please try this...

DAEMONS=(...@acpid @hal @alsa...)

and delete debus from that line.

Removed dbus from the daemons array and hal still won't start. Also, I noticed that the "Loading udev events" portion of the boot process seems to take a lot longer than it used to. Last boot, it took about 19130ms, whereas it used to take < 3000ms. Could these be related?

Also, I ran a full system upgrade, both udev and hal were upgraded.

[zen@intellitech ~]$ pacman -Q | grep hal
hal 0.5.11-4
hal-info 0.20080508-1
[zen@intellitech ~]$ pacman -Q | grep udev
klibc-udev 128-1
udev 128-5

With regards,
Adam Eberlin

My AUR Packages

Offline

#4 2008-10-08 13:09:28

koch
Member
From: Germany
Registered: 2008-01-26
Posts: 369

Re: [PARTIALLY SOLVED] HAL fails to start

have a look here

http://bugs.archlinux.org/task/11675

and search the forum for "hal". there are some more people with similar problems.

anyway...putting hal in rc.conf is enough.

have a look also here...

http://wiki.archlinux.org/index.php/Daemons

at the bottom of the page look for "essentials".

i am sort of busy now but will have a closer look at this later because i have this problem too.:(

Offline

#5 2008-10-08 14:59:54

intellitech
Member
From: Chicago, IL
Registered: 2008-09-18
Posts: 34
Website

Re: [PARTIALLY SOLVED] HAL fails to start

I seem to have come up with a temporary solution, although it causes problems with my ntfs automounting thru gnome-volume-manager. After removing custom ntfs-3g fdi policies from /etc/hal/fdi/policy/, hal seems to load without any problems. This seems to be a bug, since my policies were working at one point, and I believe a hal update caused the problem. I'll see if I can track down the exact hal version that it breaks on, and submit a respective bug report for it.

** OFF-TOPIC UPDATE: Using hwdetect --modules to get a modules list, and using these modules in rc.conf with MOD_AUTOLOAD turned off (and also in mkinitcpio.conf and running mkinitcpio -g <path>), seems to have dramatically increased my "Load udev events" processing time. After doing this, my udev speed went from almost 4000ms to literally < 500ms.


With regards,
Adam Eberlin

My AUR Packages

Offline

Board footer

Powered by FluxBB