[gentoo-user] hald won't start

[Mick]

Hi All,

I noticed on a box of mine that hald won't start:
==========================================
# /etc/init.d/hald restart
* Starting Hardware Abstraction Layer daemon ... [ !! ]
==========================================

Confusingly, the logs don't show anything. How do I troubleshoot this?
--
Regards,
Mick

[Mick]

I just noticed that I do not have CONFIG_INOTIFY_USER enabled in the kernel.
Perhaps this has something to do with it?
--
Regards,
Mick

[Dale]

I'm not positive but I had several errors because I was missing that option in my kernel.  You may want to build that one in regardless, just to prevent future problems.  I seem to recall that the error I got told me I needed that in the kernel but that was a few weeks ago and I'm not a elephant.  :/  It's a good place to start though.

Dale

:-)  :-)

[Mick]

On Saturday 14 July 2007 21:08, Dale wrote:
> Mick wrote:

> > I just noticed that I do not have CONFIG_INOTIFY_USER enabled in the
> > kernel. Perhaps this has something to do with it?
>
> I'm not positive but I had several errors because I was missing that
> option in my kernel. You may want to build that one in regardless, just
> to prevent future problems. I seem to recall that the error I got told
> me I needed that in the kernel but that was a few weeks ago and I'm not
> a elephant. :/ It's a good place to start though.

Thanks Dale, I will recompile the kernel as soon as it finishes emerging OOo
(in the next day or so . . . )

--
Regards,
Mick

[Alessandro del Gallo]

Mick ha scritto:

I had the same problem:

hald was sucking 99% cpu, removables devices didn't work
and launching hald from terminal with the option "hald --daemon=no
--verbose=yes" i discovered it was on a infinite loop echo-ing
something about rules so I got in /etc/udev/rules.d
and found (by probing one by one) 99-libgphoto2.rules that
was HALS's murder.

I hope it's useful for you

Alessandro
--
gento...@gentoo.org mailing list

[Mick]

On Monday 16 July 2007 10:26, Alessandro del Gallo wrote:
> Mick ha scritto:
> > Hi All,
> >
> > I noticed on a box of mine that hald won't start:
> > ==========================================
> > # /etc/init.d/hald restart
> > * Starting Hardware Abstraction Layer daemon ... [ !! ]
> > ==========================================
> >
> > Confusingly, the logs don't show anything. How do I troubleshoot this?
>
> I had the same problem:
>
> hald was sucking 99% cpu, removables devices didn't work
> and launching hald from terminal with the option "hald --daemon=no
> --verbose=yes" i discovered it was on a infinite loop echo-ing
> something about rules so I got in /etc/udev/rules.d
> and found (by probing one by one) 99-libgphoto2.rules that
> was HALS's murder.

Thank you Alessandro, what do you mean by probing - how did you probe them?
--
Regards,
Mick

[Alessandro del Gallo]

moving it to /tmp and restart hald,
becouse the error I got when launching
# hald --daemon=no --verbose=yes
is "Unhandled Rule (0)"
so deleting the rules (delete de rule, start hald, if it doesn't
works put the rule back , and move another) from etc/udev/rules.d
I debugged the problem, the bad rule was owned by gphoto2

sorry for my bad english

Mick ha scritto:

--
gento...@gentoo.org mailing list

[Mick]

On Tuesday 17 July 2007 10:52, Alessandro del Gallo wrote:
> moving it to /tmp and restart hald,
> becouse the error I got when launching
> # hald --daemon=no --verbose=yes
> is "Unhandled Rule (0)"
> so deleting the rules (delete de rule, start hald, if it doesn't
> works put the rule back , and move another) from etc/udev/rules.d
> I debugged the problem, the bad rule was owned by gphoto2

The thing is I do not get any errors in /var/log/messages and I do not have
gphoto2 (I do not have gnome on this box).

An interesting thing, when I emerge hal it is installed and started alright.
If I try to restart it (or during boot) it fails. No error messages other
than [!!]. I am totally in the dark with this.

> sorry for my bad english

No problemo!
--
Regards,
Mick

[Mick]

On Tuesday 17 July 2007 10:52, Alessandro del Gallo wrote:
> Mick ha scritto:
> > On Monday 16 July 2007 10:26, Alessandro del Gallo wrote:
> >> Mick ha scritto:
> >>> Hi All,
> >>>
> >>> I noticed on a box of mine that hald won't start:
> >>> ==========================================
> >>> # /etc/init.d/hald restart
> >>> * Starting Hardware Abstraction Layer daemon ... [ !! ]
> >>> ==========================================
> >>>
> >>> Confusingly, the logs don't show anything. How do I troubleshoot this?
> >>
> >> I had the same problem:
> >>
> >> hald was sucking 99% cpu, removables devices didn't work
> >> and launching hald from terminal with the option "hald --daemon=no
> >> --verbose=yes" i discovered it was on a infinite loop echo-ing
> >> something about rules so I got in /etc/udev/rules.d
> >> and found (by probing one by one) 99-libgphoto2.rules that
> >> was HALS's murder.
> >
> > Thank you Alessandro, what do you mean by probing - how did you probe
> > them?

> moving it to /tmp and restart hald,

> becouse the error I got when launching
> # hald --daemon=no --verbose=yes
> is "Unhandled Rule (0)"
> so deleting the rules (delete de rule, start hald, if it doesn't
> works put the rule back , and move another) from etc/udev/rules.d
> I debugged the problem, the bad rule was owned by gphoto2

I've had a chance to get back to this machine and this is what it is showing:
==========================================
# hald --daemon=no --verbose=yes
11:16:08.930 [I] hald.c:533: hal 0.5.9
11:16:08.931 [I] hald.c:598: Will not daemonize
11:16:08.945 [I] hald_dbus.c:4807: local server is listening at
unix:abstract=/var/run/hald/dbus-1LzyssbMzl,guid=faadd829b7aaff57dce0510046b451e8
11:16:09.174 [I] hald_runner.c:299: Runner has pid 31241
11:16:09.180 [W] ci-tracker.c:200: Could not get uid for connection:
org.freedesktop.DBus.Error.NameHasNoOwner Could not get UID of
name 'org.freedesktop.DBus': no such name
11:16:09.181 [E] hald_dbus.c:4462: Cannot get caller info for
org.freedesktop.DBus
Runner started - allowed paths
are '/usr/libexec:/usr/lib/hal/scripts:/usr/bin'
11:16:09.196 [I] hald_runner.c:180: runner connection is 0x8096988
11:16:09.330 [I] mmap_cache.c:251: cache mtime is 1184442336
*** [DIE] osspec.c:watch_fdi_files():349 : Unable to initialize inotify:
Function not implemented
==========================================

What do you make of this?
--
Regards,
Mick

[Mick]

>b451e8 11:16:09.174 [I] hald_runner.c:299: Runner has pid 31241

> 11:16:09.180 [W] ci-tracker.c:200: Could not get uid for connection:
> org.freedesktop.DBus.Error.NameHasNoOwner Could not get UID of
> name 'org.freedesktop.DBus': no such name
> 11:16:09.181 [E] hald_dbus.c:4462: Cannot get caller info for
> org.freedesktop.DBus
> Runner started - allowed paths
> are '/usr/libexec:/usr/lib/hal/scripts:/usr/bin'
> 11:16:09.196 [I] hald_runner.c:180: runner connection is 0x8096988
> 11:16:09.330 [I] mmap_cache.c:251: cache mtime is 1184442336
> *** [DIE] osspec.c:watch_fdi_files():349 : Unable to initialize inotify:
> Function not implemented
> ==========================================
>
> What do you make of this?

I removed /usr/share/hal and tried to remerge it. It complained that I "MUST"
compile pciutils without the zlib flag. Not sure why pciutils has the zlib
flag as a default, but if it is causing a problem with hal shouldn't it be
removed?

Anyway, after masking zlib for pciutils and remerging both packages it now
seems to be working again . . .

Thanks for your suggestions.
--
Regards,
Mick

[Alessandro del Gallo]

Hi,

Last 2 lines of your log:

Unable to initialize inotify:
> > Function not implemented

inotify comes from kernel, in my config:

elwood linux # zcat /proc/config.gz |grep -i notify
CONFIG_INOTIFY=y
CONFIG_INOTIFY_USER=y
CONFIG_DNOTIFY=y

in menuconfig:

┌─────── File systems

[...]

[*] Inotify file change notification support

[*] Inotify support for userspace

bye bye

Mick ha scritto:

--
gento...@gentoo.org mailing list