Bug 104715 - dbus not working
Summary: dbus not working
Status: VERIFIED FIXED
Alias: None
Product: SUSE LINUX 10.0
Classification: openSUSE
Component: Basesystem (show other bugs)
Version: unspecified
Hardware: Other All
: P5 - None : Blocker
Target Milestone: ---
Assignee: Timo Hoenig
QA Contact: Klaus Kämpf
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-08-15 15:45 UTC by Steffen Winterfeldt
Modified: 2007-06-05 09:57 UTC (History)
4 users (show)

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
strace dbus-monitor --system (5.18 KB, text/plain)
2005-08-16 08:57 UTC, Steffen Winterfeldt
Details
ldd -v dbus-monitor (1.86 KB, text/plain)
2005-08-16 08:58 UTC, Steffen Winterfeldt
Details
output of ps axu |grep defunct (1.98 KB, text/plain)
2005-08-16 09:45 UTC, Timo Hoenig
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Steffen Winterfeldt 2005-08-15 15:45:38 UTC
There seems to be a problem with dbus on pre-beta-2. 
 
While dbus-daemon is running neither 'dbus-monitor --system' nor hal 
can connect. They fail with 'No reply within specified time'.
Comment 1 Timo Hoenig 2005-08-15 16:19:06 UTC
Please provide a strace of

   $ strace dbus-monitor --system
Comment 2 Timo Hoenig 2005-08-15 16:25:36 UTC
Since "No reply within specified time" mostly happens if applications are linked against a wrong 
version of libdbus-1 please provide

   $ ldd -v /usr/bin/dbus-monitor
Comment 3 Danny Al-Gaaf 2005-08-15 17:36:17 UTC
I installed latest STABLE Autobuild snapshot via slp (with last dbus and HAL 
packages) and all works perfect (lshal, lshal --monitor, dbus-monitor --system, 
powersave, kpowersave). How installed you this pre-beta-2 ?
Comment 4 Timo Hoenig 2005-08-15 17:41:20 UTC
Danny: Thanks for testing.

Steffen: If you can reproduce this bug, I have prpared D-BUS packages with verbose debug output. 
Please install them from mbuild g0-thoenig-3 and run D-BUS with:

   $ DBUS_VERBOSE=1 dbus-daemon --system --nofork
Comment 5 Steffen Winterfeldt 2005-08-16 08:57:25 UTC
Created attachment 46135 [details]
strace dbus-monitor --system
Comment 6 Steffen Winterfeldt 2005-08-16 08:58:08 UTC
Created attachment 46136 [details]
ldd -v dbus-monitor
Comment 7 Steffen Winterfeldt 2005-08-16 08:59:46 UTC
Just checked again. It's not working. In the installation system, that is. 
Sure, could well be that something is missing, but I don't see what (and it 
worked up to beta1). 
 
You can have a look at d253. 
Comment 8 Klaus Kämpf 2005-08-16 09:19:43 UTC
Is this the cause for #104654 ? 
Comment 9 Timo Hoenig 2005-08-16 09:23:45 UTC
strace of dbus-daemon shows:

   read(6, "BEGIN\r\nl\1\0\1\0\0\0\0\1\0\0\0n\0\0\0\1\1o\0\25\0\0\0"..., 2048) = 135
   open("/etc/passwd", O_RDONLY)           = -1 EACCES (Permission denied)

though this looks sane to be:

   lrwxrwxrwx  1 root root 26 Aug 16 02:50 /etc/passwd -> /mounts/instsys/etc/passwd
   -rw-r--r--  1 root root 697 Dec 31  1969 /mounts/instsys/etc/passwd

I can not reproduce this on Danny's STABLE (as of yesterday evening, ~8pm)

Anyway, d253 does not seem to be in a proper state for debugging, there are way to many zombies:

   inst-sys:/etc # ps axu |grep Z |wc
        25     299    2102

Could we please reboot the machine?
Comment 10 Timo Hoenig 2005-08-16 09:26:28 UTC
(In reply to comment #8)
> Is this the cause for #104654 ? 

I don't know yet.  Do the systems of #104654 also have this amount of zombies? Does  lshal  or  
dbus-monitor --system  for systems which are affected by #104654?
Comment 11 Klaus Kämpf 2005-08-16 09:32:06 UTC
It seems this bug is the cause for #104654. 
While dbus is running, hwinfo --disk reports an UDI but YaST does not see the 
disks. After dbus crashes, hwinfo --disk does NOT report an UDI but YaST now 
sees the disk. 
Comment 12 Timo Hoenig 2005-08-16 09:40:18 UTC
*** Bug 104654 has been marked as a duplicate of this bug. ***
Comment 13 Timo Hoenig 2005-08-16 09:43:08 UTC
As mentioned above the affected systems suffer from having a vast amount of zombies. See 
upcoming logs.
Comment 14 Timo Hoenig 2005-08-16 09:45:40 UTC
Created attachment 46141 [details]
output of ps axu |grep defunct
Comment 15 Timo Hoenig 2005-08-16 09:47:45 UTC
Changing severity to Blocker, assigning to component Kernel.
Comment 16 Steffen Winterfeldt 2005-08-16 09:47:57 UTC
Sorry, the zombies are normal. linuxrc doesn't currently collect orphaned 
children. 
Comment 17 Klaus Kämpf 2005-08-16 10:05:05 UTC
hald is in zombie mode. This shouldn't be normal ?! 
Comment 18 Steffen Winterfeldt 2005-08-16 10:07:00 UTC
Sure it is. It coudn't connect to dbus. See $SUBJECT. 
Comment 19 Klaus Kämpf 2005-08-16 10:11:09 UTC
Timo, sitting next to me debugging this, just confirmed this... 
 
strace shows: open("/etc/passwd", O_RDONLY) = -1 EACCESS (Permission denied) 
Comment 20 Klaus Kämpf 2005-08-16 10:12:18 UTC
see comment #9 
 
Kernel bug ? 
Comment 21 Steffen Winterfeldt 2005-08-16 10:14:11 UTC
I just tested a CD with all beta2 but for the beta1 kernel and it 
doesn't work either. So it's probably not the kernel. 
Comment 22 Steffen Winterfeldt 2005-08-16 10:18:14 UTC
Another bit: 
 
dbus & hald start up ok, but hald dies with the first lshal. After that 
neither one work (even after restarting them). 
Comment 23 Steffen Winterfeldt 2005-08-16 10:29:21 UTC
Breaks even with 9.3 kernel. Hmm. 
Comment 24 Timo Hoenig 2005-08-16 11:01:08 UTC
permissions for /etc are buggy (700)

   $ chmod 755 /etc

fixes the problem.
Comment 25 Steffen Winterfeldt 2005-08-16 11:04:21 UTC
They come from the bootsplash config files. That's why vga=normal helped. :-/ 
Comment 26 Steffen Winterfeldt 2005-08-16 11:10:38 UTC
Ok, now hal & hwinfo work fine. 
 
But yast doesn't find disks. 
Comment 27 Steffen Winterfeldt 2005-08-16 11:11:06 UTC
reopen for yast 
Comment 28 Steffen Winterfeldt 2005-08-16 11:12:43 UTC
the wrong one