[OpenAFS] Linux kernel panic, OpenAFS client, gconf

[Jan-Marc Pilawa]

Am Freitag, 18. Juni 2004 03:48 schrieb chas williams (contractor):
> In message <200406180016.03970.j.pilawa@tu-bs.de>,Jan-Marc Pilawa writes:
> >This Patch. Miles Davis forwared this patch to me. Seems to be a previous
> >version. (I'll apply your newer patch in about 8 hours)
>
> the new patch is only different actually, but i suppose its possible
> that you are getting a race between taking the entry's parent and the
> lock_kernel().  this is mostly why i feel the latest version is more
> correct.
>
> we also probably dont need to call shrink_dcache_parent() on
> dentries that are not directories.

I applied that patch and rebooted the machines. But something is still not 
really good: After 3 hours on one of the machines afs stopped working. dmesg 
shows nothing (no kernel panic), but have a look: 

rzlx0004:~ # ps faxu | grep afs
root     27588  0.0  0.0  1588  468 pts/11   S    13:20   0:00              \_ 
grep afs
root      1131  0.1  0.0     0    0 ?        SW   10:21   0:21 
[afs_rxlistener]
root      1133  0.0  0.0     0    0 ?        DW   10:21   0:00 [afs_callback]
root      1135  0.0  0.0     0    0 ?        DW   10:21   0:00 [afs_rxevent]
root      1137  0.0  0.0     0    0 ?        DW   10:21   0:01 [afsd]
root      1139  0.0  0.0     0    0 ?        DW   10:21   0:00 
[afs_checkserver]
root      1142  0.1  0.0     0    0 ?        SW   10:21   0:14 
[afs_background]
root      1144  0.1  0.0     0    0 ?        SW   10:21   0:18 
[afs_background]
root      1146  0.1  0.0     0    0 ?        SW   10:21   0:16 
[afs_background]
root      1147  0.1  0.0     0    0 ?        SW   10:21   0:14 
[afs_background]
root      1150  0.1  0.0     0    0 ?        SW   10:21   0:16 
[afs_background]
root      1152  0.1  0.0     0    0 ?        SW   10:21   0:13 
[afs_background]
root      1154  0.1  0.0     0    0 ?        SW   10:21   0:13 
[afs_background]
root      1156  0.1  0.0     0    0 ?        SW   10:21   0:15 
[afs_background]
root      1157  0.1  0.0     0    0 ?        SW   10:21   0:17 
[afs_background]
root      1159  0.0  0.0     0    0 ?        SW   10:21   0:07 [afs_cachetrim]
rzlx0004:~ # uname -a
Linux rzlx0004 2.4.21-226-smp4G #1 SMP Tue Jun 15 10:28:32 UTC 2004 i686 i686 
i386 GNU/Linux
rzlx0004:~ # uptime
  1:24pm  up   3:03,  17 users,  load average: 65.62, 63.45, 56.55
rzlx0004:~ # 
rzlx0004:~ # ps faxu | grep DW 
root     28266  0.0  0.0  1592  472 pts/11   S    13:28   0:00  |           \_ 
grep DW
root      1133  0.0  0.0     0    0 ?        DW   10:21   0:00 [afs_callback]
root      1135  0.0  0.0     0    0 ?        DW   10:21   0:00 [afs_rxevent]
root      1137  0.0  0.0     0    0 ?        DW   10:21   0:01 [afsd]
root      1139  0.0  0.0     0    0 ?        DW   10:21   0:00 
[afs_checkserver]
y0003165  4417  0.0  0.0     0    0 ?        DW   10:40   0:01 [kdeinit]
y0003165  4438  0.0  0.0     0    0 ?        DW   10:40   0:03 [kdeinit]
y0003165 21616  0.0  0.0     0    0 ?        DW   12:23   0:00 [gconfd-2]
rzlx0004:~ #

I observed this kind of malfunction before with the one day older patch. With 
unpatched openafs-1.2.11 I observed only these kernel panics. 

There are no processes which were startet later than 12:38 and were accessing 
AFS. The already running Processes which were accessing AFS produce that high 
load. Any Idea how to debug that situation? As long as i can avoid touching 
something in AFS i have some Terminals open... :-/


Mit freundlichen Gruessen / Sincerely

Jan Pilawa

-- 
+ Kontakt ----------------------------------------------------+
+ Systembetreuung Rechenzentrum TU Braunschweig               +
+ Hans-Sommer-Str. 65, D-38092 Braunschweig                   +
+ Tel: +49 531 391-5548 E-Mail: j.pilawa@tu-bs.de ____________+