[passbreak2001]

Hallo,

komme mit meinem Apacheproblemchen müßig, aber immerhin step-by-step weiter.
Grundlage: Apache shutdown einmal in der Woche zu einer bestimmten Uhrzeit. War der Meinung das ein Cronscript dafür verantwortlich ist aber was hat 'mandb' mit Apache zu tun?!

Nein. In der authlog habe ich nun übereinstimmende Einträge gefunden. Kann aber nichts damit anfangen. Wer kann mir bei der Analyse helfen?

Feb 11 05:19:14 vsXXX su[28827]: + ??? root:nobody
Feb 11 05:19:15 vsXXX su[28827]: (pam_unix) session opened for user nobody by (uid=0)
Feb 11 05:19:17 vsXXX pam_limits[28827]: setrlimit limit #6 to soft=-1, hard=-1 failed: Operation not permitted; uid=0 euid=0
Feb 11 05:19:18 vsXXX su[28827]: pam_open_session: Permission denied 

--> Apache shutted down um 5.20 Uhr

Eine Woche später:

Feb 18 05:19:02 vsXXX CRON[26800]: (pam_unix) session opened for user root by (uid=0)
Feb 18 05:19:08 vsXXX CRON[26800]: (pam_unix) session closed for user root
Feb 18 05:19:15 vsXXX su[27109]: + ??? root:nobody
Feb 18 05:19:16 vsXXX su[27109]: (pam_unix) session opened for user nobody by (uid=0)
Feb 18 05:19:17 vsXXX pam_limits[27109]: setrlimit limit #6 to soft=-1, hard=-1 failed: Operation not permitted; uid=0 euid=0
Feb 18 05:19:19 vsXXX su[27109]: pam_open_session: Permission denied 
Feb 18 05:20:01 vsXXX CRON[27673]: (pam_unix) session opened for user root by (uid=0)

--> Apache shutted down um 5.21 Uhr

Das Spiel übrigens jede Woche einmal. Könnte zwar ein cronscript einrichten der Apache wieder restartet um 5.25 z.B., würde aber lieber das Problem in der Basis lösen.

Bin für jede Hilfe dankbar!

[vmk]

mhm, was knallt denn da gegen das rlimit? Ich denke, irgendwas verbraucht viel Speicher und die Hostmaschine killt daraufhin Prozeße bzw. ein Prozeß "stürzt" einfach so ab.

[passbreak2001]

Zitat von vmk 

mhm, was knallt denn da gegen das rlimit? ...

Kurz vorher wird ein Cronscript gestartet der die 'mandb' aktualisiert. Das das irgendwie, zumindest mittelbar, zusammenhängt habe ich schon gemerkt wegen der Uhrzeit. ABER WARUM ?

Was zum Geier hat Apache damit zu tun? Seh selbts (apache error.log):

[Sun Feb 18 05:21:33 2007] [warn] child process 32739 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 12447 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 23141 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 12448 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 23148 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 23150 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 14183 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 25045 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 22730 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:35 2007] [warn] child process 13044 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 32739 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 12447 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 23141 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 12448 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 23148 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 23150 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 14183 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 25045 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 22730 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:36 2007] [warn] child process 13044 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 32739 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 12447 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 23141 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 12448 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 23148 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 23150 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 14183 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 25045 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 22730 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:37 2007] [warn] child process 13044 still did not exit, sending a SIGTERM
[Sun Feb 18 05:21:39 2007] [error] child process 32739 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 12447 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 23141 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 12448 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 23148 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 23150 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 14183 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 25045 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 22730 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:39 2007] [error] child process 13044 still did not exit, sending a SIGKILL
[Sun Feb 18 05:21:40 2007] [notice] caught SIGTERM, shutting down

[vmk]

In dem Log was du zuletzt gepostet hast, sieht man das der Apache abgestürzt ist. Bi st du speichermäßig knapp dran?

Die mandb wird aktualisiert, weil du das so willst - Es ist als regelmäßiger Job eingetragen.

[passbreak2001]

Zitat von vmk 

In dem Log was du zuletzt gepostet hast, sieht man das der Apache abgestürzt ist. Bi st du speichermäßig knapp dran?

Die mandb wird aktualisiert, weil du das so willst - Es ist als regelmäßiger Job eingetragen.

Nicht ganz weil ich das will. War schon bei der Grundkonfiguration so eingestellt. Aber was ist mit der auth.log? Kann das nicht eher was mit dem 'permission denied for user nobody' zu tun haben?