Hpsmhd lässt sich nach crash nicht mehr starten

HP Insight Diagnostics Insight Diagnostics ist eine lustige Sache wenn er läuft.
wenn der server aber crasht, kann es sein, dass sich dieser nicht mehr starten lässt, das ganze schaut dann so aus...

T[root@hihhlx05 ~]# /etc/init.d/hpsmhd start
Starting hpsmhd:
...                                                        [  OK  ]
T[root@hihhlx05 ~]# /etc/init.d/hpsmhd status
hpsmhd dead but subsys locked

das ist übel, ein blick ins logfile sagt dann:

T[root@hihhlx05 /var/spool/opt/hp/hpsmh/logs]# tail -f *
==> access_log <==
10.16.9.11 - - [19/Jun/2006:11:11:45 +0200] "GET / HTTP/1.1" 302 351
10.16.9.11 - - [19/Jun/2006:11:11:45 +0200] "GET /cpqlogin.htm?RedirectUrl=/&RedirectQueryString= HTTP/1.1" 200 12724
10.16.9.11 - - [19/Jun/2006:11:11:45 +0200] "GET /hplogo.gif HTTP/1.1" 200 1807
==> error_log <==
hpsmhd PID->7979 is running now!
[Thu Jun 22 20:26:24 2006] [error] Cannot allocate shared memory: (17)File exists
hpsmhd PID->7979 is running now!
hpsmhd PID->7979 is running now!
hpsmhd PID->7979 is running now!
[Thu Jun 22 20:28:12 2006] [error] Cannot allocate shared memory: (17)File exists
hpsmhd PID->7979 is running now!
[Thu Jun 22 20:30:28 2006] [error] Cannot allocate shared memory: (17)File exists
[Sat Jun 24 01:17:08 2006] [error] Cannot allocate shared memory: (17)File exists
[Sun Jun 25 22:43:24 2006] [error] Cannot allocate shared memory: (17)File exists

T[root@hihhlx05 /opt/hp/hpsmh/logs]# cat smh.log
000054E3
ERROR LOG (severity=2)  Fri Nov 11 10:40:04 2005
LocalAccessType value "Anonymous;Admi" is not valid, please reset the value

übel übel, das sieht schlimm aus, aber es kommt besser: (google sei dank)

> [Wed Aug 20 10:33:24 2003] [notice] Digest: generating secret for digest 
> authentication ...
> [Wed Aug 20 10:33:24 2003] [notice] Digest: done
> [Wed Aug 20 10:33:26 2003] [error] Cannot allocate shared memory: (17)File 
> exists

This is a bug in mod_ssl (#78019); to work around it, after booting up
but *before* trying to start httpd, remove the file
/var/cache/mod_ssl/scache.

danach war das ganze leichtes spiel:

T[root@hihhlx05 /opt/hp/hpsmh/logs]# mv scache /tmp/
T[root@hihhlx05 ~]# /etc/init.d/hpsmhd start

Starting hpsmhd:
...                                                        [  OK  ]
T[root@hihhlx05 ~]# /etc/init.d/hpsmhd status
hpsmhd (pid 17935 17932) is running...

trallalla, es funzt... komisch aber wirkungsvoll :~)