[Systems] wiki is down

[Bernie Innocenti]

On Sat, 2010-02-20 at 20:16 -0500, Chris Ball wrote:
> Looks to me like apache died, but I don't have access to restart it or
> look at its logfiles to find out why.

Apache had exhausted sysvipc semaphores once again. I fixed it with the
same magic formula from the last time:

  for i in `ipcs -s | awk '{print $2}'`; do ipcrm -s $i; done

Unsurprisingly, the last request served by apache before dying was for
related to Trac.

Also, browsing the logs, I've found:

 [Sat Feb 20 10:20:34 2010] [notice] child pid 11555 exit signal Segmentation fault (11)
 [Sat Feb 20 17:02:51 2010] [notice] child pid 2654 exit signal Segmentation fault (11)
 [Sat Feb 20 19:02:56 2010] [error] server reached MaxClients setting, consider raising the MaxClients setting

Our MaxClient is already 256, I doubt raising it further makes much
sense.

I have no explanation for the segfaults, but cjb reports that gcc
sometimes crashes. This usually means kernel bug or unreliable hardware,
because gcc is 100% input-driven and therefore its crashes should be
deterministic.

-- 
   // Bernie Innocenti - http://codewiz.org/
 \X/  Sugar Labs       - http://sugarlabs.org/