Re: Extra processes/will not die

Try testing some theories.  Maybe process reaping (via SIGTERM) isn't happening.  Try instrumenting mod_fastcgi.c with log messages when it tries to reap a Ruby process (the code is not that bad).  Or make Ruby log whenever it receives a signal--absence of those messages could indicate a problem.  Etc.
TW

At 09:23 PM 05/26/05, Tom Reinhart wrote:
>I'm running Ruby on Rails with mod_fastcgi under Apache 1.3 and am
>having trouble keeping the server responsive for long periods of time.
>My Apache config follows:
>
><IfModule mod_fastcgi.c>
>        FastCgiIpcDir /tmp/fcgi_ipc
>        FastCgiServer /path/dispatch.fcgi -initial-env -processes 5
>-idle-timeout 60
></IfModule>
>
>After running for a while (sorry I can't be more specific on this
>point, but it takes more than a few hours and less than a week), I
>begin to receive Apache 500 errors. If I check the process list, it
>appears that there are the maximum number (as set in config) of
>running Ruby processes:
>
>nobody   31752  0.1  1.6 20592 17088 ?       S    May26   0:02
>/usr/bin/ruby /path/dispatch.fcgi
>
>There are the max number of processes before the errors start, too,
>although everything works perfectly. I believe it may have something
>to do with how FastCGI tries to kill the Ruby processes, but I cannot
>be sure, as I do not know much of the FCGI internals.
>
>Kill9'ing the processes achieves nothing, but `apachectl restart`
>helps. Obviously this isn't something I want to do often.
>
>I have heard that there is better success using the FCGI plugin for
>LigHTTPd, but using that server software isn't an option for me at the
>moment.
>
>What could be causing the errors?
>
>Sincerely,
>
>Tom Reinhart
>[email protected]
>http://AllTom.com/
>___________________________________
>fastcgi-developers mailing list
>http://fastcgi.com/fastcgi-developers/

___________________________________
fastcgi-developers mailing list
http://fastcgi.com/fastcgi-developers/