The vanishing daemons problem seems to be causing an issue on a new
server I've recently deployed.  It may have been a very rare
occurrence on an old single core machine (really can't remember), but
now random daemons are disappearing now we've switched to a more
modern quad core xeon server.

I've tried raising the debug level in Apache to see if it can show any
useful details, a typical working restart:

[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Daemon process
inactivity timer expired, stopping process 'fatfluffs'.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Shutdown
requested 'fatfluffs'.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Stopping
process 'fatfluffs'.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Destroying
interpreters.
[Sun Mar 14 01:14:02 2010] [debug] mod_wsgi.c(5172): mod_wsgi
(pid=6795): Create thread state for thread 0 against interpreter
'www.fatfluffs.com|/site.wsgi'.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Destroy
interpreter 'www.fatfluffs.com|/site.wsgi'.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Cleanup
interpreter ''.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Terminating
Python.
[Sun Mar 14 01:14:02 2010] [info] mod_wsgi (pid=6795): Python has
shutdown.
[Sun Mar 14 01:14:03 2010] [info] mod_wsgi (pid=6795): Process
'fatfluffs' has died, restarting.
[Sun Mar 14 01:14:03 2010] [info] mod_wsgi (pid=7134): Starting
process 'fatfluffs' with uid=33, gid=33 and threads=15.
[Sun Mar 14 01:14:03 2010] [info] mod_wsgi (pid=7134): Initializing
Python.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(11151): mod_wsgi
(pid=7134): Process 'fatfluffs' logging to 'www.fatfluffs.com' with
log level 7.
[Sun Mar 14 01:14:03 2010] [info] mod_wsgi (pid=7134): Attach
interpreter ''.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10662): mod_wsgi
(pid=7134): Starting 15 threads in daemon process 'fatfluffs'.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10491): mod_wsgi
(pid=7134): Enable monitor thread in process 'fatfluffs'.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10672): mod_wsgi
(pid=7134): Starting thread 1 in daemon process 'fatfluffs'.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10495): mod_wsgi
(pid=7134): Deadlock timeout is 300.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10498): mod_wsgi
(pid=7134): Inactivity timeout is 300.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10461): mod_wsgi
(pid=7134): Enable deadlock thread in process 'fatfluffs'.
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10672): mod_wsgi
(pid=7134): Starting thread 2 in daemon process 'fatfluffs'.
(removed repeated lines)
[Sun Mar 14 01:14:03 2010] [debug] mod_wsgi.c(10672): mod_wsgi
(pid=7134): Starting thread 15 in daemon process 'fatfluffs'.
[Sun Mar 14 01:14:06 2010] [debug] mod_wsgi.c(11925): mod_wsgi
(pid=7104): Request server was 'www.fatfluffs.com|0'.
[Sun Mar 14 01:14:06 2010] [debug] mod_wsgi.c(12676): mod_wsgi
(pid=7134): Server listener address '|80'.
[Sun Mar 14 01:14:06 2010] [debug] mod_wsgi.c(12685): mod_wsgi
(pid=7134): Server listener address '|80' was found.
[Sun Mar 14 01:14:06 2010] [debug] mod_wsgi.c(12697): mod_wsgi
(pid=7134): Connection server matched was 'drake.hawkz.com|80'.
[Sun Mar 14 01:14:06 2010] [debug] mod_wsgi.c(12713): mod_wsgi
(pid=7134): Request server matched was 'www.fatfluffs.com|0'.

Then this is the last restart in the logs before it started failing
requests:

[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Daemon process
inactivity timer expired, stopping process 'fatfluffs'.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Shutdown
requested 'fatfluffs'.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Stopping
process 'fatfluffs'.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Destroying
interpreters.
[Sun Mar 14 01:44:07 2010] [debug] mod_wsgi.c(5172): mod_wsgi
(pid=7471): Create thread state for thread 0 against interpreter
'www.fatfluffs.com|/site.wsgi'.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Destroy
interpreter 'www.fatfluffs.com|/site.wsgi'.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Cleanup
interpreter ''.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Terminating
Python.
[Sun Mar 14 01:44:07 2010] [info] mod_wsgi (pid=7471): Python has
shutdown.
[Sun Mar 14 01:44:08 2010] [debug] mod_wsgi.c(11925): mod_wsgi
(pid=7608): Request server was 'www.fatfluffs.com|0'.

After this point requests slowly end up with error 500 until an Apache
reload is needed to get the missing daemon back.

We're running mostly standard Debian Lenny stuff, Apache 2.2.9, tried
most versions of mod_wsgi (packaged, backport, latest) and all seem to
end up with the same problem.  WSGIRestrictEmbedded is turned on with
mod_wsgi 3.2, all daemons are standard 1 process, 15 thread, 1000 max
requests, 300 sec timeout.

It's certainly occurring every few days now and is getting highly
annoying. :(

If there's any other details needed to figure this one out then let me
know.

-- 
You received this message because you are subscribed to the Google Groups 
"modwsgi" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/modwsgi?hl=en.

Reply via email to