this is a strange situation i thought i'd post and try to get some
expertise on.

the short version is, i do something to get mod_python to crash.  i
then try to clean up all the tmp files, look for open files, sysv ipc
stuff, etc.  but even after all that i can't get apache to run again
without a full reboot.  it seems very strange that the runs of apache
should be any different!  i'm looking for ways to clean up the system
without rebooting, ala ipcrm, etc.

the details:

i'm debugging something right now which seems to crash apache 2.0.48
(pre-fork) and mod_python 3.1.3 on FreeBSD 4.9-STABLE x86.  i'm not
terribly worried about that.  but after i crash it, when i try to run
it again i get log msgs like:

[Thu Mar 11 10:08:51 2004] [notice] mod_python: Creating 32 session mutexes
based on 50 max processes and 0 max threads.
(24)Too many open files: mod_python: Failed to reinit global mutex
/tmp/mpmtx84815.
'import site' failed; use -v for traceback
[Thu Mar 11 10:08:52 2004] [error] make_obcallback: could not import
mod_python.apache.
ImportError: No module named mod_python.apache
[Thu Mar 11 10:10:38 2004] [error] (24)Too many open files: apr_accept: (client
socket)

i've cleaned up /tmp/mpm*, /var/run/httpd.scoreboard, etc.  i've also
verified that in fact

kern.openfiles: 68
kern.maxfiles: 12328
kern.maxfilesperproc: 11095

and raised the openfiles limit on the www process, and checked lsof
which says there isn't much open.  so it looks like some kind of
shared memory / locking thing, except that ipcs tells me not much: i
guess it's not a sysv ipc thing.

does anyone know how to clean house for this kind of stuff?

thanks in advance, B

_______________________________________________
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Reply via email to