httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael <mich...@asstr.org>
Subject [users@httpd] apache 2.0.40 children not restarting on request?
Date Tue, 27 Aug 2002 20:43:00 GMT
	I have an apache 2.0.40 server using the worker MPM that's been running
on a linux 2.2 kernel for about 4 days now, and every day I run
apache -k graceful after rotating the logs (the program is named apache
instead of httpd). It looks like that's not working for killing the child
processes, though. A look at ps shows 271 apache processes running, most
of which have been started sometime today. There are 6, though, that have
not:

root     12526  0.0  0.2  5164 2152 ?        S    Aug24   0:51 lt-apache
web      23633  0.0  1.7 37328 16668 ?       S    Aug25   0:00 lt-apache
web      23635  0.0  1.7 37328 16668 ?       S    Aug25   0:04 lt-apache
web      23681  0.0  1.7 37328 16668 ?       S    Aug25   0:00 lt-apache
web       2754  0.0  1.6 37128 16372 ?       S    Aug25   0:00 lt-apache
web       2755  0.0  1.6 37128 16372 ?       S    Aug25   0:04 lt-apache
web       2797  0.0  1.6 37128 16372 ?       S    Aug25   0:00 lt-apache

	One of them is my root process, one of them is presumably cgid, and
the other five look to be the control processes for the threads I have
running. A look in /proc for these processes shows that the root process
has all my current log files open. The other six, though, have the log
files from the day they were started open (a look in /proc/2797/fd shows
that most of the log files it has open are deleted).

	This doesn't cause any problems with logging, since the threads that are
spawned are using the correct log files. The problem it causes is that
the 500MB that the deleted logfiles take up won't be freed from the
filesystem until I stop and restart apache.

	I've tried apache -k graceful, apache -k restart, and kill -HUP 12526
and none of them have caused the children to close their logfiles.

	More info: this may be related to virtual hosts, since a deeper look
at the child process fd table shows that the main error.log file is the
current error.log file; it's just the access and error logs defined in
the virtual hosts that aren't getting reopened correctly.

	Has anyone seen this issue before?

- Michael


---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message