Hi I have been seeing very similar symptoms as you in the last couple of weeks.We have a high volume web site, running Linux 2.6.9 and Apache 2.2.3, and the hardware specification is very similar to yours.
We are using mod_php, not mod_python.About once a day, the CPU load on the web server drops a lot, and is almost 0%. The reason is that there are no available worker threads, and the reason for that is that most worker threads are struck in the "W" state (sending reply).
The only way of freeing up the worker threads is to restart apache.I suggest you use the Apache mod_status / server-status to check the state of your worker threads. The "ExtendedStatus" needs to be "On". Are there a lot of threads in "W" state, with "SS - Seconds since beginning of most recent request" values like 500, 2000 or even more ?
In fact, we have seen the SS number as high as 30000. We think we are experiencing something very similar to http://mail-archives.apache.org/mod_mbox/httpd-users/200611.mbox/%3C20061114084333.GC10809@xxxxxxxxxxxxxxxxxxxxxxxx%3EI.e., we think that is might explain why the threads ends up in the "W" state.
When running netstat, do you see a lot of sockets in CLOSE_WAIT state ? Regards Alf Hogemark --------------------------------------------------------------------- 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@xxxxxxxxxxxxxxxx " from the digest: users-digest-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|