DO NOT REPLY [Bug 46215] Race condition in bybusyness algorithm

[bu...@apache.org]

https://issues.apache.org/bugzilla/show_bug.cgi?id=46215





--- Comment #1 from Thomas Binder <bi...@arago.de>  2009-04-21 06:20:17 PST ---
Created an attachment (id=23519)
 --> (https://issues.apache.org/bugzilla/attachment.cgi?id=23519)
Patch using apr_atomic functions, also adding "busyness" column in
balancer-manager

I've experienced the same issue, but the proposed patch did not completely
resolve it for me. Running something like

ab -k -c 20000 -n 1000 http://example.com/

would still cause the "busy" counter to stay greater zero from time to time.

I've therefore tried a different approach, using functions from apr_atomic.h to
increase / decrease the "busy" counter instead of locking the balancer. So far,
the problem hasn't reappeared even after running the "benchmark" above several
times.

The patch also adds a "busyness" column to the output of balancer-manager, to
make it easier to check the current behaviour of the balancer. If you don't
want / need that, simply remove the hunks "@@ -830,6 +831,7 @@" and "@@ -861,6
+863,7 @@" from the patch.

-- 
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org