DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://nagoya.apache.org/bugzilla/show_bug.cgi?id=16872>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=16872 Hang upon heavy load Summary: Hang upon heavy load Product: Apache httpd-2.0 Version: 2.0.44 Platform: PC OS/Version: Windows NT/2K Status: NEW Severity: Normal Priority: Other Component: Core AssignedTo: [email protected] ReportedBy: [EMAIL PROTECTED] Win 2K, no-ssl standard build (apache_2.0.44-win32-x86-no_ssl.msi). Hit a static HTML page using 'ab' and the machine after, say, 2000 hits gets 100% CPU saturation on the Apache process, cannot serve any more pages and does not recover. It complains "Server ran out of threads to serve requests". The log file (given I started Apache in Console mode) says: [Thu Feb 06 14:01:03 2003] [notice] Parent: Created child process 724 [Thu Feb 06 14:01:03 2003] [notice] Child 724: Child process is running [Thu Feb 06 14:01:03 2003] [notice] Child 724: Acquired the start mutex. [Thu Feb 06 14:01:03 2003] [notice] Child 724: Starting 250 worker threads. [Thu Feb 06 14:05:02 2003] [warn] Server ran out of threads to serve requests. Consider raising the ThreadsPerChild setting *****and now I press CTRL-C to halt Apache****** [Thu Feb 06 14:05:39 2003] [notice] Parent: Received shutdown signal -- Shutting down the server. [Thu Feb 06 14:05:39 2003] [notice] Child 724: Exit event signaled. Child process is ending. [Thu Feb 06 14:05:40 2003] [notice] Child 724: Released the start mutex [Thu Feb 06 14:05:41 2003] [notice] Child 724: Waiting for 250 worker threads to exit. [Thu Feb 06 14:06:09 2003] [notice] Parent: Forcing termination of child process 184 My ThreadsPerChild is the default for the Win build... ie 250. This seemed very similar to bug 13991; however this case was reported on 2.0.42 and "fixed" in 2.0.43 or rather the problem didn't reoccur. I downgraded to 2.0.43 and the problem persisted. Trying on a second Win2K box does not show the problem, therefore I have to assume some other environmental conditions are at play. The failing box is Win2K (I will supply Service Pack info later) at a major ISP behind a firewall (unknown type). The loadtest was run on a Linux box with highbandwidth to the Windows box. The load command was: ./ab -n 10000 -c 400 http://111.222.333.444/ The ok Win2K box is on the same local network as the Linux box as is Service Pack 3. Would this be some IP stack issue? Incidentally, I ran spawn-of-Satan IIS on the failing box; it performed fine under the load test. What other information would be useful? --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
