Re: connection pool issues with mod_proxy_ajp when tomcat is remote
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Russell, On 4/13/2010 3:53 PM, Russell Uman wrote: > tomcat server.xml: > enableLookups="false" redirectPort="8443" protocol="AJP/1.3" > minSpareThreads="5" maxThreads="300" > connectionTimeout="12" URIEncoding="UTF-8" /> > > apache 2.0.52/mod_jk 1.2.19 workers.properties > worker.default.port=8009 > worker.default.type=ajp13 > worker.default.connect_timeout=12 > worker.default.connection_pool_timeout=120 > worker.default.prepost_timeout=12 > worker.default.reply_timeout=180 > worker.default.retries=1 > > apache 2.2.3/mod_proxy_ajp > ProxyPass / ajp://example.org:8009/ timeout=1800 keepalive=on 120 seconds is a long time to wait for an HTTP request to show up. Is that really what you wanted? Also, 1800 seconds (aka 30 minutes) is a long time for a response. Are you sure these are reasonable values for your application? You have set prepost_timeout, but haven't set a ping_mode. Are you seeing pings being sent? 120 seconds is a long time to wait for an AJP connection to be made (or to reply to a ping). Again, are these values sane? Glad to see that mod_jk is working well for you. Pid is right: 2.2.3 was a lng time ago in terms of mod_proxy_ajp. It's unfortunate that package managers (RedHat, CentOS, Debian, etc.) move relatively slowly when it comes to incorporating features such as these. It's understandable, as most admins want stability over everything else. Using mod_jk will give you the best experience if you are stuck with a package-managed version of httpd, but are free to install hand-rolled modules such as mod_jk. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkvFJ1kACgkQ9CaO5/Lv0PB7vQCeOZa3xHx7UlsyhZkrl92UspSq VZwAn2pxWrRVhn1r/8BoS0XE/MtEZefK =Uujk -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: connection pool issues with mod_proxy_ajp when tomcat is remote
On 13/04/2010 23:25, Russell Uman wrote: we're trying to move from apache 2.0.52 with mod_jk 1.2.19 to apache 2.2.3 with mod_proxy_ajp in front of tomcat 5.5.28 mod_proxy_ajp was fairly new in 2.2.3, there are significant improvements in it since then. Upgrading is essential if you want to use it in production. Otherwise, go back to mod_jk - but do try to get the latest version. thanks! i'm testing now with mod_jk and apache 2.2.3 and things seem to be more cooperative. 2.2.3 is quite old, 28 Jul 2006 according to the archive, 2.2.15 is the current release - you should definitely use the latest version and if not that a very recent one. p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: connection pool issues with mod_proxy_ajp when tomcat is remote
we're trying to move from apache 2.0.52 with mod_jk 1.2.19 to apache 2.2.3 with mod_proxy_ajp in front of tomcat 5.5.28 mod_proxy_ajp was fairly new in 2.2.3, there are significant improvements in it since then. Upgrading is essential if you want to use it in production. Otherwise, go back to mod_jk - but do try to get the latest version. thanks! i'm testing now with mod_jk and apache 2.2.3 and things seem to be more cooperative. -- russell uman firebus d-_-b - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: connection pool issues with mod_proxy_ajp when tomcat is remote
On 13/04/2010 20:53, Russell Uman wrote: howdy! we're trying to move from apache 2.0.52 with mod_jk 1.2.19 to apache 2.2.3 with mod_proxy_ajp in front of tomcat 5.5.28 mod_proxy_ajp was fairly new in 2.2.3, there are significant improvements in it since then. Upgrading is essential if you want to use it in production. Otherwise, go back to mod_jk - but do try to get the latest version. and we've run into a troubling bug/configuration/networking problem with apache 2.0 and mod_proxy, things are very well behaved. apache has a connection pool of tomcat threads, and when apache kills a child process the corresponding tomcat thread is closed. with apache 2.2 and mod_proxy_ajp, when apache kills a child process, the sometimes the tomcat thread is left hanging in the keep alive state. tomcat waits 120 seconds before terminating the hanging thread. this seems to happen more often under higher traffic, and more often when tomcat is not running on the same server as apache. under traffic, this results in a large number of tomcat threads that have lost their connection to apache and are waiting to timeout, eventually leaving no free threads for legitimate requests. can anyone suggest what could be causing this, or how to troubleshoot? i've tried using the keepalive=on parameter to ProxyPass with no luck. here are some relevant configuration details: with both versions of apache we're using the worker MPM tomcat server.xml: apache 2.0.52/mod_jk 1.2.19 workers.properties worker.default.port=8009 worker.default.type=ajp13 worker.default.connect_timeout=12 worker.default.connection_pool_timeout=120 worker.default.prepost_timeout=12 worker.default.reply_timeout=180 worker.default.retries=1 apache 2.2.3/mod_proxy_ajp ProxyPass / ajp://example.org:8009/ timeout=1800 keepalive=on thanks in advance for any help, and if there's another mailing list for mod_proxy_ajp questions that would be a better recipient please let me know! Nope, hereabouts is good enough. p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
connection pool issues with mod_proxy_ajp when tomcat is remote
howdy! we're trying to move from apache 2.0.52 with mod_jk 1.2.19 to apache 2.2.3 with mod_proxy_ajp in front of tomcat 5.5.28 and we've run into a troubling bug/configuration/networking problem with apache 2.0 and mod_proxy, things are very well behaved. apache has a connection pool of tomcat threads, and when apache kills a child process the corresponding tomcat thread is closed. with apache 2.2 and mod_proxy_ajp, when apache kills a child process, the sometimes the tomcat thread is left hanging in the keep alive state. tomcat waits 120 seconds before terminating the hanging thread. this seems to happen more often under higher traffic, and more often when tomcat is not running on the same server as apache. under traffic, this results in a large number of tomcat threads that have lost their connection to apache and are waiting to timeout, eventually leaving no free threads for legitimate requests. can anyone suggest what could be causing this, or how to troubleshoot? i've tried using the keepalive=on parameter to ProxyPass with no luck. here are some relevant configuration details: with both versions of apache we're using the worker MPM tomcat server.xml: apache 2.0.52/mod_jk 1.2.19 workers.properties worker.default.port=8009 worker.default.type=ajp13 worker.default.connect_timeout=12 worker.default.connection_pool_timeout=120 worker.default.prepost_timeout=12 worker.default.reply_timeout=180 worker.default.retries=1 apache 2.2.3/mod_proxy_ajp ProxyPass / ajp://example.org:8009/ timeout=1800 keepalive=on thanks in advance for any help, and if there's another mailing list for mod_proxy_ajp questions that would be a better recipient please let me know! -- russell uman firebus d-_-b - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org