{"id":1150,"date":"2012-08-27T17:05:54","date_gmt":"2012-08-27T09:05:54","guid":{"rendered":"http:\/\/rmohan.com\/?p=1150"},"modified":"2012-08-27T17:09:55","modified_gmt":"2012-08-27T09:09:55","slug":"tuning-apache-tomcat","status":"publish","type":"post","link":"https:\/\/mohan.sg\/?p=1150","title":{"rendered":"Tuning Apache + Tomcat"},"content":{"rendered":"

ATG: Tuning Apache + Tomcat<\/h1>\n

I did some tuning on Apache which vastly decreased response times.<\/p>\n

The tests show increase vast decrease in response. (see table below) which is due to the following improvements made effective by the tuning (Increased maxclients, KeepAlive on. KeepAliveTimeout etc):<\/p>\n

1) Reusable connections to handle multiple requests from the same client, which in return causes less connections from client to the web-server and decreases response times.<\/p>\n

2) Because of Point 1, apache now needs less Httpd clients to server more load, hence decrease in resource usage (Memory).<\/p>\n

4) MaxClients has been increased to 1000, hence if required apache can now expand to serve more load without the queue buildup which causes high latencies.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Request<\/strong><\/td>\n100<\/strong><\/td>\n100 (Tuned)<\/strong><\/td>\nPerf gain (%)<\/strong><\/td>\n400<\/strong><\/td>\n400(Tuned)<\/strong><\/td>\nPerf gain (%)<\/strong><\/td>\n<\/tr>\n
1-homePage<\/strong><\/td>\n850.21<\/td>\n253.52<\/td>\n70.18<\/td>\n4812.72<\/td>\n533.55<\/td>\n88.91<\/td>\n<\/tr>\n
2-login<\/strong><\/td>\n1132.27<\/td>\n593.72<\/td>\n47.56<\/td>\n5411.01<\/td>\n1186.64<\/td>\n78.07<\/td>\n<\/tr>\n
3-1-search<\/strong><\/td>\n1698.79<\/td>\n1616.14<\/td>\n4.87<\/td>\n4303.77<\/td>\n4366.99<\/td>\n-1.47<\/td>\n<\/tr>\n
3-2-searchPagination<\/strong><\/td>\n1822.23<\/td>\n1005.70<\/td>\n44.81<\/td>\n8810.29<\/td>\n3429.22<\/td>\n61.08<\/td>\n<\/tr>\n
4-productDetails<\/strong><\/td>\n1043.62<\/td>\n398.56<\/td>\n61.81<\/td>\n5454.44<\/td>\n686.17<\/td>\n87.42<\/td>\n<\/tr>\n
5-cartOverviewDeleteAllProducts<\/strong><\/td>\n328.95<\/td>\n320.88<\/td>\n2.45<\/td>\n712.25<\/td>\n514.25<\/td>\n27.80<\/td>\n<\/tr>\n
5-cartOverviewViaMiniCart<\/strong><\/td>\n1026.01<\/td>\n421.22<\/td>\n58.95<\/td>\n5248.49<\/td>\n743.63<\/td>\n85.83<\/td>\n<\/tr>\n
6-cartOverviewAddProduct<\/strong><\/td>\n351.52<\/td>\n336.90<\/td>\n4.16<\/td>\n873.58<\/td>\n824.36<\/td>\n5.63<\/td>\n<\/tr>\n
7-cartPreview<\/strong><\/td>\n1352.69<\/td>\n665.76<\/td>\n50.78<\/td>\n6218.38<\/td>\n1162.07<\/td>\n81.31<\/td>\n<\/tr>\n
8-cartConfirm<\/strong><\/td>\n908.65<\/td>\n848.04<\/td>\n6.67<\/td>\n1662.67<\/td>\n1251.99<\/td>\n24.70<\/td>\n<\/tr>\n
9-logout<\/strong><\/td>\n840.03<\/td>\n214.11<\/td>\n74.51<\/td>\n5095.36<\/td>\n523.79<\/td>\n89.72<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The changes will require the following changes in the HTTP.cong (\/etc\/httpd\/conf\/httpd.conf)<\/p>\n

\u2014-
KeepAlive on
MaxKeepAliveRequests 100
KeepAliveTimeout 2<\/p>\n

<IfModule prefork.c>
StartServers 10
MinSpareServers 10
MaxSpareServers 20
ServerLimit 1000
MaxClients 1000
MaxRequestsPerChild 4000
<\/IfModule><\/p>\n

<IfModule worker.c>
StartServers 10
MaxClients 1000
MinSpareThreads 25
MaxSpareThreads 75
ThreadsPerChild 25
MaxRequestsPerChild 1000
<\/IfModule>
\u2014<\/p>\n

Reference: http:\/\/blog.springsource.org\/2008\/10\/14\/optimising-and-tuning-apache-tomcat-part-2\/<\/p>\n

Note: the above improves performance but exausts mod_jk thread pools, as the response are shorter now, hence more load.<\/p>\n

Read the following on MPMs to fine tune it further:<\/p>\n

Choosing an MPM<\/h3>\n

Apache 2.x supports pluggable concurrency models, called\u00a0Multi-Processing Modules<\/a> (MPMs). When building Apache, you must choose an MPM to use. There are platform-specific MPMs for some platforms:\u00a0beos<\/a><\/code>,\u00a0mpm_netware<\/a><\/code>,\u00a0mpmt_os2<\/a><\/code>, and\u00a0mpm_winnt<\/a><\/code>. For general Unix-type systems, there are several MPMs from which to choose. The choice of MPM can affect the speed and scalability of the httpd:<\/p>\n