This version has a known memory leak in the AJP/1.3 Connector. The work-around is to set: request.registerRequests=false in your $CATALINA_HOME/conf/jk2.properties file.
"Carroll, Simon, CND" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] Hi, I have the following problem and I was wondering if anyone else had a similar issue which they managed to resolve. I have a web application which is installed on 2 tomcat instances (version 4.1.30 using jdk 1.4.2_05), each on separate boxes. I have 2 apache (version 2) instances, also on separate boxes, which are configured to pass requests to the tomcat instances via the mod_jk plugin. The configuration works as expected and the requests from apache are equally balanced between the two tomcat instances. The issue I have is that the web application installed on tomcat seems to have a memory leak and if it is left to run for a prolonged period the tomcat processes begin to hog the CPU. To counter this I have set up a cron job which restarts the tomcat instances every morning, the restarts are staggered so that the first instance is restarted at 07:00 and the second at 07:20. However after the restarts I can see from the apache logs that a number of HTTP requests are taking 190 seconds to process. From the logging I have determined that the hold up occurs when the request is passed to tomcat. I see this issue for approximately 3 hours and then it clears itself up until the following day when we restart the tomcat instances again. Does anyone know why this is occuring? Should I be restarting the apache instances after I have restarted the tomcat instances? Is there anyway I can clear down the CPU usage without restarting tomcat? If anyone can help me with this issue I would be very grateful. Thanks in advance Simon --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
