I have an old Red Hat 9 server where a Tomcat 5.5 is running with 10 big sites. Those sites use Hibernate, Lucene, have a complex CMS and many other big libraries.
I’m running out of memory, even if the server has 4 giga: I can give to JVM max 1.5 giga total. It’s a known limit of JVM 32 bit on some systems (there are a lot of information about that on the net).
There are a couple solutions one can use to solve the problem. Seen that server has memory, one can use two IP address and two Tomcat with connectors binded on different IP. The Tomcat are so responding on the same port (if it’s a problem to change the port number).
Other way is to sun two Tomcat instaces on two different ports (like 8080 and 8081) and reconfigure the firewall/apache to route some virtual host on first port and some other on second port. That one is my case.
Clearly the portals have been grouped and distributed on the two Tomcat instances.
Remember to change the “Tomcat server port” (8006) for a second instance: that port is used to shutdown the server instance.
