Chuck, On Tue, Mar 18, 2008 at 2:39 AM, Kevin King <[EMAIL PROTECTED]> wrote:
> Write a quick JSP page that returns the amount of server memory available > and monitor this when the system starts to dog; sounds like maybe the Java > garbage collector might need to be run more frequently. > This page has source code for a thread-dump jsp page so you can see what is going on. http://www.javaspecialists.eu/archive/Issue132.html IIRC the Tomcat manager should also tell you memory available. I'm not sure GC would be a problem these days - possibly on older jvm's but unless you are doing climate modelling, I'm not sure that would be a problem. However, if a reference has been left to the objects then they can't be GC'ed anyway. This issue has some test code on GC but they use "an array of the last million created objects." for testing to give you an idea of performance. The page also has some pointers to tools to see what GC is doing. I've never had to do that myself so no idea how difficult it is. http://www.javaspecialists.eu/archive/Issue115.html Any locking or synchronized code going on that could be holding things up? Has the JSP been set to service only one request at a time? Can you change the start up parameters for Tomcat to give it more memory etc? How are the connections to U2 being managed - could there be a connection pool or similar with locking issues or running out of connections? As there is nothing untoward in the Tomcat logs I presume you aren't getting exceptions thrown all over the place when it grinds to a halt. Please let us know how you get on with this one. Regards, Adrian Auckland, NZ ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
