Page Comparison

[11:22:07 CDT(-0500)] <lfuller1> Got pulled in to look over a problem where a university using uportal is running out of memory. I showed them how to take a heap dump, which they did the next time they ran into trouble. At present it seems to take them roughly 24 hours to get in a bad situation.
[11:22:15 CDT(-0500)] <lfuller1> Looking at the heap dump they saw this:
[11:22:18 CDT(-0500)] <lfuller1> One instance of "org.jasig.portal.events.handlers.QueueingEventHandler" loaded by "org.apache.catalina.loader.WebappClassLoader @ 0x2aaab3b25ba0" occupies 1,814,492,936 (77.51%) bytes.
[11:22:43 CDT(-0500)] <lfuller1> Which... looks as if it could be related to the stats. Anyone run into a similar issue?
[11:22:43 CDT(-0500)] <EricDalquist> is this the same issue that was talked about on the user list?
[11:22:59 CDT(-0500)] <lfuller1> no idea. Will go look at the user list.
[11:23:02 CDT(-0500)] <EricDalquist> where it sounded like Quartz magically stopped calling the stats flush list
[11:23:06 CDT(-0500)] <EricDalquist> er flush trigger
[11:23:14 CDT(-0500)] <lfuller1> lame.
[11:23:25 CDT(-0500)] <lfuller1> this has been happening pretty consistently for these guys
[11:23:59 CDT(-0500)] <EricDalquist> yeah
[11:24:11 CDT(-0500)] <EricDalquist> I'm trying to remember who it was happening too
[11:24:12 CDT(-0500)] <lfuller1> They just discovered the problem this morning though.
[11:24:14 CDT(-0500)] <EricDalquist> was it parker grimes?
[11:24:50 CDT(-0500)] <EricDalquist> nope ... Paul Gazda
[11:25:06 CDT(-0500)] <EricDalquist> hrm
[11:25:09 CDT(-0500)] <EricDalquist> maybe not that either
[11:25:26 CDT(-0500)] <lfuller1> did they end up resolving the issue or just punt and comment out the stats?
[11:25:56 CDT(-0500)] <EricDalquist> ah it was Curtis
[11:25:59 CDT(-0500)] <EricDalquist> finally found the thread
[11:26:43 CDT(-0500)] <EricDalquist> http://jasig.275507.n4.nabble.com/uportal-statistics-td2262749.html#a2262749
[11:27:07 CDT(-0500)] <EricDalquist> I'm not sure I ever got a really good answer about what debugging they added
[11:27:40 CDT(-0500)] <EricDalquist> my thought was to turn quartz logging to debug
[11:27:49 CDT(-0500)] <EricDalquist> and add logging at the entry point to the stats flush call
[11:27:49 CDT(-0500)] <lfuller1> yeah, once you said the name I found the thread..
[11:29:19 CDT(-0500)] <EricDalquist> and see if it could be figured out why the flush call wasn't triggering
[15:22:56 CDT(-0500)] <EricDalquist> lfuller1: what ever happened with that portlet entity cleanup patch?
[15:23:03 CDT(-0500)] <EricDalquist> I can't remember where we left that discussion