Ok I just eyeballed the Verbose GC log and there is definitely something amiss. In 8.9 hours of uptime there were 786 Full GC's, that is one every 41 seconds the latest ones are taking 3.8 seconds and as I mentioned earlier everything stops dead in a Full GC. So we know that there are far too many Full GC's, they should occur when necessary but not at this frequency. The next step is to add this argument to jvm.config:
-XX:+DisableExplicitGC This will stop any explicit Full GC's. Explicit Full GC's are those called by either the application or an outside server such as RMI and they are almost always unnecessary and disruptive. Those are the findings so far. Kind Regards - Mike Brunt CTO and Lead Architect - Webapper Services LLC Tel 562.243.6255 Web Site: http://www.webapper.com Blog: http://www.webapper.net Tools: http://www.seefusion.com -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sean Corfield Sent: Wednesday, May 24, 2006 4:49 PM To: [email protected] Subject: Re: [Reactor For CF] Been out of touch, but trying to get back On 5/24/06, Doug Hughes <[EMAIL PROTECTED]> wrote: > I understand that it's natural behavior. However, based on past > experience, a site as small as my blog shouldn't as much RAM as it's using. OK, just wanted to check (and to make the point for others here since I see this assumption that "there is a memory leak" based on natural JVM memory usage far too often). Having seen the graph, I'd agree with Mike that there's something seriously wrong here: those repeated full GCs are very worrying! -- Sean A Corfield -- http://corfield.org/ Got frameworks? "If you're not annoying somebody, you're not really alive." -- Margaret Atwood -- Reactor for ColdFusion Mailing List -- [email protected] -- Archives at http://www.mail-archive.com/reactor%40doughughes.net/ -- Reactor for ColdFusion Mailing List -- [email protected] -- Archives at http://www.mail-archive.com/reactor%40doughughes.net/
