Matthew mentions the challenge of getting thread dumps in this later note. Just to clarify (as I wrote my last note before reading this one), the "thread dump" referred to here is in fact the same general idea as the stack trace I referred to. The good news again, if you use any of the 3 monitors, is that it's just a push-button operation, or you can even get it sent to you in an email when a problem situation is detected (in FusionReactor and CF8 Monitor, I know).
Different folks use "thread dump" and "stack trace" to mean the same thing, or they can differ. In my sense of the words (and which FusionReactor and the CF8 Monitor uses), a thread dump is a stack trace for every thread, and a stack trace for one thread is what tells you what line of code is running and what java objects/methods are being called on the basis of that line of code. Without those tools, as has been alluded to here, you can get a full thread dump by some other methods that are a bit more complicated, and then you have to hunt through them to find the stack trace for a thread of interest. All doable, just easier with the CF monitor tools. Hope that's helpful. To answer your last question, Matthew, the ability to view how large variables are (or all vars in a given scope) is something that's enabled only in the CF8 Monitor (which runs only on CF8 Enterprise and Developer, sadly). Neither FusionReactor nor SeeFusion will tell you. There are also ways to get at the information by using the undocumented objects like coldfusion.runtime.SessionTracker. A google search for those will turn up ways to access that info. /charlie -----Original Message----- From: cfaussie@googlegroups.com [mailto:cfaus...@googlegroups.com] On Behalf Of Matthew Sent: Tuesday, January 06, 2009 8:21 PM To: cfaussie Subject: [cfaussie] Re: JRUN hanging Hi Kai Thanks for the tip. I have already tried connecting the JConsole from my local dev machine to the live server but can't get it to connect. Perhaps I've got the port wrong. How do you work out which port the CF JVM is listening on? I'm looking into the Thread Dump idea (I've just finished reading the Adobe article you spoke of) however there is a lot to understand. Anyway, I've switched everything on so I can take a Thread Dump next time the server appears to hang (or what is really called SPINNING) however I don't think this will help in any great way. I'm pretty certain that whatever monitoring solution I find it will tell me what I already know, which is that because a big part of the website relies on a web service call to a 3rd party which goes down sometimes and because CFINVOKE won't obey the timeout parameter then I'll always have this problem. I need to work out a way to detect that the web service is out and disable it before too many users on the website instantiate a connection which won't timeout properly and cause the server to go into a spin. I will continue to investigate. If anyone else wants to chime in please feel free! By the way; is there anyway to get a report/graph which shows how much memory each variable is taking up i.e. can you take a snapshot at a moment in time to show all the session variables and how much memory they are costing? Cheers Matthew --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "cfaussie" group. To post to this group, send email to cfaussie@googlegroups.com To unsubscribe from this group, send email to cfaussie+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/cfaussie?hl=en -~----------~----~----~----~------~----~------~--~---