So, we have seen a series of problems lately that have been very frustrating to diagnose. I keep dreadfully remembering Dave Livingston's statement Tuesday night regarding one of the times to suspect you may have been breached ("If you find unexplained behavior on your server") but I am finding no evidence of this as of yet.
Last week we began seeing some request timeouts which I finally determined was a memory problem on the SQL Server (more here...). Once that problem was in the rear view mirror, we began seeing occasional spikes on the Jrun process where it would peg out the processor until the CF service was restarted. We have a fairly high-traffic application and we are seeing this once or twice a day. Each time I have found it to be (much like the SQL Server issue) a case where the available threads would fill up, and you could see the queued thread number continue to rise behind it. I have found no consistent instigator of that problem. Although a pain, and something that needs to be addressed, I am finding a new and more mind-boggling problem...
Since Monday, we are seeing problems where a page load happens but it behaves as if not all the HTML code in the page was there. This only happens in IE, and seems to happen for user. For instance, on a log-in screen.. The user submits... they go to the authenticated page, and some _javascript_-dependent menu is not working. Or they might go to a page and it will be just a blank screen (including the CF administrator). Then when refreshed, the page works fine. When this happens, it is often accompanied by a little yellow triangle on the status bar saying that an error has occurred at position 1 of line 2, which has often been a blank line. An interesting point in this is that if you view source on the page that failed, copy the HTML into a new blank template, and run it, it behaves just fine. The code is there (kind of ruling out CF at that point), but somewhere in the presentation of that between the web server and the browser, something just gets goofed up. This problem is consistent and reproduce-able, but only on that machine. The same code on the development server does not behave that way. Does anyone have any clue or advice at all?
Thanks!
--
~Dave Shuck
[EMAIL PROTECTED]
www.daveshuck.com
_______________________________________________ Reply to DFWCFUG: [email protected] Subscribe/Unsubscribe: http://lists1.safesecureweb.com/mailman/listinfo/list List Archives: http://www.mail-archive.com/list%40list.dfwcfug.org/ http://www.mail-archive.com/list%40dfwcfug.org/ DFWCFUG Sponsors: www.HostMySite.com www.teksystems.com/
