Well the only thing I've seen on our boxes that is similar is if a parsing error occurs (like  a cfset missing an opening or closing bracket).  What happens when that page is run, JRun starts eating RAM until the server grinds to a halt.

This happens fairly quickly though and I would suspect that that you've got something else going on...because that kind of traffic with the issue we've seen would toast your servers much faster than they seem to be dying.

HTH

Cheers

Bryan Stevenson B.Comm.
VP & Director of E-Commerce Development
Electric Edge Systems Group Inc.
t. 250.920.8830
e. [EMAIL PROTECTED]

---------------------------------------------------------
Macromedia Associate Partner
www.macromedia.com
---------------------------------------------------------
Vancouver Island ColdFusion Users Group
Founder & Director
www.cfug-vancouverisland.com
  ----- Original Message -----
  From: Jeremy Bruck
  To: CF-Talk
  Sent: Monday, May 17, 2004 9:18 AM
  Subject: Maxing out running requests then server hanging -- Help

  Everyone,

  We put out a new release of our web-application and have been having
  server problems problems ever since.  All will run fine and then all of
  a sudden running requests will start creeping up and then CFMX hangs --
  about every 4/6ish hours.  Before the new release the most problems we
  would ever have was database locks and we could either wait for the
  database lock to finish or kill it -- none of this has been database
  locks.  Majority of the time both boxes (2 web servers hardware
  loadbalanced) will be using up about 700ish meg of ram (both have 1 gig
  of ram) when they are hung.  We tried even running CFMX from a command
  line to get a java stack trace when it hangs, but it never would create
  a dump of the stack trace for us even -- toooo hung I am assuming.

  Then this past week for a solid hour the running requests would keep
  running up to the max (25/box) and then would hang -- all we could do
  was continually restart the service which we did about 20 times on both
  boxes -- we tried rebooting in the middle of all this as well which
  didn't help either.  The weird thing was after about an hour it all went
  back to the 4/6ish hour hand routine.

  Any ideas on what could be causing the hangs?  Since the new release
  used CFC's heavily could this be causing any of the issues?  Any ideas
  on better ways to trace/debug it?

  Here are the details on our setup (all software is fully patched):
      2 hardware loadbalanced webservers running Win2K with CFMX 6.1 with
  an Apache webserver (2.046)
      Seperate box running MS SQL Server
      Approximate traffic is 12K to 15K unique visitors/day

  Thanks,
  Jeremy

  ----------------------------------
  Jeremy Bruck
  ManufacturingQuote
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]

Reply via email to