4 Replies Latest reply on Jun 18, 2008 8:46 PM by Philhurrell

    jrun.exe uses 100% of CPU

    Philhurrell
      Anyone got any pointers as to this issue.

      Win NT2003 server using Apache + mysql DB + CF8 Standard
      jrun.exe has recently begun to use up 100% of CPU time as monitored in taskmgr

      We thought this was related to using client variables and filling up the registry, but have turned this off in application and cleaned the registry back from 500mb to 18mb! (also tried using DB to store them)

      Have searched google - lots of examples of issue in older versions of CF, but this issue is specifically with MX8

      It is possible that some runaway process / query is causing the issue but finding it is proving elusive, is there a way to get more info from coldfusion as to what it is doing ?

      Restarting Coldfsion service fixes issue for about 2 hours then it re-occurs.
        • 1. Re: jrun.exe uses 100% of CPU
          Dan Bracuk Level 5
          We've had that happen to us from time to time. The most frequent cause was a combination of a slow page and inpatient user who kept clicking the button/link.
          • 2. Re: jrun.exe uses 100% of CPU
            speedpedal Level 1
            Since you are on CF8, this would be a good time to use the Server Monitoring Tools in CFADMIN. Also, adjust your request tuning inside CFADMIN. You can make your adjustments based on the data gathered on the server monitor tool.

            HTH
            • 3. Re: jrun.exe uses 100% of CPU
              Philhurrell Level 1
              We're going to look into slow pages, but most of them appear to work quickly.
              Unfortunalty this is CF8 Standard and doesnt allow access to
              Server Monitoring > Server Monitor

              "This feature is not available in this edition of the ColdFusion server. "

              Unless there is a way around this.....
              • 4. Re: jrun.exe uses 100% of CPU
                Philhurrell Level 1
                Ok, the problem is resolved. For anyone who is interested. We looked at Apache access.log and noticed a HUGE number of PROPFIND requests narrowing this down it seems opening a word document from our CF application then closing it caused a loop to be set up.
                Essentially when the doc closed it tried to open the index.cfm which redirected to home.cfm and somehow went back to index.cfm

                Changing apache set up to use home.cfm instead of index.cfm fixed the issue.