3 Replies Latest reply on May 1, 2017 10:46 AM by greglane

    FileMaker WebDirect excessive polling

    greglane

      Hello,

       

      A client's solution is hosted with FileMaker Cloud. As we were updating the solution, we found that a large amount of disk space was being consumed by log files. Since January, the /FileMakerData/Logs/HTTPServer/ folder contains about 17 GB of log files. That seems odd because there are only a handful of WebDirect users and they each typically connect for less than 10 minutes a few times per week. The FileMaker WebDirect Session Timeout is set for 60 minutes.

       

      Digging into the ssl_request_log files shows thousands of log entries like the following with between 10 and 20 nearly identical entries every second. This appears to happen after a user's session is "dead" (after the last 'POST /fmi/webd/HEARTBEAT/' for that user) and goes on for hours at a time. It doesn't happen every day or with every session. It does happen with multiple users. On days when the problem occurs, there are between 2 and 30 ten MB ssl_request_log files and a similar number of ssl_access_log files. 

       

      [20/Apr/2017:17:32:04 -0500] 209.63.##.### TLSv1.2 ECDHE-RSA-AES128-SHA256 "GET /fmi/webd/PUSH?v-uiId=0&v-csrfToken=af875###-4594-449e-96d3-164c36b#####&X-Atmosphere-tracking-id=f3e9f8fb-94e9-4b7b-bb33-21fcb05#####&X-Atmosphere-Framework=2.2.6.vaadin5-jquery&X-Atmosphere-Transport=long-polling&X-Atmosphere-TrackMessageSize=true&Content-Type=application%2Fjson%3B%20charset%3DUTF-8&X-atmo-protocol=true&_=1492727524459 HTTP/1.1" 2256

       

      The log entries for a given user are identical except for the timestamp and the last "_" parameter, which tends to be a series like this:

       

      &_=1492727535644

      &_=1492727535697

      &_=1492727535755

      &_=1492727535812

      &_=1492727535868

       

      So far, we haven't been able to reliably reproduce it or determine a cause. Is this a known issue? Is there something in our solution that might be triggering it? We upgraded this FileMaker Cloud instance to version 1.15.1.35 today, so this has all been happening with the previous version.

       

      Also, does FileMaker Cloud ever prune the log files or should that be done manually via ssh?

       

      Thanks,

      Greg

        • 1. Re: FileMaker WebDirect excessive polling
          TSGal

          greglane:

           

          Thank you for your post.

           

          This is not a known issue.  Our Development and Testing departments would like to look at your instance.  I have sent you a private message requesting your instance ID that I will then forward on to Development and Testing for review.

           

          TSGal

          FileMaker, Inc.

          • 2. Re: FileMaker WebDirect excessive polling
            TSGal

            greglane:

             

            Our Testing department has been able to replicate the issue, as the issue appears to be related to clients timing out (the comments from Testing are a bit cryptic).  All information has been sent to Development for further review.

             

            TSGal

            FileMaker, Inc.

            1 of 1 people found this helpful
            • 3. Re: FileMaker WebDirect excessive polling
              greglane

              Thanks TSGal. We’ll add a step to our maintenance routine until the issue is fixed or until FileMaker Cloud begins purging old logs. We used the following command to delete about 17 GB of ssl log files that were over 30 days old:

               

              sudo find /FileMakerData/Logs/HTTPServer/ssl* -mtime +30 -exec rm {} \;

               

              2 of 2 people found this helpful