Checking the Communications Process

When you take the data capture, you should also check that the communications process (MFCS) is still running. If it is, you can then check what it is currently doing. To do this:

  1. Record the hostname of the machine where the enterprise server instance is running.
  2. Go to the Enterprise Server Administration Web interface. In the list of servers on the Home page, find the enterprise server instance that you want to check. The Communications Processes column lists one or more communications processes for that server. Record each of the port numbers shown in the column. The port number is the last part of the address, after the second colon.
  3. In a Web browser, go to the following URL:
    http://hostname:port/MF_KEEPALIVE
    

    where hostname is the hostname (or IP address) from step 1 and port is the first port number from step 2.

    If MFCS is working correctly, you will see a response similar to the following:

    Server=ESDEMO
    ServerType=GKC
    Version=1.2.5
    Status=Started
  4. Repeat step 3 using "MF_STATISTICS" instead of "MF_KEEPALIVE". You then see a response showing what MFCS is currently doing, how many conversations it has processed, and other information.

If you are running multiple communications processes, you can check each one in turn using this method. Each of the port numbers you recorded in step 2 corresponds to a communications process.

Note: Some of your communications processes for a given server might be disabled; they do not need to be checked.

You can also append /file to URL for the MF_STATISTICS request (for example, http://hostname:port/MF_STATISTICS/file). This will cause MFCS to append the statistics report to a file named statistics.txt in the system directory, the file is created if it does not already exist. The system directory is the same location as the MFCS log files log*.html. MFCS does this before attempting to send the HTTP response, so it might work even in circumstances where MFCS is unable to send the response to the browser. This can also be used to periodically collect information about the MFCS workload.

For UNIX, you can also direct MFCS to update the statistics.txt file by sending it the SIGUSR1 signal, for example using the command kill -USR1 MFCS-pid.