Production System Failure Data Capture

For all troubleshooting exercises, it is essential you gather as much information as possible about the state of the enterprise server region when the problem occurred and about the events leading up to the problem. This should include the date/time the problem occurred, and observations on what was happening in the system, how long it had been running, what the symptoms were, and how/where these were observed.

When a problem occurs, you must capture and provide specific logs, traces and dumps, and ideally the contents of various configuration files, directories (for example, MFDS), and the output from a number of operating system tools. This should be done as soon as possible after the failure occurs.

Diagnostic Collection Script

Use the MFESdiags diagnostic script to collect this information automatically.

For Windows environments : The script is MFESdiags.cmd which can be obtained from Customer Care. You can run it from any command prompt or from Windows Explorer. You will be prompted for the region name, and you should zip up the resultant data collection directory and attach it to an incident.

For UNIX environments : The script is mfesdiags.sh which can be obtained from Customer Care. Transfer it to the UNIX machine and give it the 'execute' permission (for example, chmod +x mfesdiags.sh). Run the script in the COBOL/Enterprise Server environment. It will prompt for the relevant region name for which to collect diagnostics, and will generate mfesdiags.trz which you should attach to an incident.

Both the UNIX and Windows scripts invoke the relevant 'mfsupport' utility which collects information about the operating system and machine hardware, along with details relating to Micro Focus products you have installed.

As a minimum, for example if MFESdiags is not used to collect the data, you should collect and provide the following items as soon as possible after a failure. You can do this simply by zipping or 'tar'ing the contents of the system/region directory.

  • console.log
  • log.html or log-*.html (for communications problems)
  • journal.log (for MFDS/security problems)
  • Any casdump or aux traces
  • MFSupportInfo (on Windows)
  • mfsupport (on Unix)

The MFESdiags data collection script collects all of the above, and other useful information details, as follows:

  • All files from the Enterprise Server's System/Region Directory, including:
    • console.log and console.bak. The communications process log log.html or log-*.html
    • Any trace diagnostic data sets, casauxta.rec and casauxtb.rec
    • Any system abend dumps, casdumpa.rec, casdumpb.rec or casdumpx.rec
    • Any HSF output files (*.csv)
  • The output from MFSupportInfo (Windows) or mfsupport (UNIX), which contains product and system information
  • The Directory Server (MFDS) log file
  • The Directory Server configuration directory
  • Additional configuration files (if they exist) including:
    • mf-server.dat
    • mf-client.dat
    • CTF Trace configuration file
    • COBCONFIG (core_on_error) configuration file
    • EXTFH file handler configuration file
  • A list of system processes running on the machine
  • A list of network connections and listening sockets using netstat
  • Windows Service information for MFDS (Windows)
  • Windows System and Application Event logs (Windows)
  • A list of shared memory areas and semaphores (UNIX)
  • A list of open files (UNIX)
Note: The MFESdiags script will collect CTF trace logs if the region is configured to generate CTF trace. See the CTF Trace topic for configuration details.