LSF Troubleshooting

The following are general troubleshooting steps:

1. Ensure the LSF lsrun command is enabled.
2. Look for user errors.

   For example:

   • Are the executable path and project path correct and complete?
   • Are there sufficient resources (CPU/Memory/Disk) allocated to the job?
   • Is the project available on the execution host?
   • Does the job submitter have read/writer permissions on the project directory and read/execute permissions on the installation directory?
   • Is the project locked?
3. Determine whether this is a standalone product issue.
   • Run Electronics Desktop on the machine outside of the scheduler and see if it opens and analyzes.
4. Examine outputs and logs.
   • Output of the LSF batch job. Obtain this using LSF commands: "bacct -l <jobid>"
   • Batch log (typically <projectname>.log, located in the project directory.
5. Enable additional debug logs using the steps below.

   In the job submission window, set the following environment variables:

   • ANSOFT_DEBUG_MODE = 1
   • ANSOFT_DEBUG_LOG = <path to directory accessible by all machines in the cluster>
   • ANSOFT_DEBUG_LOG_SEPARATE = 1
   • ANSOFT_LSF_LOG = <path to a specific .log file in the directory set under ANSOFT_DEBUG_LOG>
6. For each pair of machines between which remote analysis fails, run ping remote-machine and note the output.
7. For each machine in the network, dump network interfaces (for example, run ifconfig -a) and note the output.