Getting help

The first port of call for help on Apocrita should be this site especially the section on Using the cluster.

For standard linux commands and SGE commands, the Linux manual pages provide full information on available options and how to use the tools.

If you can't solve your issue using the available documentation you should contact ITS Research Support who will be able to assist you with your issue; emails to this address will automatically create tickets.

Recommendations when requesting help

If you are having trouble running a particular job, please check the job detail on the stats pages to check if you are exceeding resources, or any error messages returned. Typical issues are non-existent files and folders used for input and output, and failure to load the module before using the application.

When raising a support ticket with us, if your query regards jobs on the cluster, please supply all relevant info such as:

  • jobid
  • the location/path of your submission script
  • your username on the cluster
  • any error messages or unusual behaviour noted.

This information will allow us to start looking into your issue immediately, without having to ask further questions. If you subsequently get the job working, please notify us as soon as you can, so that we do not spend time looking into a job script that has subsequently been fixed.

We also recommend using version control for your code and scripts. The college Github Enterprise instance is ideal for this. This will help you track changes and revert non-working code easily.

Recording Sessions

If you are experiencing issues with certain commands you may be asked to record your session, this can be done with the record_session command.

Here's an example of using the record_session command:

Once you have recorded your session notify ITS Research Support and they will continue investigation of your issue.

This enables us to see the exact errors / issues you are experiencing and the precise steps to reproduce them.