What to do if you have an error

Sometimes, the diagnostics may crash and you will not know why. This section will give you a procedure to follow before reporting the issue. This procedure is intended to solve some common problems or, at least, to help you in creating good issue reports. Remember: a good issue report reduces the time required to solve it!

Hint

Please, read carefully the error message. Most times the error message will point you to the problem’s source and sometimes even give you a hint of how to solve it by yourself. And if this it not the case or if you find it obscure, even if it was helpful, please contact the developers so it can be improved in further versions

Try this simple steps BEFORE reporting an issue

  • Clean scratch folder
  • Update to the latest compatible tag: maybe your issue is already solved in it
  • If you get the error for the first chunk of a given diagnostic, change the number of chunks to 1
  • Call the diags with the -lc DEBUG -log log.txt options

Now, you have two options: if everything is fine, the error was probably due to some corrupted files or some unstable machine state. Nevertheless, try running the diagnostic with -lc DEBUG -log log.txt for all the chunks. If everything it’s fine that’s all.

If you experienced the same problem again, go to the GitLab portal and look into the open issues ( https://earth.bsc.es/gitlab/es/ocean_diagnostics/issues ). If you find your issue or a very similar one, use it to report your problems. If you can not find an open one that suites your problem, create a new one and explain what is happening to you. In any case, it will be very useful if you can attach your diags.conf and log.txt files and specify the machine you were using.

After that, it’s just a matter of waiting for the developers to do their work and answering the questions that they may have. Please, be patient.

Caution

Of course, there is a third option: you keep experiencing an error that appears randomly on some executions but you are not able to reproduce it in a consistent manner. Report it and attach as much logs and configuration files as you have, along with the date and time of the errors.