ecl_run.py does not always find error messages

[eivindjahren]

ecl_run.py will sometimes fail with the following stderr message:

job ECLIPSE100 failed with: 'Process exited with status code 1'
stderr file: '/.../realization-15/pred/ECLIPSE100.stderr.129',
its contents:
Traceback (most recent call last):
File "/.../ert/resources/forward-models/res/script/ecl100.py", line 9, in <module>
run(config, [arg for arg in sys.argv[1:] if len(arg) > 0])
File "/.../ert/resources/forward-models/res/script/ecl_run.py", line 537, in run
run.runEclipse(eclrun_config=eclrun_config)
File "/.../ert/resources/forward-models/res/script/ecl_run.py", line 412, in runEclipse
raise err from None
File "/../ert/resources/forward-models/res/script/ecl_run.py", line 390, in runEclipse
self.assertECLEND()
File "/.../ert/resources/forward-models/res/script/ecl_run.py", line 459, in assertECLEND
raise RuntimeError(
RuntimeError: Eclipse simulation failed with:1 errors:

But no errors are found with self.parseErrors(). Likely the parsing is not correct, but unsure in which way

[berland]

This requires access to the PRT files of the failed Eclipse runs.

If there is usually a match between result.errors and len(error_list) we can log the PRT filename when it mismatches, and then we need to get access to the file afterwards.

[berland]

Or possibly, whenever error_list is empty, we could log the last say 2000 bytes of the PRT file.

[eivindjahren]

I think both approaches are valid @berland. We could when len(error_list) < result.errors then we print the name of the PRT file and the last characters of it.