Open lupeterm opened 1 year ago
Good thing to remove obr find references, however, the idea of obr find extended obr query. Currently, obr query allows to query key value pairs in job docs and so to find the log file of the failed process (the exact query might be difficult though). Obr find was meant to display the content of said log file.
Maybe I could change the exception to be more specific, for example
'['blockMesh']' returned non-zero exit status 1 for job with id XY. check: 'obr query -q "{key:'job_id', value:'XY} --log blockMesh" for more info
Thus, i would also add a --log
--get-log <operation>
flag that would additionally print the most recent log file of the specified operation.
That is a great idea, but I guess --log needs a pager function like git log to handle large log files. Also, we would need to clarify how to handle many exceptions.
For instance:
Probably just replace this with the equivalent call to
obr query