search

aodn-archive / DELETE_ME_go-go-duck

NetCDF aggregation service
1 stars 0 forks source link

issue 1728 reason #105

Closed danfruehauf closed 9 years ago

danfruehauf commented 9 years ago

Fixes https://github.com/aodn/aodn-portal/issues/1728 by providing a reason to the user as to why the aggregation failed.

The crux was to find out when a timeout occurred I think. Since waitForOrKill could not tell if a timeout occurred or it's just another failure. So now it's measuring the time it took for the job to run. If it actually exceeded the expectation it shows a timeout error.

gogoduck.sh now also returns a value of 3 when the file limit is reached. We can display that to the user in the nice error log, rather than let him figure it out from the report.

I hope that's the last GGD tweak we do before we rewrite in Java.

kereid commented 9 years ago

I'd say something along the lines of TIMEOUT_EXPIRED= Maximum processing time limit exceeded, TOO_MANY_FILES = Maximum file limit exceeded

Did you want a suggestion for none as well? What is the context of a "none" job reason?

danfruehauf commented 9 years ago

@kereid I think that's good. NONE is... when there are no problems :)

The 2 others are for when the actual aggregation fails in an unexpected manner. UNKNOWN, well :)

kereid commented 9 years ago

Oh right :) for unknown would it be worth having UNKNOWN = Unexpected error; please report to info@emii.org.au

danfruehauf commented 9 years ago

@kereid No worries. I'll fix that too.

jonescc commented 9 years ago

I'll review

danfruehauf commented 9 years ago

@joness Ready again.