Investigate the way the transaction happens from a job being submitted to the
job executing: there should be a way of making the transaction atomic
atomic: able to tell user with certainty that the job was submitted from
execution, in the case of disruption.
without the need to save any intermediary file/output to disk
metrics/performance: minimize time needed for such transaction to be executed
Original issue reported on code.google.com by geo...@marsel.is on 25 May 2013 at 11:01
Original issue reported on code.google.com by
geo...@marsel.ison 25 May 2013 at 11:01