Hi Sami,
I quess you refer to these:
• LocalJobRunner:
• Run as kind of singelton
• Have a kind of jobQueue
• Implement JobSubmissionProtocol status-report
methods
• implement killJob method
Right!
-how about writing a nutchrunner that just extends the
functionality of localjobrunner?
That would be one solution, however I still hope that the hadoop
developer understand that it would be general benefit to improve the
local jobrunner.
Since it would be somehow duplicated code it does not feel right, but
I also think better this way as never get this issue solved.
-scheduling (jobQueue) could be completely outside of jobrunner?
We solved that with Quarz and file based JobStore we implemented back
than.
Stefan
