Hi,
I have a strange timeout issue with a scenario where i'm running multiple copies of the job executor.
The job itself runs a mapreduce, then runs a trans to update a log table.
The problem is that the DB connection gets timed out.
To me this makes no sense, becuase the update of the log table does not START until AFTER the mapreduce finishes. So it should connect, and update immediately.
However; Is the multiple copies thing getting in the way? Is it somehow sharing that database connection between all those copies? I would have assumed that each copy would have its own connection. I'm 99% sure this is how it is, but just want to rule this out!
Thanks,
Dan