Hey, I finally solved it! It was to do with my Cassandra cluster. My hadoop and cassandra clusters were in two different datacenters. This caused cassandra requests to timeout. And that meant the generate phase didn’t have any input!
Works like a charm now :) Regards -- Manikandan Saravanan Architect - Technology TheSocialPeople On 6 June 2014 at 10:19:45 pm, Lewis John Mcgibbney ([email protected]) wrote: Hi, On Thu, Jun 5, 2014 at 11:19 PM, <[email protected]> wrote: > > # skip URLs containing certain characters as probable queries, etc. > #-[?*!@=] > > Is this your preference? Anyways, you need to try and debug why there ends up being no Map Input records for the Generate phase. I suggest you run the ant eclipse target and debug a generate step on one URL on a fresh inject on a fresh Keyspace (or truncated on atleast). Lewis
