Also, you may want to keep an eye on the TEZ-690 Ease of Use jira. Its got some breaking changes. And creates Abstract classes for Processors etc. Removes confusing helper methods like doJobClientMagic :) Reduces DAG creation verbosity.
Bikas -----Original Message----- From: Bikas Saha [mailto:[email protected]] Sent: Sunday, May 18, 2014 10:05 AM To: [email protected] Subject: RE: spill files Chris, MRInputLegacy is specifically written for MR legacy and is not something we recommend to use. It has a bunch of workarounds for MR idiosyncroacies. MRInput is what we recommend to use. Please switch to that. Bikas -----Original Message----- From: Chris K Wensel [mailto:[email protected]] Sent: Saturday, May 17, 2014 4:46 PM To: [email protected] Subject: Re: spill files > Where did you need to invoke #init() ? That shouldn't be the case, except maybe for specific inputs. > The processor should not need to invoke initialize or close on Inputs > / Outputs. "start()" is something that was added to allow Processors to skip Inputs if required. MRInputLegacy#init must be called before I can call #getReader() which must be called before I can get #getOldRecordReader(). At least that's how I got that chain to work. fwiw, by not calling #close() I have things working now for a simple grouping case. will start on co-grouping/joins next. ckw -- Chris K Wensel [email protected] http://concurrentinc.com -- CONFIDENTIALITY NOTICE NOTICE: This message is intended for the use of the individual or entity to which it is addressed and may contain information that is confidential, privileged and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient, you are hereby notified that any printing, copying, dissemination, distribution, disclosure or forwarding of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and delete it from your system. Thank You.
