I'm guessing that was just a slip with ROUND_ROBIN_BACKOFF => RANDOM.
Rahul: Both RANDOM and ROUND_ROBIN now have backoff semantics in them.
So RANDOM_BACKOFF became RANDOM and ROUND_ROBIN_BACKOFF became
ROUND_ROBIN. It as processors fail, they will be temporarily blacklisted
and the random/round_robin semantics will be carried out on the
remaining sinks as one would expect.
On 11/17/2012 05:56 PM, Alexander Alten-Lorenz wrote:
Hi,
Yes, thats the correct configuration. We had such an parameter in Flume, but
removed them later as we clean up the code. RANDOM is working like
ROUND_ROBIN_BACKOFF.
SideNote:
Please use our wiki for the newest guides:
https://cwiki.apache.org/confluence/display/FLUME/Index
and the actual user guide:
http://flume.apache.org/FlumeUserGuide.html
cheers,
Alex
On Nov 16, 2012, at 10:45 PM, Rahul Ravindran <[email protected]> wrote:
Hi ,
The documentation at
http://archive.cloudera.com/cdh/3/flume-ng/FlumeUserGuide.html indicates that
there is a Round_robin_backoff but this threw an error. It looks like there is
a constant with this type defined in code, but it is not used anywhere. The
code seemed to indicate that the below should achieve a backoff for Round Robin
though there is not documentation about the processor.backoff parameter. Is the
below the right way to perform a round robin with backoff?
Thanks,
~Rahul.
agent1.sinkgroups = group1
agent1.sinkgroups.group1.sinks = avroSink1 avroSink2
agent1.sinkgroups.group1.processor.type = load_balance
agent1.sinkgroups.group1.processor.selector = ROUND_ROBIN
agent1.sinkgroups.group1.processor.backoff = true
--
Alexander Alten-Lorenz
http://mapredit.blogspot.com
German Hadoop LinkedIn Group: http://goo.gl/N8pCF
