Hi Folks. I don't think I paid enough attention to YARN vs. MR1 when I
built Giraph 1.0.0 for our system. How much better is Giraph on YARN?
Thank you.
--
Matthew Cornell | m...@matthewcornell.org
Theoretically, Giraph on YARN would be much better (actual resource
request rather than mapper hack). That being said, Eli is the best
person to talk about that. We haven't tried YARN.
Avery
On 10/6/14, 8:51 AM, Matthew Cornell wrote:
Hi Folks. I don't think I paid enough attention to YARN
Dear Mathew,
Pardon the late reply.
In my work, every vertex will have a score, I want to find the top-K scoring
vertexes.I planned to use an aggregator for this, here, I store my aggregated
value as Text, which stores the top-K vertex-id and their scores.
By the way, I have solved my problem in
Thanks Charith, but my main question still remains, even with the examples that
comes with Giraph, such as simple shortest path computation example, the
System.out.println or the Log.Debug (or I also tried Log.Info) they all do not
print out customer messages that I write in the compute method.
Hi Tamer,
if you run Giraph on YARN you can use the log aggregation feature. If you
try to write to HDFS you should consider the HDFS API, but many mappers
would have to write into individual files. Why not writing all logs via
Log4j into Flume and from here to HDFS?
There is a Log4J appender