|
|
+
Oleg Zhurakousky 2012-10-04, 19:58
-
Re: Submitting a job to a remote clusterHemanth Yamijala 2012-10-05, 04:08
Hi,
Could you please share your setup details - i.e. how many slaves, how many datanodes and tasktrackers. Also, the configuration - in particular hdfs-site.xml ? To answer your question: the datanode address is picked up from hdfs-site.xml, or hdfs-default.xml from the property dfs.datanode.address. This is generally left as default value, unless you want to change the port number and things will work fine. Thanks Hemanth On Fri, Oct 5, 2012 at 1:28 AM, Oleg Zhurakousky <[EMAIL PROTECTED] > wrote: > Trying to submit a Job to a remote Hadoop instance. Everything seem to > start fine but then I am seeing this: > > 2012-10-04 15:56:32,617 INFO [org.apache.hadoop.mapred.JobClient] - < map > 0% reduce 0%> > > 2012-10-04 15:56:32,621 INFO [org.apache.hadoop.mapred.MapTask] - <data > buffer = 79691776/99614720> > > 2012-10-04 15:56:32,621 INFO [org.apache.hadoop.mapred.MapTask] - <record > buffer = 262144/327680> > > 2012-10-04 15:56:32,641 WARN [org.apache.hadoop.hdfs.DFSClient] - <Failed > to connect to /127.0.0.1:50010, add to deadNodes and > continuejava.net.ConnectException: Connection refused> > > > How can I specify datanode address to use? > > Thanks > > Oleg > |