|
|
-
MapReduce vs hosts (Cannot resolve the host name)Jean-Marc Spaggiari 2012-10-11, 13:17
Hi,
I'm facing a small issue, most probably configuration related, that I'm not able to solve. I'm trying to run the rowcounter. Here is the command line: export HADOOP_HOME=/home/hadoop/hadoop-1.0.3/; export HBASE_HOME=/home/hbase/hbase-0.94.0/; HADOOP_CLASSPATH=`${HBASE_HOME}/bin/hbase classpath` ${HADOOP_HOME}/bin/hadoop jar ${HBASE_HOME}/hbase-0.94.0.jar rowcounter work_proposed -Dhbase.client.scanner.cachi=10 I'm getting errors for each of my regions: 12/10/11 09:10:45 ERROR mapreduce.TableInputFormatBase: Cannot resolve the host name for /192.168.23.9 because of javax.naming.NameNotFoundException: DNS name not found [response code 3]; remaining name '9.23.168.192.in-addr.arpa' I have 25 regions for this table, so I have 25 lines like that. And then it's blocking at: 12/10/11 09:10:45 INFO mapred.JobClient: Running job: job_201209151131_0005 12/10/11 09:10:46 INFO mapred.JobClient: map 0% reduce 0% Here is an extract of my host file: 192.168.23.9 node5 192.168.23.8 node4 192.168.23.7 node3 192.168.23.6 node2 192.168.23.5 node1 Hadoop is working fine. HBase is working fine. If I try the shell and do a count, it's working fine. If I try with bin/hbase org.apache.hadoop.hbase.mapreduce.RowCounter work_proposed it's working fine, but it's doing the regions one by one, and not in parallel. So it's very slow. I can ping the nodes, by name or IPs, I can ssh them too. Any idea where I can start to look at? Thanks, JM |