I had been getting this exact same problem on my cluster. The solution I
*finally* came up with (after 1-2 weeks of messing) was to manually
insert a unique datanode ID into the datadir/current/VERSION file. If
your problem is the same as mine your VERSION file may look like:
#Fri Apr 04 00:43
Hi,
I am configuring a cluster using 0.16.2. But when the cluster starts, it can
not find any live datanode. I check the logs of one datanodes, which seems
hanging in startup:
2008-04-05 17:48:17,967 INFO org.apache.hadoop.dfs.DataNode: Registered
FSDatasetStatusMBean
2008-04-05 17:48:17,968 IN
