Sounds like a problem with the hostnames of your datanodes.
Check that your are able to ping all the datanodes with the hostnames
they had send to the namenode.
check:
bin/nutch ndfs -report to see the hostnames.
Stefan
Am 24.11.2005 um 16:04 schrieb Anton Potehin:
When we start namenode
Hi,
with the latest mr branch I have a strange problem copy a file from
local to the ndfs.
Has anyone manage to use the ndfs client with this branch until last
days?
The problems occurs until connecting to a datanode to get the next
free block:
cluster2:~/mr_nutch-0.8-dev myuser$