hi.
i have a small hadoop grid connected with a 1g network.
when servers are configured to use the local dns server the jobs are
running without a problem and copy speed during reduce is tens on MB.
once i change the servers to work with a cache only named server on each
node, i start to get fa
Looks like your caching DNS servers aren't really functioning as you'd
expect them to?
> org.apache.hadoop.hbase.ZooKeeperConnectionException:
> java.net.UnknownHostException: s06.xxx.local
(That .local also worries me, you probably have a misconfiguration in
resolution somewhere.)
On Wed, Jan 4
so it seems but doing a dig from terminal command line returns the
results correctly.
the same setting are running in production servers (not hadoop) for
months without problems.
clarification - i changed servers names in logs, domain isn't xxx.local
originally..
On 01/04/2012 05:19 PM, Har
Hi,
Please ping the host you want to reach and check your hosts-file and your
resolve.conf
- Alex
Alexander Lorenz
http://mapredit.blogspot.com
On Jan 4, 2012, at 7:28 AM, Oren wrote:
> so it seems but doing a dig from terminal command line returns the results
> correctly.
> the same settin
I'm not sure if java is using the system's libc resolver, but assuming it is,
you cannot use utilities like nslookup or dig because their use their own
resolver. Ping usually uses the libc resolver. If you are on linux, you can
use "getent hosts $hostname" to definitively test the libc resolve
