That's expected and unfortunately there is no easy way to change the behaviour.
See: https://issues.apache.org/jira/browse/HADOOP-1202 They are binding to the internal subnet because the hostname resolves to the internal IP address (inside the Amazon cloud) + because of NAT the only NIC available on the VMs is the one connected to the internal network. Cheers, -- Andrei Savu / axemblr.com / Tools for Clouds On Tue, May 1, 2012 at 8:31 AM, Alex Heneveld < [email protected]> wrote: > hi, > > i'm trying to access the namenode externally, in a simple hadoop cluster > essentially built with: > > whirr.instance-templates=1 hadoop-jobtracker+hadoop-**namenode,1 > hadoop-datanode+hadoop-**tasktracker > whirr.provider=aws-ec2 > > the web ui's are correctly on 0.0.0.0 but the namenode and jobtracker seem > to be bound to the internal subnet only. > (see netstat below.) > > is there an easy way to s/10.192.135.52/0.0.0.0/ ? > > thanks, > alex > > > alex@domU-12-31-39-0F-84-C6:~$ netstat -an > Active Internet connections (servers and established) > Proto Recv-Q Send-Q Local Address Foreign Address State > tcp 0 0 0.0.0.0:50030 0.0.0.0:* LISTEN > tcp 0 0 10.193.135.52:8020 0.0.0.0:* LISTEN > tcp 0 0 10.193.135.52:8021 0.0.0.0:* LISTEN > tcp 0 0 0.0.0.0:50070 0.0.0.0:* LISTEN > tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN > tcp 0 0 10.193.135.52:8020 10.214.11.37:37576 > ESTABLISHED > tcp 0 0 10.193.135.52:8021 10.214.11.37:44212 > ESTABLISHED > tcp 0 0 10.193.135.52:22 208.51.138.10:58791 > ESTABLISHED > tcp6 0 0 :::22 :::* LISTEN > udp 0 0 0.0.0.0:68 0.0.0.0:* > > >
