How to expose the task tracker/Job tracker webinterface to the public in hadoop?

I'm trying to monitor different cluster nodes, but everytime I have to ssh -X to the node and start the browser to take a look at the status information.

Is there anyway to make theses http://jobtracker:50030, http://namenode:50070 etc, open to public access ? Anyone can look at them in the browser.

I guess this probably need to be done by setup a server开发者_如何转开发. But I haven't find any useful information by googleing it.

Why don't you run a proxy like NginX on the master?

Nginx would run on port 80, and allow you to forward one domain to port 50030, and the other to 50070.

It's used in production web-application deployments by thousands of companies to do exactly that:

http://nginx.org/