Hi Keith: On Fri, Jul 20, 2012 at 3:28 PM, Keith Wiley <kwi...@keithwiley.com> wrote:
> I'm curious about the relationship between the namenode/job/task trackers > and the machine's web server? Do the former require the latter? The hadoop daemons embed a jetty instance to serve their user interfaces over http. You will see the jetty instance come online in your logs with lines like this: 2012-07-20 22:55:14,519 INFO org.apache.hadoop.http.HttpServer: Port returned by webServer.getConnectors()[0].getLocalPort() before open() is -1. Opening the listener on 50060 2012-07-20 22:55:14,519 INFO org.apache.hadoop.http.HttpServer: listener.getLocalPort() returned 50060 webServer.getConnectors()[0].getLocalPort() before open() is -1. Opening the listener on 50060 2012-07-20 22:55:14,519 INFO org.apache.hadoop.http.HttpServer: listener.getLocalPort() returned 50060 webServer.getConnectors()[0].getLocalPort() returned 50060 2012-07-20 22:55:14,519 INFO org.apache.hadoop.http.HttpServer: Jetty bound to port 50060 Does successful connection to the trackers imply that the machine has a web > server up and running? A tasktracker that is "up" should have it's web interface served by it's embedded jetty instance. > I realize the ports are totally different (web is generally port 80), but > the trackers are headed with the "http" URI, so I'm a little unsure if that > means anything w.r.t. my question. Can you run and access the trackers on > a machine that doesn't have a webserver installed or running? > > The hadoop daemons do not require a web server to be running on any grid nodes. > I know, it's a weird question. Thanks for any quick response. > > HTH Regards, Al