Thanks for the help HarshJ.
You are correct in that the master is not seeing any of the TTs. The
webui shows 0 active nodes. I ran start-all and assumed it started
everything. It did, but for some reason they are not communicating.
Running start-all again shows:
fish12.cs.uwyo.edu: tasktracker running as process 1979. Stop it first.
fish21.cs.uwyo.edu: tasktracker running as process 938. Stop it first.
fish13.cs.uwyo.edu: tasktracker running as process 1859. Stop it first.
fish22.cs.uwyo.edu: tasktracker running as process 24126. Stop it first.
fish14.cs.uwyo.edu: tasktracker running as process 1905. Stop it first.
fish16.cs.uwyo.edu: tasktracker running as process 32282. Stop it first.
fish19.cs.uwyo.edu: tasktracker running as process 32489. Stop it first.
fish17.cs.uwyo.edu: tasktracker running as process 5261. Stop it first.
fish15.cs.uwyo.edu: tasktracker running as process 1595. Stop it first.
fish18.cs.uwyo.edu: tasktracker running as process 1653. Stop it first.
fish20.cs.uwyo.edu: tasktracker running as process 23028. Stop it first.
So the processes are running correctly on all the nodes, right? I just
have a misconfiguration that is preventing them from talking to each
other? What should mapred-site.xml look like for a typical cluster?
Thanks,
Travis
On 5/3/2011 11:28 AM, Harsh J wrote:
Hello Travis,
On Tue, May 3, 2011 at 10:53 PM, Travis Bolinger<tboli...@gmail.com> wrote:
11/05/03 10:44:39 INFO mapred.JobClient: Running job: job_201105031040_0001
11/05/03 10:44:40 INFO mapred.JobClient: map 0% reduce 0%
It just sits at 0% forever (well ~30 minutes before I killed it). When I
tested in pseudo-distributed mode it was nearly instant, so I'm assuming I
have a setting wrong.
Are all your TTs up and functional (reported live in JT's web UI)?
