Yes : great start Dave , here are two things you should update: 1) please specify the in the title that this is deploying bigtop 0.7.0... on the custom slides that this change is only necessary for deploying 0.7.0 rpms. The puppet recipes on head target for 0'8.0, but you are using 0.7.0 rpms, which require "mapreduce.shuffle" hence the need for that mapreduce_shuffle parameter change .
2) Can you also include debugging info about how to cat the yarn binary logs for when containers fail? > On Aug 7, 2014, at 10:27 AM, Sean Mackrory <[email protected]> wrote: > > Great to see this being done - the project could really use nicer > documentation for users. Regarding the mapreduce_shuffle -> mapreduce.shuffle > issue - have you created or seen a JIRA for that? Haven't run into it myself > but we should get that fixed... > > One nitpick: when invoking scripts in /etc/init.d, it's usually recommended > to invoke them using the 'service' command. It cleans out the environment to > make sure the scripts are invoked in a nice predictable way. It also looks a > little more polished :) > > e.g > > /etc/init.d/hadoop-hdfs-namenode init -> service hadoop-hdfs-namenode init > > >> On Thu, Aug 7, 2014 at 7:23 AM, David Fryer <[email protected]> wrote: >> Hi Bigtop! >> I think I'm going to do a web presentation on setting up a bare-metal hadoop >> cluster. The presentation slides that I will be using can be found here: >> https://docs.google.com/presentation/d/1cEkq2x-Q_n7nbFCttHIn3luoOkz3UMlEEtbudea9l1Y/edit?usp=sharing >> If anyone has any edits or comments, I'd be happy to hear your feedback. >> Thank You, >> David Fryer >
