Reputation: 1294
I know hadoop 2.6 cluster can be configured to run 'yarn' or 'local', where 'yarn' is mapreduce v2 and 'local' is just local mode. And I learnt from this thread (What is the difference between classic, local for mapreduce.framework.name in mapred-site.xml?) that it can also be configured to run in 'classic' framework, which is mapreduce v1. But I cannot run any job if I simply change 'mapreduce.framework.name' from 'yarn' (or 'local') to 'classic'. So, is it possible to do that? How can I configure it?
My another thought is, I'm using apache hadoop 2.6 distro, does that come with mapreduce v1 framework? If not, I should not be able to configure the cluster to run v1 framework.
Note, my question is not to run a mapreduce v1 job on hadoop 2.6.0, but configure the cluster in some way, (not 'yarn', not 'local'), running mapreduce v1 framework when got some job.
Upvotes: 0
Views: 314
Reputation: 934
You can use hadoop 2.6.0 to configure MR1 or MR2. I have not used Apache, but i use CDH distributon of it where i configured my cluster as MRV1.
Upvotes: 1