SPARK-4447. Remove layers of abstraction in YARN code no longer needed after dropping yarn-alpha

[sryza]

No description provided.

[vanzin]

nit: should go after java and scala imports.

[SparkQA]

Test build #24273 has started for PR 3652 at commit 17628f8.

• This patch merges cleanly.

[vanzin]

nit: should be right after java imports.

[vanzin]

I assume most here is just code motion, so I didn't pay that close attention to the changes in the container allocation code. If that's true, then this LGTM.

[SparkQA]

Test build #24273 has finished for PR 3652 at commit 17628f8.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24273/
Test PASSed.

[andrewor14]

delete

[andrewor14]

This looks fine on a first glance. I plan to take a detailed pass in the next day or two to make sure we didn't change any logic.

[andrewor14]

Just took a deeper look comparing the old code with the new code and couldn't find any glaring problems. I'll test and merge this once you address the comments.

[sryza]

Cool, updated patch reflects Marcelo and Andrew's comments.

[SparkQA]

Test build #24676 has started for PR 3652 at commit 2791158.

• This patch merges cleanly.

[SparkQA]

Test build #24676 has finished for PR 3652 at commit 2791158.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24676/
Test PASSed.

[andrewor14]

Cool I just tested this on a Hadoop 2.4 cluster and it worked like a charm. I'm merging this into master thanks @sryza @vanzin.