[SPARK-4837] NettyBlockTransferService should use spark.blockManager.port config

[aarondav]

This is used in NioBlockTransferService here:
https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/network/nio/NioBlockTransferService.scala#L66

[SparkQA]

Test build #24430 has started for PR 3688 at commit ebd2007.

• This patch merges cleanly.

[SparkQA]

Test build #24430 has finished for PR 3688 at commit ebd2007.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24430/
Test FAILed.

[aarondav]

Jenkins, retest this please.

[SparkQA]

Test build #24441 has started for PR 3688 at commit ebd2007.

• This patch merges cleanly.

[SparkQA]

Test build #24441 has finished for PR 3688 at commit ebd2007.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24441/
Test FAILed.

[rxin]

Jenkins, retest this please.

[SparkQA]

Test build #24503 has started for PR 3688 at commit ebd2007.

• This patch merges cleanly.

[SparkQA]

Test build #24503 has finished for PR 3688 at commit ebd2007.

• 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/24503/
Test PASSed.

[rxin]

LGTM.

[JoshRosen]

LGTM, too, so I'm going to merge this into master (1.3.0) and branch-1.2 (1.2.1). Thanks!

[dizzy2]

this fix doesn't work when both executor and driver are running on the same machine (as they are trying to acquire the same port)

[aarondav]

@dizzy2 Sorry, thought I responded earlier -- you are quite right. It should be a straightforward fix of adding a Utils.startServiceOnPort call to do the same incrementing we do elsewhere.