[SPARK-12170] [SparkR] Deprecate the JAVA-specific deserialized storage levels

[yanboliang]

Since the data is always serialized on the R side, the storage levels of JAVA-specific deserialization should be deprecated. We should automatically convert MEMORY_ONLY to MEMORY_ONLY_SER and deprecated MEMORY_ONLY_SER at SparkR side. Please refer the similar change for PySpark at #10092.

[sun-rui]

@yanboliang, thanks for the PR. But for SparkR RDD API, still discussion how to evolve it. So maybe we need more discussion before submitting a PR for all RDD related JIRA issues (Those JIRAs are for tracking purpose for now)

@shivaram, @felixcheung.

[SparkQA]

Test build #47312 has finished for PR 10189 at commit fa88169.

• This patch fails R style tests.
• This patch merges cleanly.
• This patch adds no public classes.

[yanboliang]

@sun-rui Got it, thanks for your kindly remind.

[felixcheung]

yes, in fact we are instructed not to work on Spark 2.0 release yet

[SparkQA]

Test build #47314 has finished for PR 10189 at commit 588acaf.

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

[shivaram]

@yanboliang Could you close this PR for now ? It just makes the reviewing queue easier

[yanboliang]

OK, I will close this PR.