[GOBBLIN-386] Optimize the memory usage of schema fetch in Hive Registration

[autumnust]

This PR is addressing the problem when hive registration is blowing up the memory since the olde schema-fetching process requires (1)Getting all .avro files within a dataset folder (2). Sort all these files by their modification time.

The solution is to conduct DFS while comparing the timestamp on-the-fly.

JIRA

• [x] My PR addresses the following Gobblin JIRA issues and references them in the PR title. For example, "[GOBBLIN-386] My Gobblin PR"
  • https://issues.apache.org/jira/browse/GOBBLIN-386

Description

• [x] Here are some details about my PR, including screenshots (if applicable):

Tests

• [x] My PR adds the following unit tests OR does not need testing for this extremely good reason:

Commits

• [x] My commits all reference JIRA issues in their subject lines, and I have squashed multiple commits if they address the same issue. In addition, my commits follow the guidelines from "How to write a good git commit message":
  1. Subject is separated from body by a blank line
  2. Subject is limited to 50 characters
  3. Subject does not end with a period
  4. Subject uses the imperative mood ("add", not "adding")
  5. Body wraps at 72 characters
  6. Body explains "what" and "why", not "how"

[autumnust]

@ibuenros @erwa Can you help review ? Thanks.

[erwa]

Looks good to me.