Add caching file system to hive connector

[jainxrohit]

== RELEASE NOTES ==

Hive Changes
* Allow reading data from HDFS while caching the fetched data on local disks. Turn on the feature by specifying the cache directory config `cache.base-directory`.

[highker]

some comments

[highker]

Could you remove the period in the commit title? https://chris.beams.io/posts/git-commit/ is a good commit message guideline.

[jainxrohit]

Could you remove the period in the commit title? https://chris.beams.io/posts/git-commit/ is a good commit message guideline.

Nice article, fixed the commit message.

[shixuan-fan]

LGTM % nits

[highker]

The test failure is due to permission/auth setting. Try overriding the following function in CachingFileSystem

    @Override
    public void setPermission(Path path, FsPermission permission)
            throws IOException
    {
        dataTier.setPermission(path, permission);
    }

But in general, I would suggest overriding all default functions from FileSystem. A good example is FilterFileSystem

[highker]

coding style comments

[highker]

some comments

[highker]

LGTM. @shixuan-fan, could you give it a final review and merge it?

[shixuan-fan]

LGTM, ideally we would want to have three commits:

• Fixing caching file system
• Raptor side change
• Hive side change

But since it is already reviewed, I won't bother breaking it down. I'll merge it once we've completed the internal repo pull request that adapt to this one.