Configure block and inode metastore separately

[bzheng888]

What changes are proposed in this pull request?

Separate configuration for block and inode is make sense, because in some cases, blockstore use ROCKS is enough, but inodestore must use HEAP, we want to distinguish the difference between these two types.

Why are the changes needed?

Sometimes we want to config inode metastore type and block metastore type seperately.

Does this PR introduce any user facing changes?

Please list the user-facing changes introduced by your change, including

1. change in user-facing APIs Some configuration changes

[alluxio-bot]

Automated checks report:

• Commits associated with Github account: PASS
• PR title follows the conventions: FAIL
  • The title of the PR does not pass all the checks. Please fix the following issues:
    • First word of title ("Seperate") is not an imperative verb. Please use one of the valid words

Some checks failed. Please fix the reported issues and reply 'alluxio-bot, check this please' to re-run checks.

[alluxio-bot]

Automated checks report:

• Commits associated with Github account: PASS
• PR title follows the conventions: PASS

All checks passed!

[codecov-commenter]

Codecov Report

Merging #14565 (e8f64e9) into master (ea4d766) will decrease coverage by 24.87%. The diff coverage is 73.91%.

@@              Coverage Diff              @@
##             master   #14565       +/-   ##
=============================================
- Coverage     41.83%   16.96%   -24.88%     
+ Complexity     9326     2702     -6624     
=============================================
  Files          1494     1494               
  Lines         87696    87841      +145     
  Branches      10451    10461       +10     
=============================================
- Hits          36691    14900    -21791     
- Misses        48021    71698    +23677     
+ Partials       2984     1243     -1741     

Impacted Files	Coverage Δ
...ommon/src/main/java/alluxio/util/FeatureUtils.java	0.00% <0.00%> (ø)
...ster/src/main/java/alluxio/master/MasterUtils.java	55.17% <66.66%> (-4.09%)	:arrow_down:
...common/src/main/java/alluxio/conf/PropertyKey.java	98.07% <100.00%> (-1.38%)	:arrow_down:
...mon/src/main/java/alluxio/shell/CommandReturn.java	0.00% <0.00%> (-100.00%)	:arrow_down:
...mon/src/main/java/alluxio/util/ExceptionUtils.java	0.00% <0.00%> (-100.00%)	:arrow_down:
...n/src/main/java/alluxio/wire/AlluxioProxyInfo.java	0.00% <0.00%> (-100.00%)	:arrow_down:
.../src/main/java/alluxio/wire/AlluxioMasterInfo.java	0.00% <0.00%> (-100.00%)	:arrow_down:
...src/main/java/alluxio/job/meta/JobIdGenerator.java	0.00% <0.00%> (-100.00%)	:arrow_down:
...n/src/main/java/alluxio/stress/BaseParameters.java	0.00% <0.00%> (-100.00%)	:arrow_down:
...src/main/java/alluxio/client/UnderStorageType.java	0.00% <0.00%> (-100.00%)	:arrow_down:
... and 691 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update ea4d766...e8f64e9. Read the comment docs.

[bzheng888]

@maobaolong Done, PTAL!

[bzheng888]

It would be better if you KEEP MASTER_METASTORE propertykey and if it setByUser, the original logic should honor the configured value and ignored the MASTER_BLOCK_METASTORE and MASTER_INODE_METASTORE config

Thanks for pointing this, update this pr, PTAL!

[bzheng888]

Hi, @jiacheliu3, can you take some time to review this pr?

[jiacheliu3]

@bzheng888 @maobaolong Can you please clarify on the use case? I'm not following. This is a very big change.

FYI: @apc999 @yuzhu @ggezer

[bzheng888]

@jiacheliu3 Thank you for your reply, the main idea for this pr is to config master inode metastore and block metastore separately. Talking about cases, if our case are big files, and the amount of inode metadata is not that much, we can config inode metastore as heap, but the block metadata will be very large, it's better to use rocks metastore to support large namespace.

[bzheng888]

@maobaolong @jiacheliu3 @dbw9580 Update this pr, PTAL!

[maobaolong]

@bzheng888 Will take a quick look soon

[jiacheliu3]

@bzheng888 @maobaolong Please see my previous comments above. Thx!

[bzheng888]

Don't know why failed this test, seems not relate to this pr?

[bzheng888]

@jiacheliu3 @elega PTAL

[bzheng888]

Resolve the conflicts

[jiacheliu3]

alluxio-bot, merge this please