Speed ​​up rocksdb checkpoint

[adol001]

Is your feature request related to a problem? Please describe. We now store 10 billion alluxio inode metadata using rocksdb. The current alluxio checkpoint will tar.gz rocksdb data file by single thread, alluxio restart or re-election leader will untar.gz by single thread. This speed is not fast.

Describe the solution you'd like To divide the rocksdb data into parts and tar.gz each part separately, so that rocksdb write to checkpoint or restore from checkpoint can speed up the speed

[adol001]

In addition to checkpoint, backup is also single-threaded gz. It should also be optimized too.

[yuzhu]

@jenoudet FYI as well.

[adol001]

Refine checkpoint by Parallel zip compression and decompression https://github.com/Alluxio/alluxio/pull/16151

The current implementation of backup, instead of executing backup, it is better to reload it

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in two weeks if no further activity occurs. Thank you for your contributions.