feat(compaction): add vnode watermark picker

[zwang28]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

close #17157

• Add a new picker that picks trivial reclaim compaction task. For simplicity, this picker only takes bottommost level into consideration.
• Extract SkipWatermarkState::from_safe_epoch_watermarks and HummockVersion::safe_epoch_table_watermarks into pure functions for easier reuse.
• Add a CompactionSelectorContext to convey parameters to compaction selectors.

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added test labels as necessary. See details.
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

• My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

• My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users)

Release note

If this PR includes changes that directly affect users or other significant modifications relevant to the community, kindly draft a release note to provide a concise summary of these changes. Please prioritize highlighting the impact these changes will have on users.

[zwang28]

One issue is compaction tasks of dynamic type has the highest priority, potentially starving space reclaim tasks.

[hzxa21]

One issue is compaction tasks of dynamic type has the highest priority, potentially starving space reclaim tasks.

The assumptions are:

1. dynamic task should be done ASAP because it will cause write stall
2. dynamic task will also reclaim tombstones/space if it can

2 is not 100% valid because dynamic task will not take reclaim releated statistics like tombstone ratio into consideration. Also, for trivial reclaim task, which uses very few resource, it can be done relatively fast and won't block dynamic tasks.

However, this is a separate issue pre-exists prior to this PR so we don't need to block this PR because of that.

[hzxa21]

LGTM.

Do we have a integration test (like the one in compaction test pipeline maybe?) to verify the effectiveness of the new picker?

[cyliu0]

/test

[risingwave-ci]

Failed: 1
Passed:0
FailedExecutions:
https://buildkite.com/risingwave-test/compaction-test/builds/176
PassedExecutions:

Please ignore this, it's failed because of the toml issue in the pipeline

[risingwave-ci]

Failed: 1
Passed:0
FailedExecutions:
https://buildkite.com/risingwave-test/compaction-test/builds/176
PassedExecutions:

[cyliu0]

/test

[risingwave-ci]

Failed: 0
Passed:2
FailedExecutions:

PassedExecutions:
https://buildkite.com/risingwave-test/compaction-test/builds/177
https://buildkite.com/risingwave-test/compaction-test/builds/177

[zwang28]

Do we have a integration test (like the one in compaction test pipeline maybe?) to verify the effectiveness of the new picker?

I'm still considering how to add the corresponding tests.

[zwang28]

Do we have a integration test (like the one in compaction test pipeline maybe?) to verify the effectiveness of the new picker?

I'm still considering how to add the corresponding tests.

I ended up with adding a simulation integration test.
I didn't add one in compaction test pipeline because I needed risectl to move the table to dedicated compaction group, in order to prevent other table's compaction from complicating the tests. The compaction test pipeline doesn't support risectl yet.