feat(barrier): support database failure isolation (part 2, local)

[wenym1]

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

What's changed and what's your intention?

After #19664. The local barrier manager part of database failure isolation.

Previously in ManagedBarrierState, all databases are in a Running status, as a DatabaseManagedBarrierState, because in case of any failure, we will wait for a global recovery. To support database failure isolation, besides the Running status, we will have the following statuses.

pub(crate) enum DatabaseStatus {
    Running(DatabaseManagedBarrierState),
    Suspended(SuspendedDatabaseState),
    Resetting(ResettingDatabaseState),
}

The lifecycle of a DatabaseStatus will be

1. Get created when the control stream is reset, or when handling add_partial_graph request. The status is at Running.
2. In case of actor failure, it will send to the meta global barrier manager with a ReportDatabaseFailureResponse, and enter the Suspended.
3. Wait for the ResetDatabaseRequest from the meta global barrier manager, and then enter Resetting, which spawns a task to clear all actors and clear hummock uploader state of the state tables. Note that we may enter the Resetting directly from Running when the database reset is triggered by failure reported from other CNs.
4. After the task finishes, get the DatabaseStatus removed from ManagedBarrierState, and will be recreated by the subsequent add_partial_graph request from the meta global barrier manager.

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.

[wenym1]

• feat(barrier): support database failure isolation (part 2, local) #19579 👈 (View in Graphite)
• feat(barrier): support database failure isolation (part 1, meta) #19664
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[hzxa21]

The logic LGTM. I think we need some tests to trigger per DB failure and make sure other DBs are unaffected. Simulation tests are preferred.

[hzxa21]

We skip try_find_root_actor_failure here. Is it intentional?

[wenym1]

Yes, because this error is likely to be caused by hummock sync error, and is not related to actor failure.