Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

refactor(meta): build actors in inject barrier #18270

Merged
merged 5 commits into from
Aug 29, 2024

Conversation

wenym1
Copy link
Contributor

@wenym1 wenym1 commented Aug 27, 2024

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

What's changed and what's your intention?

After #18104, when handling BuildActor request, we simply spawn the actors and save the join handle. After the process of building actors is simplified, we can further postpone spawning actor till inject barrier, so that

  • stream service rpc broadcast_actor_info_table, update_actors and build_actors can be removed, and the works of these rpc can be done together in InjectBarrier.
  • on CN side, all spawned actors are part of the streaming graph, and we don't need to handle the case of actors that are spawned but not joined in the streaming graph
  • we can avoid callingdrop_actors while cancelling a command, because for a create streaming job command that is enqueued but not handled yet, its actor is not spawned yet, and there is no need to further drop the actors.

In this PR, for simplicity, most changes are on the meta node. The actor locations to be broadcast and the BuildActorInfo is included in the InjectBarrierRequest. On local barrier manager, we still reuse the current implemented methods. The local barrier manager will be refactored in future PRs.

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.

@MrCroxx
Copy link
Contributor

MrCroxx commented Aug 28, 2024

AMAZING 🥵

let mut map = match job_type {
CreateStreamingJobType::Normal => HashMap::new(),
CreateStreamingJobType::SinkIntoTable(replace_table) => {
replace_table.new_table_fragments.actors_to_create()
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Great job!! Looks like #14904 can be resolved naturally!

Copy link
Contributor

@shanicky shanicky left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

scaling part lgtm

Copy link
Member

@yezizp2012 yezizp2012 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!! It definitely simplify the barrier processing and actor building. 🤔 We need to rethink about #16742 and find another solution.

Copy link
Member

@BugenZhao BugenZhao left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

My only concern is that this may increase the latency for injecting barriers, which could slow down the barrier manager loop. Rest LGTM.

src/meta/src/barrier/creating_job/status.rs Outdated Show resolved Hide resolved
src/meta/src/model/stream.rs Outdated Show resolved Hide resolved
@wenym1 wenym1 added this pull request to the merge queue Aug 29, 2024
Merged via the queue into main with commit fac2904 Aug 29, 2024
33 of 34 checks passed
@wenym1 wenym1 deleted the yiming/build-actor-in-barrier branch August 29, 2024 05:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants