fix(connector): unblock fetch_watermarks in rdkafka

[wangrunji0408]

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

What's changed and what's your intention?

This PR bumps madsim-rdkafka. Operations in method fetch_watermarks is now wrapped in tokio::task::spawn_blocking to avoid blocking the current thread. https://github.com/madsim-rs/madsim/pull/196/files https://github.com/madsim-rs/madsim/pull/200/files

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.

[tabVersion]

lgtm as no business logic change

[xxchan]

I'm not sure if I understand it correctly:

• Upgrading rdkafka (chore(deps): update librdkafka to v2.3.0 #15313) already fixed the hanging issue
• This PR (spawan_blocking) isn't a must-have now, but it can avoid hanging even if there's bug in rdkafka. (can make async timeout work)

Right?

[wangrunji0408]

Correct

[wangrunji0408]

Meta node exited with this error:

2024-03-05T10:04:50.38429495Z ERROR risingwave_meta::stream::source_manager: error happened when tick from connector source worker error=failed to fetch metadata from kafka (message_queue:29092): Kafka error: Meta data fetch error: BrokerTransportFailure (Local: Broker transport failure)

Don't know why. 😕

[tabVersion]

Meta node exited with this error:

2024-03-05T10:04:50.38429495Z ERROR risingwave_meta::stream::source_manager: error happened when tick from connector source worker error=failed to fetch metadata from kafka (message_queue:29092): Kafka error: Meta data fetch error: BrokerTransportFailure (Local: Broker transport failure)

Don't know why. 😕

seems a network issue? not sure rerun helps

[tabVersion]

got an error from inner

ERROR librdkafka: librdkafka: DESTROY [thrd:app]: Failed to join internal main thread: Resource temporarily unavailable (was process forked?)  

seems we break the rdkafka inner @wangrunji0408

[wangrunji0408]

Humm, looks like the raw kafka client can not be sent to another thread. But librdkafka claims it can be called from any thread.

Yes, librdkafka is completely thread-safe (unless otherwise noted in the API documentation). Any API, short of the destructor functions, may be called at any time from any thread.

[tabVersion]

Humm, looks like the raw kafka client can not be sent to another thread. But librdkafka claims it can be called from any thread.

Yes, librdkafka is completely thread-safe (unless otherwise noted in the API documentation). Any API, short of the destructor functions, may be called at any time from any thread.

Seems a kafka issue. Let's leave this in backlog.

[BugenZhao]

Humm, looks like the raw kafka client can not be sent to another thread. But librdkafka claims it can be called from any thread.

Yes, librdkafka is completely thread-safe (unless otherwise noted in the API documentation). Any API, short of the destructor functions, may be called at any time from any thread.

https://github.com/madsim-rs/madsim/pull/196/files?diff=split#r1535237591 👀

[yuhao-su]

Any reason this pr not merged?