[SPARK-40844][SS] Flip the default value of Kafka offset fetching config #38306
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Author
gaborgsomogyi
approved these changes
Oct 19, 2022
Contributor
gaborgsomogyi
left a comment
gaborgsomogyi
left a comment
There was a problem hiding this comment.
LGTM. I would mention one minor thing somewhere. There will be time when Kafka is not going to provide the old way of offset fetching so migration highly is advised. When the consumer is not going to be able to fetch offsets then the community has basically 2 options:
- Drop the old feature
- Not upgrading Kafka version
Temporarily not upgrading Kafka can be done but the megatrend is that the old feature will be dropped. Maybe we can deprecate the old fetching feature now?!
Contributor
Author
|
Kafka community deprecated the method |
Contributor
|
Sounds fair, ship it then :) |
dongjoon-hyun
approved these changes
Oct 19, 2022
Member
|
cc @viirya , too |
viirya
approved these changes
Oct 19, 2022
Contributor
Author
|
https://github.com/HeartSaVioR/spark/actions/runs/3279380674/jobs/5400649338 Tests are failing from below suites
which none of them are related to this change. |
Contributor
Author
|
Thanks! Merging to master. |
SandishKumarHN
pushed a commit
to SandishKumarHN/spark
that referenced
this pull request
Dec 12, 2022
### What changes were proposed in this pull request? This PR proposes to flip the default value of Kafka offset fetching config (`spark.sql.streaming.kafka.useDeprecatedOffsetFetching`) from `true` to `false`, which enables AdminClient based offset fetching by default. ### Why are the changes needed? We had been encountered several production issues with old offset fetching (e.g. hang, issue with Kafka consumer group rebalance) which could be mitigated with new offset fetching. Despite the breaking change on the ACL, there is no need for moderate users to suffer from the old way. The discussion went through the dev. mailing list: https://lists.apache.org/thread/spkco94gw33sj8355mhlxz1vl7gl1g5c ### Does this PR introduce _any_ user-facing change? Yes, especially users who relies on Kafka ACL based on consumer group. They need to either adjust the ACL to topic based one, or set the value to `true` for `spark.sql.streaming.kafka.useDeprecatedOffsetFetching` to use the old approach. ### How was this patch tested? Existing UTs. Closes apache#38306 from HeartSaVioR/SPARK-40844. Authored-by: Jungtaek Lim <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]>
HeartSaVioR
pushed a commit
that referenced
this pull request
Dec 16, 2024
…afkaOffsetReaderConsumer` ### What changes were proposed in this pull request? This was missed when the default value of configuration is changed via the following. - #38306 KafkaMicroBatchSourceSuite consists of set of different suites, where KafkaMicroBatchSourceSuiteBase based suite is defined. There are 4 implementations of this abstract class for now: 1. `KafkaMicroBatchV1SourceSuite` - V1 source that supposes to use `KafkaOffsetReaderConsumer` as `KafkaOffsetReader`. 2. `KafkaMicroBatchV2SourceSuite` - V2 source that supposes to use `KafkaOffsetReaderConsumer` as `KafkaOffsetReader`. 3. `KafkaMicroBatchV1SourceWithAdminSuite` - V1 source that uses `KafkaOffsetReaderAdmin` as `KafkaOffsetReader`. 4. `KafkaMicroBatchV2SourceWithAdminSuite` - V2 source that uses `KafkaOffsetReaderAdmin` as `KafkaOffsetReader`. But `KafkaMicroBatchV1SourceSuite` and `KafkaMicroBatchV2SourceSuite` are still running based on `KafkaOffsetReaderAdmin`, as `USE_DEPRECATED_KAFKA_OFFSET_FETCHING` is `false` be default. By switching it to `true` in `beforeAll`, we make sure that corresponding `KafkaOffsetReader` is in use. ### Why are the changes needed? To improve unit tests coverage for `KafkaOffsetReaderConsumer` ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Unit Tests. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #49164 from ostronaut/hotifx/KafkaMicroBatchSourceSuite-cover-KafkaOffsetReaderConsumer. Authored-by: Dima <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
This PR proposes to flip the default value of Kafka offset fetching config (
spark.sql.streaming.kafka.useDeprecatedOffsetFetching) fromtruetofalse, which enables AdminClient based offset fetching by default.Why are the changes needed?
We had been encountered several production issues with old offset fetching (e.g. hang, issue with Kafka consumer group rebalance) which could be mitigated with new offset fetching. Despite the breaking change on the ACL, there is no need for moderate users to suffer from the old way.
The discussion went through the dev. mailing list: https://lists.apache.org/thread/spkco94gw33sj8355mhlxz1vl7gl1g5c
Does this PR introduce any user-facing change?
Yes, especially users who relies on Kafka ACL based on consumer group. They need to either adjust the ACL to topic based one, or set the value to
trueforspark.sql.streaming.kafka.useDeprecatedOffsetFetchingto use the old approach.How was this patch tested?
Existing UTs.