[FLINK-36114] Make SchemaRegistryRequestHandler thread safe by blocking more schema change events
#3563
Conversation
This file contains 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
…king more schema change events
leonardBang
reviewed
Aug 22, 2024
...-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/SchemaOperator.java
Outdated
Show resolved
Hide resolved
.../org/apache/flink/cdc/runtime/operators/schema/coordinator/SchemaRegistryRequestHandler.java
Outdated
Show resolved
Hide resolved
.../org/apache/flink/cdc/runtime/operators/schema/coordinator/SchemaRegistryRequestHandler.java
Outdated
Show resolved
Hide resolved
|
Thanks @yuxiqian for the contribution, I left some comments @loserwang1024 Would you like to take a look at this PR whrn you have time? |
...-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/SchemaOperator.java
Show resolved
Hide resolved
...-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/SchemaOperator.java
Outdated
Show resolved
Hide resolved
...-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/SchemaOperator.java
Outdated
Show resolved
Hide resolved
| TableId tableId = event.tableId(); | ||
| Optional<Schema> latestSchema = getLatestOriginalSchema(tableId); | ||
| return Boolean.TRUE.equals( | ||
| SchemaChangeEventVisitor.visit( |
There was a problem hiding this comment.
I wonder we should do it in registry or sink? @leonardBang , WDYT?
leonardBang
approved these changes
Aug 22, 2024
There was a problem hiding this comment.
Thanks @yuxiqian for the cooperation fix, LGTM, I believe your effort will make the schema evolution more stable.
yuxiqian
added a commit
to yuxiqian/flink-cdc
that referenced
this pull request
Aug 22, 2024
…afe by blocking subsequent schemaChangeEvent This closes apache#3563. Co-authored-by: Hongshun Wang <loserwang1024@gmail.com> (cherry picked from commit ee843e2)
leonardBang
pushed a commit
to yuxiqian/flink-cdc
that referenced
this pull request
Aug 27, 2024
…afe by blocking subsequent schemaChangeEvent This closes apache#3563. Co-authored-by: Hongshun Wang <loserwang1024@gmail.com> (cherry picked from commit ee843e2)
qiaozongmi
pushed a commit
to qiaozongmi/flink-cdc
that referenced
this pull request
Sep 23, 2024
…afe by blocking subsequent schemaChangeEvent This closes apache#3563. Co-authored-by: Hongshun Wang <loserwang1024@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This closes FLINK-36114.
Currently, SchemaRegistry asynchronously receives schema change requests from SchemaOperator, and results of multiple requests might got mixed up together, causing incorrect logic flow in multiple parallelism cases.
Changing SchemaRegistry's behavior to accept requests in serial should resolve this problem.
This PR also changes MySQL pipeline source behavior.
Previously, some subTask might not emit CreateTableEvents if there are no corresponding SnapshotSplit assigned to them. This might cause problems if they are assigned some StreamSplits later, since their downstream Transform nodes never received CreateTableEvent and doesn't know about the schema.
By forcing MySQL source emitting CreateTableEvents when transiting from Snapshot to BinLog stage should resolve this problem.
This PR also adjusts Pipeline E2e test case parallelism to 1 to 4 to verify this change.