[flink] Immediately throw Exception in mailbox executor rather than wait until next record in flink sink writer. #736
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
…ait until next record.
|
@wuchong @swuferhong , CC |
luoyuxia
reviewed
Apr 14, 2025
| @@ -123,6 +130,18 @@ public void write(RowData value, Context context) throws IOException, Interrupte | |||
|
|
|||
| InternalRow internalRow = sinkRow.replace(value); | |||
| CompletableFuture<?> writeFuture = writeRow(value.getRowKind(), internalRow); | |||
| writeFuture.whenComplete( | |||
There was a problem hiding this comment.
Just wondering what's the problem if checkAsyncException in writing new rows?
There was a problem hiding this comment.
Two reasons:
- case 1: The KafkaWriter flushing needs to wait for all in-flight records to send successfully. What if one record is stuck for some reason, the job will be stuck.
- case 2: If data is little, the failover will be thrown at a long time later. (if the former is a window, a heavy calculate will be replay.)
wuchong
reviewed
Apr 14, 2025
| @@ -16,7 +16,7 @@ | |||
|
|
|||
| # Set root logger level to OFF to not flood build logs | |||
| # set manually to INFO for debugging purposes | |||
| rootLogger.level = OFF | |||
| rootLogger.level = DEBUG | |||
| } | ||
|
|
||
| // Checking for exceptions from previous writes | ||
| // Checking for exceptions from previous writes |
| // Checking for exceptions from previous writes | ||
| mailboxExecutor.execute(this::checkAsyncException, "Update error metric"); | ||
| } | ||
| }); | ||
| writeFuture.exceptionally( |
| } | ||
|
|
||
| private void testExceptionWhenFlussUnavailable( | ||
| BiConsumer<FlinkSinkWriter, MailboxExecutor> actionAfterFlussUnavailable) |
There was a problem hiding this comment.
The actionAfterFlussUnavailable is not used.
| BiConsumer<FlinkSinkWriter, MailboxExecutor> actionAfterFlussUnavailable) | ||
| throws Exception { | ||
| FlussClusterExtension flussClusterExtension = FlussClusterExtension.builder().build(); | ||
| flussClusterExtension.start(); |
There was a problem hiding this comment.
use try ... catch ... finally to make sure the flussClusterExtension is closed if any exception is thrown in the middle.
| testExceptionWhenFlussUnavailable( | ||
| (writer, mailboxExecutor) -> { | ||
| writer.getTableWriter().flush(); | ||
| writer.getTableWriter().flush(); |
| testExceptionWhenFlussUnavailable( | ||
| (writer, mailboxExecutor) -> { | ||
| writer.getTableWriter().flush(); | ||
| assertThatThrownBy(() -> writer.flush(true)) |
There was a problem hiding this comment.
Add some comment for the tests that why the lines will throw exception, but tableWriter#flush() doesn't.
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.
Purpose
Linked issue: close #735
Brief change log
Tests
Add some failover test in FlinkSinkWriterTest.
API and Format
Documentation