KAFKA-10370: consumer.seek() with SinkTaskContext's offsets when initialize #9145
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
ning2008wisc
changed the title
ning2008wisc
force-pushed
the
kafka-10370
branch
2 times, most recently
from
f794ad5
to
0000fc5
Compare
ning2008wisc
changed the title
ning2008wisc
force-pushed
the
kafka-10370
branch
from
0000fc5
to
8e9ccb3
Compare
…er when initialize
ning2008wisc
force-pushed
the
kafka-10370
branch
from
8e9ccb3
to
9fcb309
Compare
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Outdated suggestions cannot be applied.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.
In WorkerSinkTask.java, when we want the consumer to consume from certain offsets, rather than from the last committed offset, WorkerSinkTaskContext provided a way to supply the offsets from external (e.g. an implementation of SinkTask) to rewind the consumer.
In the poll() method, it first calls rewind() to read the offsets from WorkerSinkTaskContext, if the offsets are not empty, then in the rewind() method, run consumer.seek(tp, offset) to rewind the consumer.
As a part of WorkerSinkTask initialization, when the SinkTask starts, we can supply the specific offsets by context.offset(supplied_offsets); in start() method, so that when the consumer does the first poll in WorkerSinkTask, it should rewind to the specific offsets in rewind() method. However in practice, we saw the following IllegalStateException when running consumer.seek(tp, offsets);
[2020-08-07 23:53:55,752] INFO WorkerSinkTask{id=MirrorSinkConnector-0} Rewind test-1 to offset 3 (org.apache.kafka.connect.runtime.WorkerSinkTask:648) [2020-08-07 23:53:55,752] INFO [Consumer clientId=connector-consumer-MirrorSinkConnector-0, groupId=connect-MirrorSinkConnector] Seeking to offset 3 for partition test-1 (org.apache.kafka.clients.consumer.KafkaConsumer:1592) [2020-08-07 23:53:55,752] ERROR WorkerSinkTask{id=MirrorSinkConnector-0} Task threw an uncaught and unrecoverable exception (org.apache.kafka.connect.runtime.WorkerTask:187) java.lang.IllegalStateException: No current assignment for partition test-1 at org.apache.kafka.clients.consumer.internals.SubscriptionState.assignedState(SubscriptionState.java:368) at org.apache.kafka.clients.consumer.internals.SubscriptionState.seekUnvalidated(SubscriptionState.java:385) at org.apache.kafka.clients.consumer.KafkaConsumer.seek(KafkaConsumer.java:1597) at org.apache.kafka.connect.runtime.WorkerSinkTask.rewind(WorkerSinkTask.java:649) at org.apache.kafka.connect.runtime.WorkerSinkTask.poll(WorkerSinkTask.java:334) at org.apache.kafka.connect.runtime.WorkerSinkTask.iteration(WorkerSinkTask.java:229) at org.apache.kafka.connect.runtime.WorkerSinkTask.execute(WorkerSinkTask.java:198) at org.apache.kafka.connect.runtime.WorkerTask.doRun(WorkerTask.java:185) at org.apache.kafka.connect.runtime.WorkerTask.run(WorkerTask.java:235) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) [2020-08-07 23:53:55,752] ERROR WorkerSinkTask{id=MirrorSinkConnector-0} Task is being killed and will not recover until manually restarted (org.apache.kafka.connect.runtime.WorkerTask:188)As suggested in https://stackoverflow.com/questions/41008610/kafkaconsumer-0-10-java-api-error-message-no-current-assignment-for-partition/41010594, the resolution (that has been initially verified) proposed in this PR is to use consumer.assign() with consumer.seek(), instead of consumer.subscribe, to handle the initial position of the consumer, when specific offsets are provided by external through WorkerSinkTaskContext