Multiple consumers read from the same partition after rebalance #1200
Comments
|
Might be related to the #127 |
7 tasks
erikvanoosten
added a commit
to zio/zio-kafka
that referenced
this issue
Nov 5, 2023
Fixes #590 "Many records duplicately processed after rebalancing" In this change we introduce a new mode that holds up a rebalance until all messages that were provided to the stream of a revoked partition, have been committed. ### Motivation Here is a common (single partition) scenario around rebalances: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100) 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance 1. the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets 1. _at the same time,_ another consumer on another instance, starts consuming from the last committed offset (which is 50) and will process the same messages with offsets 50 to 100 Messages with offsets 50 to 100 are being processed by both consumers simultaneously. Note that both consumers will try to commit these offsets. Until the first consumer is ready, the stored offsets can go up and down and are therefore unreliable. After merging this change, the scenario will unfold as follows: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100). Zio-kafka keeps track of the highest provided offset 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance * the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets * inside the onRevoked callback, zio-kafka continues to process commit commands from the user * zio-kafka continues to do so until the commit with the highest provided offset (offset 100) completes * the onRevoked callback completes, signalling to Kafka that the next consumer may start consuming from the partition 1. another consumer on another instance, starts consuming from the last committed offset (which is now 100, problem solved!) ### Commit queue Because both the main runloop, and the rebalance listener need to process (and thus receive) commits commands, the commit commands were moved to a separate queue. Because the main runloop may still need to be kickstarted when it is no longer polling, a new command `CommitAvailable` was introduced. ### Complications 1. The chosen solution is not suitable for all consumers. - There are use cases where not all messages are read from the stream. For example, some want to read exactly 100 messages from a topic and then stop consuming. In that case the user has no intention to commit all messages, and therefore we should not wait for that to happen. Since stream consumers can basically do whatever they want, the only way we can support such use cases is by letting the consumer tell zio-kafka that they are done with committing. This requires an API change. For example, we can let the user tell zio-kafka that a given commit is the last one. - Not all consumers commit offsets (to Kafka) in the first place. In a future change we could make it work for commits to other stores though. As a workaround, these users can commit to both places. 1. It requires Kafka client 3.6.0. In earlier versions there was no way to wait for async commits to complete. ### Same thread executor The Kafka client requires that any nested invocations (that is, from the rebalance listener callback) to the java consumer happens from the same thread. This is very much at odds with how ZIO works. Attempts to convince the Kafka committers to relax this requirement failed; they could not be convinced that this is a problem. This is circumvented by using a special same-thread-runtime which runs on the thread of the caller. However, some operations such as `ZIO.timeout` and anything with `Schedules` will still shift work to another thread. We work around this by using blocking time. ### Collateral This change also: - fixes order of `private` and `final` - removes some completely useless tests ### Related The same issue is present in: - f2s-kafka: fd4s/fs2-kafka#1200 - alpakka-kafka: akka/alpakka-kafka#1038 In fact, every program that does polls and commits asynchronously is likely affected. ### Non-goals This change does not try to solve the following goals. However, these can be addressed in future PRs. - Awaiting commits after stopping the consumer, e.g. due to program shutdown (see #1087). - Support consumers that want to commit only a portion of the given messages. - Support transactional consumer/producer. - Support external commits. This branch is based on the work of abandoned PRs #788 and #830 and builds on preparatory work in PRs #744, #1068, #1073 #1086, #1089 and #1097.
erikvanoosten
added a commit
to zio/zio-kafka
that referenced
this issue
Nov 16, 2023
Fixes #590 "Many records duplicately processed after rebalancing" In this change we introduce a new experimental mode that holds up a rebalance until all messages that were provided to the stream of a revoked partition, have been committed. ### Motivation Here is a common (single partition) scenario around rebalances: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100) 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance 1. the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets 1. _at the same time,_ another consumer on another instance, starts consuming from the last committed offset (which is 50) and will process the same messages with offsets 50 to 100 Messages with offsets 50 to 100 are being processed by both consumers simultaneously. Note that both consumers will try to commit these offsets. Until the first consumer is ready, the stored offsets can go up and down and are therefore unreliable. After merging this change, the scenario will unfold as follows: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100). Zio-kafka keeps track of the highest provided offset 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance * the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets * inside the onRevoked callback, zio-kafka continues to process commit commands from the user * zio-kafka continues to do so until the commit with the highest provided offset (offset 100) completes * the onRevoked callback completes, signalling to Kafka that the next consumer may start consuming from the partition 1. another consumer on another instance, starts consuming from the last committed offset (which is now 100, problem solved!) ### Commit queue Because both the main runloop, and the rebalance listener need to process (and thus receive) commits commands, the commit commands were moved to a separate queue. Because the main runloop may still need to be kickstarted when it is no longer polling, a new command `CommitAvailable` was introduced. ### Complications 1. The chosen solution is not suitable for all consumers. - There are use cases where not all messages are read from the stream. For example, some want to read exactly 100 messages from a topic and then stop consuming. In that case the user has no intention to commit all messages, and therefore we should not wait for that to happen. Since stream consumers can basically do whatever they want, the only way we can support such use cases is by letting the consumer tell zio-kafka that they are done with committing. This requires an API change. For example, we can let the user tell zio-kafka that a given commit is the last one. - Not all consumers commit offsets (to Kafka) in the first place. In a future change we could make it work for commits to other stores though. As a workaround, these users can commit to both places. 1. It requires Kafka client 3.6.0. In earlier versions there was no way to wait for async commits to complete. ### Same thread executor The Kafka client requires that any nested invocations (that is, from the rebalance listener callback) to the java consumer happens from the same thread. This is very much at odds with how ZIO works. Attempts to convince the Kafka committers to relax this requirement failed; they could not be convinced that this is a problem. This is circumvented by using a special same-thread-runtime which runs on the thread of the caller. However, some operations such as `ZIO.timeout` and anything with `Schedules` will still shift work to another thread. We work around this by using blocking time. ### Experimental Because holding up the rebalance may have unforeseen consequences, this feature is marked as experimental. This allows us to collect experiences before we recommend this mode to all users. ### Collateral This change also: - fixes order of `private` and `final` - removes some completely useless tests ### Related The same issue is present in: - f2s-kafka: fd4s/fs2-kafka#1200 - alpakka-kafka: akka/alpakka-kafka#1038 In fact, every program that does polls and commits asynchronously is likely affected. ### Non-goals This change does not try to solve the following goals. However, these can be addressed in future PRs. - Awaiting commits after stopping the consumer, e.g. due to program shutdown (see #1087). - Support consumers that want to commit only a portion of the given messages. - Support transactional consumer/producer. - Support external commits. This branch is based on the work of abandoned PRs #788 and #830 and builds on preparatory work in PRs #744, #1068, #1073 #1086, #1089 and #1097.
erikvanoosten
added a commit
to zio/zio-kafka
that referenced
this issue
Nov 16, 2023
Fixes #590 "Many records duplicately processed after rebalancing" In this change we introduce a new experimental mode that holds up a rebalance until all messages that were provided to the stream of a revoked partition, have been committed. ### Motivation Here is a common (single partition) scenario around rebalances: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100) 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance 1. the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets 1. _at the same time,_ another consumer on another instance, starts consuming from the last committed offset (which is 50) and will process the same messages with offsets 50 to 100 Messages with offsets 50 to 100 are being processed by both consumers simultaneously. Note that both consumers will try to commit these offsets. Until the first consumer is ready, the stored offsets can go up and down and are therefore unreliable. After merging this change, the scenario will unfold as follows: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100). Zio-kafka keeps track of the highest provided offset 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance * the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets * inside the onRevoked callback, zio-kafka continues to process commit commands from the user * zio-kafka continues to do so until the commit with the highest provided offset (offset 100) completes * the onRevoked callback completes, signalling to Kafka that the next consumer may start consuming from the partition 1. another consumer on another instance, starts consuming from the last committed offset (which is now 100, problem solved!) ### Commit queue Because both the main runloop, and the rebalance listener need to process (and thus receive) commits commands, the commit commands were moved to a separate queue. Because the main runloop may still need to be kickstarted when it is no longer polling, a new command `CommitAvailable` was introduced. ### Complications 1. The chosen solution is not suitable for all consumers. - There are use cases where not all messages are read from the stream. For example, some want to read exactly 100 messages from a topic and then stop consuming. In that case the user has no intention to commit all messages, and therefore we should not wait for that to happen. Since stream consumers can basically do whatever they want, the only way we can support such use cases is by letting the consumer tell zio-kafka that they are done with committing. This requires an API change. For example, we can let the user tell zio-kafka that a given commit is the last one. - Not all consumers commit offsets (to Kafka) in the first place. In a future change we could make it work for commits to other stores though. As a workaround, these users can commit to both places. 1. It requires Kafka client 3.6.0. In earlier versions there was no way to wait for async commits to complete. ### Same thread executor The Kafka client requires that any nested invocations (that is, from the rebalance listener callback) to the java consumer happens from the same thread. This is very much at odds with how ZIO works. Attempts to convince the Kafka committers to relax this requirement failed; they could not be convinced that this is a problem. This is circumvented by using a special same-thread-runtime which runs on the thread of the caller. However, some operations such as `ZIO.timeout` and anything with `Schedules` will still shift work to another thread. We work around this by using blocking time. ### Experimental Because holding up the rebalance may have unforeseen consequences, this feature is marked as experimental. This allows us to collect experiences before we recommend this mode to all users. ### Collateral This change also: - fixes order of `private` and `final` - removes some completely useless tests ### Related The same issue is present in: - f2s-kafka: fd4s/fs2-kafka#1200 - alpakka-kafka: akka/alpakka-kafka#1038 In fact, every program that does polls and commits asynchronously is likely affected. ### Non-goals This change does not try to solve the following goals. However, these can be addressed in future PRs. - Awaiting commits after stopping the consumer, e.g. due to program shutdown (see #1087). - Support consumers that want to commit only a portion of the given messages. - Support transactional consumer/producer. - Support external commits. This branch is based on the work of abandoned PRs #788 and #830 and builds on preparatory work in PRs #744, #1068, #1073 #1086, #1089 and #1097.
erikvanoosten
added a commit
to zio/zio-kafka
that referenced
this issue
Nov 16, 2023
Fixes #590 "Many records duplicately processed after rebalancing" In this change we introduce a new experimental mode that holds up a rebalance until all messages that were provided to the stream of a revoked partition, have been committed. ### Motivation Here is a common (single partition) scenario around rebalances: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100) 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance 1. the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets 1. _at the same time,_ another consumer on another instance, starts consuming from the last committed offset (which is 50) and will process the same messages with offsets 50 to 100 Messages with offsets 50 to 100 are being processed by both consumers simultaneously. Note that both consumers will try to commit these offsets. Until the first consumer is ready, the stored offsets can go up and down and are therefore unreliable. After merging this change, the scenario will unfold as follows: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100). Zio-kafka keeps track of the highest provided offset 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance * the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets * inside the onRevoked callback, zio-kafka continues to process commit commands from the user * zio-kafka continues to do so until the commit with the highest provided offset (offset 100) completes * the onRevoked callback completes, signalling to Kafka that the next consumer may start consuming from the partition 1. another consumer on another instance, starts consuming from the last committed offset (which is now 100, problem solved!) ### Commit queue Because both the main runloop, and the rebalance listener need to process (and thus receive) commits commands, the commit commands were moved to a separate queue. Because the main runloop may still need to be kickstarted when it is no longer polling, a new command `CommitAvailable` was introduced. ### Complications 1. The chosen solution is not suitable for all consumers. - There are use cases where not all messages are read from the stream. For example, some want to read exactly 100 messages from a topic and then stop consuming. In that case the user has no intention to commit all messages, and therefore we should not wait for that to happen. Since stream consumers can basically do whatever they want, the only way we can support such use cases is by letting the consumer tell zio-kafka that they are done with committing. This requires an API change. For example, we can let the user tell zio-kafka that a given commit is the last one. - Not all consumers commit offsets (to Kafka) in the first place. In a future change we could make it work for commits to other stores though. As a workaround, these users can commit to both places. 1. It requires Kafka client 3.6.0. In earlier versions there was no way to wait for async commits to complete. ### Same thread executor The Kafka client requires that any nested invocations (that is, from the rebalance listener callback) to the java consumer happens from the same thread. This is very much at odds with how ZIO works. Attempts to convince the Kafka committers to relax this requirement failed; they could not be convinced that this is a problem. This is circumvented by using a special same-thread-runtime which runs on the thread of the caller. However, some operations such as `ZIO.timeout` and anything with `Schedules` will still shift work to another thread. We work around this by using blocking time. ### Experimental Because holding up the rebalance may have unforeseen consequences, this feature is marked as experimental. This allows us to collect experiences before we recommend this mode to all users. ### Collateral This change also: - fixes order of `private` and `final` - removes some completely useless tests ### Related The same issue is present in: - f2s-kafka: fd4s/fs2-kafka#1200 - alpakka-kafka: akka/alpakka-kafka#1038 In fact, every program that does polls and commits asynchronously is likely affected. ### Non-goals This change does not try to solve the following goals. However, these can be addressed in future PRs. - Awaiting commits after stopping the consumer, e.g. due to program shutdown (see #1087). - Support consumers that want to commit only a portion of the given messages. - Support transactional consumer/producer. - Support external commits. This branch is based on the work of abandoned PRs #788 and #830 and builds on preparatory work in PRs #744, #1068, #1073 #1086, #1089 and #1097.
erikvanoosten
added a commit
to zio/zio-kafka
that referenced
this issue
Nov 18, 2023
Fixes #590 "Many records duplicately processed after rebalancing" In this change we introduce a new experimental mode that holds up a rebalance until all messages that were provided to the stream of a revoked partition, have been committed. ### Motivation Here is a common (single partition) scenario around rebalances: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100) 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance 1. the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets 1. _at the same time,_ another consumer on another instance, starts consuming from the last committed offset (which is 50) and will process the same messages with offsets 50 to 100 Messages with offsets 50 to 100 are being processed by both consumers simultaneously. Note that both consumers will try to commit these offsets. Until the first consumer is ready, the stored offsets can go up and down and are therefore unreliable. After merging this change, the scenario will unfold as follows: 1. a consumer polls some messages and puts them in the streams (let's say messages with offsets 0 to 100). Zio-kafka keeps track of the highest provided offset 1. asynchronously, the user processes these messages. Some of them are committed (let's say up to offset 50), the rest is still being processed when... 1. a rebalance happens, the partition is revoked and assigned to another consumer on another instance * the consumer continues to process the remaining messages with offsets 50 to 100, and tries to commit those offsets * inside the onRevoked callback, zio-kafka continues to process commit commands from the user * zio-kafka continues to do so until the commit with the highest provided offset (offset 100) completes * the onRevoked callback completes, signalling to Kafka that the next consumer may start consuming from the partition 1. another consumer on another instance, starts consuming from the last committed offset (which is now 100, problem solved!) ### Commit queue Because both the main runloop, and the rebalance listener need to process (and thus receive) commits commands, the commit commands were moved to a separate queue. Because the main runloop may still need to be kickstarted when it is no longer polling, a new command `CommitAvailable` was introduced. ### Complications 1. The chosen solution is not suitable for all consumers. - There are use cases where not all messages are read from the stream. For example, some want to read exactly 100 messages from a topic and then stop consuming. In that case the user has no intention to commit all messages, and therefore we should not wait for that to happen. Since stream consumers can basically do whatever they want, the only way we can support such use cases is by letting the consumer tell zio-kafka that they are done with committing. This requires an API change. For example, we can let the user tell zio-kafka that a given commit is the last one. - Not all consumers commit offsets (to Kafka) in the first place. In a future change we could make it work for commits to other stores though. As a workaround, these users can commit to both places. 1. It requires Kafka client 3.6.0. In earlier versions there was no way to wait for async commits to complete. ### Same thread executor The Kafka client requires that any nested invocations (that is, from the rebalance listener callback) to the java consumer happens from the same thread. This is very much at odds with how ZIO works. Attempts to convince the Kafka committers to relax this requirement failed; they could not be convinced that this is a problem. This is circumvented by using a special same-thread-runtime which runs on the thread of the caller. However, some operations such as `ZIO.timeout` and anything with `Schedules` will still shift work to another thread. We work around this by using blocking time. ### Experimental Because holding up the rebalance may have unforeseen consequences, this feature is marked as experimental. This allows us to collect experiences before we recommend this mode to all users. ### Collateral This change also: - fixes order of `private` and `final` - removes some completely useless tests ### Related The same issue is present in: - f2s-kafka: fd4s/fs2-kafka#1200 - alpakka-kafka: akka/alpakka-kafka#1038 In fact, every program that does polls and commits asynchronously is likely affected. ### Non-goals This change does not try to solve the following goals. However, these can be addressed in future PRs. - Awaiting commits after stopping the consumer, e.g. due to program shutdown (see #1087). - Support consumers that want to commit only a portion of the given messages. - Support transactional consumer/producer. - Support external commits. This branch is based on the work of abandoned PRs #788 and #830 and builds on preparatory work in PRs #744, #1068, #1073 #1086, #1089 and #1097.
|
We had very similar issue at my workplace that old consumer kept consuming after being revoked. This property of always being guaranteed that only one consumer is consuming a partition is very important to my use case (enabled me to store state in memory and avoid conflicts, very useful for event sourcing). In my use case I do not really care if message will be processed multiple times, as long as only one consumer is doing it at a time. One thing ZIO kafka does differently I think is that it stores promise notifying the "actor" that partition stream finished operating, that might be a change needed here to allow the semantics described above (but it all hangs on the assumption that new assignment is blocked by revoke callback) |
wookievx
added a commit
to wookievx/fs2-kafka
that referenced
this issue
Aug 4, 2024
This is potential solution for fd4s#1200. Allowing for opting-in for graceful revoke handling (waiting for all the streams to finish, which should imply all the commits as well)
wookievx
added a commit
to wookievx/fs2-kafka
that referenced
this issue
Sep 29, 2024
This is potential solution for fd4s#1200. Allowing for opting-in for graceful revoke handling (waiting for all the streams to finish, which should imply all the commits as well)
wookievx
added a commit
to wookievx/fs2-kafka
that referenced
this issue
Nov 1, 2024
This is potential solution for fd4s#1200. Allowing for opting-in for graceful revoke handling (waiting for all the streams to finish, which should imply all the commits as well)
wookievx
added a commit
to wookievx/fs2-kafka
that referenced
this issue
Nov 13, 2024
This is potential solution for fd4s#1200. Allowing for opting-in for graceful revoke handling (waiting for all the streams to finish, which should imply all the commits as well)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
With fs2-kafka
3.0.1and also on2.x.x, we faced with the issue where multiple consumers that share the same group id start reading from the same partition after rebalance.The flow looks as follows:
We have already checked the broker and found no issue. Especially the logs confirm that the broker issues the right events. There might be something with the
ConsumerRebalanceListenerimplementation in the library.The text was updated successfully, but these errors were encountered: