ducktape: fix bad test case with acks=1 #26878
Merged
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
wdberkeley
force-pushed
the
too-relaxed-acks
branch
from
106093d to
01b5c96
Compare
CI test resultstest results on build#69205
test results on build#69384
|
| @@ -88,7 +88,7 @@ def test_consumer_interruption(self): | |||
| @parametrize(write_caching=True, disable_batch_cache=False) | |||
| @parametrize(write_caching=True, disable_batch_cache=True) | |||
| @parametrize(write_caching=False) | |||
| def test_relaxed_acks(self, write_caching, disable_batch_cache=False): | |||
| def test_acks(self, write_caching, disable_batch_cache=False): | |||
There was a problem hiding this comment.
can we keep the test but just change the validation, it is important to test redpanda with relaxed acks, with this change we are loosing coverage
There was a problem hiding this comment.
Yes, I can do that. Is it important to set write_caching=false with relaxed acks, as this test does? Should we test relaxed acks with write caching on and off?
`test_relaxed_acks` had one parameterization running with `acks=1`, but it expected that the consumer would see exactly the set of offsets that were acked to the producers, even with injected node failures. This is not always the case: 1. If the leader acks a message but doesn't have time to replicate it before failing, and a new leader is elected while the original leader is down, the original leader will truncate the acked message from its log and no consumer will see it. 2. If the leader acks a message but doesn't have time to replicate or flush it to disk before failing, and the leader is reelected when it restarts, it may lose the message and no consumer will see it. Situation 2 was recently seen in CI. This change removes the `acks=1` case and renames the test, because it's no longer testing relaxed acks, just default `acks=-1`, with different levels of caching enabled.
wdberkeley
force-pushed
the
too-relaxed-acks
branch
from
01b5c96 to
0e8ec1d
Compare
|
Force-push to address Michal's feedback: keep the test case, just relax the validation when |
mmaslankaprv
approved these changes
Jul 22, 2025
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.
test_relaxed_ackshad one parameterization running withacks=1, but it expected that the consumer would see exactly the set of offsets that were acked to the producers, even with injected node failures. This is not always the case:Situation 2 was recently seen in CI.
This change adjust validation to allow for messages to go missing when
acks=1.Fixes CORE-12793
Backports Required
Release Notes