ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-810654638
Some unrelated failures:
`streams.integration.EosIntegrationTest.shouldNotViolateEosIfOneTaskFailsWithState[exactly_once_beta]`
`kafka.clients.admin.KafkaAdminClientT
ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-804392576
Thanks @guozhangwang !
> If users mistakenly use the same stat dir, is that undefined behavior or
do we fail fast in a deterministic way? If it's the former do we ha
ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-803050846
One unrelated test failure:
`kafka.network.ConnectionQuotasTest.testListenerConnectionRateLimitWhenActualRateAboveLimit`
Merging to trunk, I'll eventually cherrypick
ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-802542776
@wcarlson5 I ended up putting a minor guard against orphaned task
directories into this PR. It's not perfect but basically we just let the
cleaner thread verify that any lo
ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-802350667
@wcarlson5 that's a good question, and I totally agree we should be more
careful about that. In theory we should always clean up all state including
unlock task directories
ableegoldman commented on pull request #10342:
URL: https://github.com/apache/kafka/pull/10342#issuecomment-801499667
Call for review any of @cadonna @wcarlson5 @lct45 @guozhangwang @rodesai
@swist
This is an automated mes
