urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1305253651
@artemlivshits The scenario you mentioned is already covered, even without
this change - when a transaction times out, the transaction coordinator bumps
the epoch, so it already fences o
urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1295780702
@hachikuji Thanks for the feedback.
Yes, the essence of the change is bumping the epoch - but only in case of
timeouts - so delivery timeout or request timeout.
Overall I agree
urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1261958861
@jolshan It is an idea, the first version of the PR was trying to implement
that, but the current state of the PR is based on the fatal state.
The idea about keeping the producer i
urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1259383078
@dajac @hachikuji Any chance you can take a look at this? This is a painful
issue in transactional producers, with some serious consequences (partition
corruption).
--
This is an auto
urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1229897726
@artemlivshits @ijuma @hachikuji Can you please take a look at this PR?
Trying to fix a bug in the transactional producer. Thanks in advance!
--
This is an automated message from the A
urbandan commented on PR #12392:
URL: https://github.com/apache/kafka/pull/12392#issuecomment-1190135063
@showuon @artemlivshits Can you please take a look at this PR? This is the
issue we had a thread about on the dev list.
--
This is an automated message from the Apache Git Service.
To