ableegoldman opened a new pull request #8996: URL: https://github.com/apache/kafka/pull/8996
We have this asymmetry in how the ProcessorStateManager handles in-memory stores: we explicitly skip over them when writing offsets to the checkpoint file, but don't do the same when reading from the checkpoint file to initialize offsets. With eos, this is taken to mean that the state is dirty, and thus we mistakenly mark the entire task as corrupted. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
