Hi all, I have done some work on the following Jira ticket:'Rewrite HighAvailabilityTaskAssignor to implement the new TaskAssignor interface'.
I've created a proof of concept on a local branch (happy to raise a WIP PR if that would help) and have added some thoughts on my approach in the linked Jira ticket. I would like to request feedback on the approach if possible; this is due to this assignor currently being reliant on the StandbyTaskAssignor, which currently uses ClientState objects, and KafkaStreamsState objects for Rack Optimisation. I realise my current approach is unlikely to be correct given the brittle nature of the data synchronisation I'm doing, but was unsure of carrying out large changes outside of the new HighAvailabilityTaskAssignor class. https://issues.apache.org/jira/browse/KAFKA-16869 Thanks, Lorcan
