John Spray

I think we can switch this one back on https://github.com/redpanda-data/redpanda/pull/7089

CC @VladLazar If we move feature_table up to the top level then it also becomes practical to early-populate it from a kvstore snapshot before controller construction.

Sorry, that was ambiguously worded -- it's the the "early" part that is enabled by moving it out of the controller, rather than the access to kvstore.

I concur.

This looks the same as https://github.com/redpanda-data/redpanda/issues/6601 Was fixed by https://github.com/redpanda-data/redpanda/pull/6794

This hasn't reoccurred in the last 30 days. We have tracking elsewhere for limiting reader concurrency: - https://github.com/redpanda-data/redpanda/issues/6024 - https://github.com/redpanda-data/redpanda/issues/6023 - https://github.com/redpanda-data/redpanda/pull/6112 So I think we can safely drop this...

This remains a rather rare failure. last 30 days: OFAIL test: ConsumerOffsetsMigrationTest.test_migrating_consume_offsets.failures=True.cpus=1 (2/769 runs) failure at 2022-10-27T07:24:08.891Z: TimeoutError("Consumer failed to consume up to offsets ... in job https://buildkite.com/redpanda/redpanda/builds/17411#018417da-00d8-435b-95c1-c905f841eec1 failure at...

This test will go away in https://github.com/redpanda-data/redpanda/pull/7507

I did this recently.

This ticket is quite old, and described an earlier migration approach with downtime, that already exists. Closing in favour of https://github.com/neondatabase/neon/issues/5199 for the ongoing work to enable seamless migration.