Andrew Wong

It looks like the errors are being logged while the node is still on v22.1.4. I suspect this is caused by the fact that https://github.com/redpanda-data/redpanda/pull/5097 seems to only live only...

Saw another instance of this ``` FAIL test: PartitionBalancerTest.test_full_nodes (1/24 runs) failure at 2022-08-16T16:42:59.830Z: AssertionError() in job https://buildkite.com/redpanda/redpanda/builds/14220#0182a71a-48cc-4534-8c04-7d2bea40908c ``` Stack trace: ``` ==================================================================================================== test_id: rptest.tests.partition_balancer_test.PartitionBalancerTest.test_full_nodes status: FAIL run time: 1...

Another instance https://ci-artifacts.dev.vectorized.cloud/redpanda/0182ae3b-de90-4596-bfb2-bd4ad36a228c/vbuild/ducktape/results/2022-08-18--001/report.html

Regarding the `cannot find consensus group` error logs, I looked into it a bit and I think it's worth reducing severity (at least in some cases). See https://github.com/redpanda-data/redpanda/pull/5742 for more...

Seen again here FAIL test: HighThroughputPartitionMovementTest.test_interrupting_partition_movement_under_load.replication_factor=3 (1/1 runs) failure at 2022-08-17T07:38:31.224Z: in job https://buildkite.com/redpanda/vtools/builds/3271#0182a9c3-a7aa-4850-a6b7-65bee8152d80

CI failure is https://github.com/redpanda-data/redpanda/issues/5858

> I have a question isn't the assertion in `redpanda.py:394` incorrect, it seems to me that even if node is stopped we should be able to access its address ?...

I updated the ones that seem like they should work. We do lose out on some developer-error-catching if we are unintentionally calling these methods with stopped nodes, but I think...

@dotnwat Thanks for the look! Fixed

Thanks for the reminder, been retrying on what appear to be unrelated issues, but wanted at least a clean run since this touches some widely used methods: - 8/18: https://github.com/redpanda-data/redpanda/issues/5884...