Kamil Braun
Kamil Braun
> The node n4 will learn from the gossip entry for 127.0.0.3 that node 127.0.0.3 is in shutdown status which means 127.0.0.3 is still part of the ring. When n3...
I don't understand why n4's handling of gossip state is different than n1 and n2's in this example. All nodes should handle it in the same way; so how come...
Due to 5ca73019dd5a08e335daa26a868ad45fb59def6d
We suspect this PR will fix it: https://github.com/scylladb/scylladb/pull/11691 It's currently in review.
> The checks for when to go into local mode seem a little too loose for me: I was expecting us looking at the replication strategy, although this would disallow...
Note that this commit unmasks some other problems. `partitioner_tests.py` will fail with this commit because it depended on the fact that we immediately enabled all cluster features while it was...
I sent a change to `partitioner_tests.py` which will cause it to pass even with this fix: https://github.com/scylladb/scylla-dtest/pull/2804
Well, this doesn't seem right, CI should fail. I ran `partitioner_tests.py` locally again and it also passed. It seems that the features are being enabled early, before the nodes gossip...
> The question is why do we even start the upgrade procedure since not all nodes reports that they support raft. Indeed. That's what I was trying to solve. But...