Jakub Smolar

@guicassolato I am still seeing this in RC4 

Scylla version: `6.3.0~dev-20250115.72f28ce81e17` with build-id `d583b7e768996945fc24bf7f697f5ec41306253a` Kernel Version: `6.8.0-1020-gcp` ## Issue description Reproduced again. Cassandra stress test tries to connect to decommissioned nodes. - node 10.142.0.40 - terminated at 02:24:06...

Reproduced in 2025.1.2 upgrade test: https://argus.scylladb.com/tests/scylla-cluster-tests/4f6c04b3-1b78-416c-a3f4-2c0dbb574e55

Replicated in https://argus.scylladb.com/tests/scylla-cluster-tests/2c56d937-4ad2-4fd3-9079-85863cc9a771 Also led to CoreDump, like in https://github.com/scylladb/scylladb/issues/24201 ``` !INFO | scylla[16660]: [shard 3:sl:d] database - Truncating truncate_ks.truncate_table2 with auto-snapshot !INFO | scylla[10965]: [shard 3:stmt] database - Truncating...

Reproduced on 2024.1.4 https://argus.scylladb.com/tests/scylla-cluster-tests/019b7eee-e5c9-4be0-ac2c-8f36e6357ee0 I understand that this was not backported at the time of 2024.1.4 build

I’ve created a working POC. Setup: - 10 [PyKMIP](https://github.com/OpenKMIP/PyKMIP) instances behind HAProxy acting as a load balancer, with a PostgreSQL database running on the loader node. - The entire setup...

@david-martin yes. The hack folder is ideal for this kind of script for now

Replicated during `disrupt_add_remove_mv` nemesis: https://argus.scylladb.com/tests/scylla-cluster-tests/1c3c11c1-6316-4521-ac29-633002e3b6ab The node is not dead. Errors occur during MV creation

Replicated again https://argus.scylladb.com/tests/scylla-cluster-tests/2fb486a8-ff99-4c17-94ee-d894c37c4a8f

Is this a duplicate of https://github.com/scylladb/scylladb/issues/23456 (or vice versa)?