[CASSANDRA-18764] CEP-15 (C*) When a host replacement happens don't loose the peer mapping right away

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 5.x
• Component/s: Accord
• Bug Category: Correctness - Unrecoverable Corruption / Loss
• Severity: Critical
• Complexity: Low Hanging Fruit
• Discovered By: Unit Test
• Platform: All
• Since Version: 5.0

Source Control Link

https://github.com/apache/cassandra-accord/commit/2ad55e03c43ce074cdf5e36cfa14cb4278c2dc0f

Test and Documentation Plan

existing tests

• Assignee: David Capwell
• Reporter: David Capwell

Authors

David Capwell

Reviewers

Blake Eggleston

• Created: 15/Aug/23 23:46
• Updated: 16/Aug/23 21:17
• Resolved: 16/Aug/23 21:17

Description

When a host replacement happens the latest epoch will no longer have the removed peer, but we may have in-flight messages that still reference it. When C* restarts we need to also filter out these peers so we don’t wait for an epoch ack.

Issue Links

• links to: Accord CI: trunk
• links to: Accord PR: trunk