[CASSANDRA-18816] Add support for repair coordinator to retry messages that timeout

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 5.0-alpha2, 5.0
• Component/s: Consistency/Repair
• Labels: pull-request-available
• Change Category: Operability
• Complexity: Normal
• Platform: All

Source Control Link

https://github.com/apache/cassandra/commit/164fa7949d1437528363465d14ed20898f1daa9e

Test and Documentation Plan

simulation testing

• Assignee: David Capwell
• Reporter: David Capwell

Authors

David Capwell

Reviewers

Caleb Rackliffe, Marcus Eriksson

• Created: 01/Sep/23 18:44
• Updated: 01/Nov/23 08:54
• Resolved: 27/Sep/23 23:23

Description

Now that CASSANDRA-15399 is in, most of the repair messages have a state that they can check against to make message delivery idempotent, allowing the coordinator to retry such messages; a few of the most critical messages to retry are: PREPARE_MSG, VALIDATION_REQ, VALIDATION_RSP, SYNC_REQ, and SYNC_RSP.

With this I propose making the coordinator able to retry these key messages to try and make repair more resilient to ephemeral issues.

Issue Links

• causes: CASSANDRA-18890 Test failure: org.apache.cassandra.repair.ConcurrentIrWithPreviewFuzzTest
• links to: CI: 5.0
• links to: CI: trunk
• links to: GH: 5.0
• links to: GH: trunk
• links to: GitHub Pull Request #68
• links to: GitHub Pull Request #2660
• links to: GitHub Pull Request #2732