[CASSANDRA-16146] Node state incorrectly set to NORMAL after nodetool disablegossip and enablegossip during bootstrap

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 4.0, 3.0.x, 3.11.x, 4.0-beta3
• Component/s: Cluster/Gossip
• Bug Category: Correctness - Consistency
• Severity: Low
• Complexity: Low Hanging Fruit
• Discovered By: Code Inspection
• Platform: All
• Since Version: 3.0.0

Source Control Link

https://github.com/apache/cassandra/commit/fee7a10823da1e29bd0e6504fea9679389180c9e

Test and Documentation Plan

ci, jvm dtest

• Assignee: Yifan Cai
• Reporter: Yifan Cai

Authors

Yifan Cai

Reviewers

Blake Eggleston, Brandon Williams

• Created: 28/Sep/20 17:23
• Updated: 03/Jan/21 17:03
• Resolved: 30/Oct/20 19:41

Description

At high level, StorageService#setGossipTokens set the gossip state to NORMAL blindly. Therefore, re-enabling gossip (stop and start gossip) overrides the actual gossip state.
 
It could happen in the below scenario.

1. Bootstrap failed. The gossip state remains in BOOT / JOINING and code execution exits StorageService#initServer.
2. Operator runs nodetool to stop and re-start gossip. The gossip state gets flipped to NORMAL