[CASSANDRA-12084] Host comes up during replacement when all replicas are down

• Type: Bug
• Status: Resolved
• Priority: Low
• Resolution: Duplicate
• Severity: Low
• Assignee: Unassigned
• Reporter: Sankalp Kohli
• Created: 23/Jun/16 21:01
• Updated: 16/Apr/19 09:30
• Resolved: 08/Jul/16 00:52

Description

Description: Steps to reproduce
1) Setup a 3 instance cluster
2) Create a Rf=3 keyspace
3) Bring non seed instance down.
4) Start 4th instance with -Dcassandra.replace_address args set.
5) When it is sleeping after “JOINING: calculation complete, ready to bootstrap”, kill the other two instances
6) This node comes up with no data.

This is happening due to code in RangeStreamer.getRangeFetchMap where it does not throw an exception if localhost is part of the sources.

Issue Links

• duplicates: CASSANDRA-11848 replace address can "succeed" without actually streaming anything