[CASSANDRA-15192] Why a added new node stop an old node to start?

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Information Provided
• Component/s: Cluster/Membership
• Platform: All
• Assignee: Unassigned
• Reporter: gloCalHelp.com
• Created: 02/Jul/19 08:50
• Updated: 01/Aug/19 21:53
• Resolved: 01/Aug/19 21:53

Description

I have a running 3 datanodes with cassandra3.11.3 with a keyspace having a table over 10millions row in Centos6.9, after I added a node in win10 with cassandra3.11.4 for a whole night, this afternoon, one of the old 3 nodes can not start, and I use "nodetool decommission" on the new node and "nodetool removenode (new node's uuid)", but one of the 3 old node can not start with an no error system.log and debug.log as the attachment.

     Finally, I setting the failing old node's start JVM_OPTS to have -Dcassandra.reset_bootstrap_progress=true and it become live.StrangeNoErrorStopLogs.zip