[
https://issues.apache.org/jira/browse/CASSANDRA-16278?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17281954#comment-17281954
]
Blake Eggleston commented on CASSANDRA-16278:
---------------------------------------------
I guess that's true... the case I'm trying to address is where a host
replacement fails, and then attempts to replace the replacement fail. Although
in that case, I suppose you should just disregard the failed replacement and
try to replace the orignal host again?
> Bootstrapping nodes can become unreplaceable
> --------------------------------------------
>
> Key: CASSANDRA-16278
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16278
> Project: Cassandra
> Issue Type: Bug
> Components: Cluster/Gossip
> Reporter: Blake Eggleston
> Assignee: Blake Eggleston
> Priority: Normal
>
> Bootstrapping nodes (including replacements) don't gossip their tokens until
> after they've received their schema. If the node experiences a failure before
> gossiping tokens, it causes the same problem CASSNDRA-15335 fixed for normal
> nodes - they cannot be replaced without gossiped tokens. Including tokens in
> the initial gossip round prevents nodes from getting into this state,
> although we still have to wait for the schema when the node is configured to
> allocate tokens for a keyspace.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
