[
https://issues.apache.org/jira/browse/CASSANDRA-4589?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13445260#comment-13445260
]
Vijay commented on CASSANDRA-4589:
----------------------------------
{quote}
that's how you add them for new DCs, with an rf of zero until you have the
nodes in place
{quote}
if we do this patch then we will start failing reads and writes (because
writeEndpoints will be < actual end points for the whole cluster) when the user
is expanding the clusters to other DC's.
{quote}
CL.ONE write, when there is >= 1 replica available, sounds like correct
behavior to me.
{quote}
+1
> writes are allowed when RF>N
> ----------------------------
>
> Key: CASSANDRA-4589
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4589
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 1.1.2
> Reporter: Brandon Williams
> Assignee: Vijay
> Priority: Minor
> Fix For: 1.2.0
>
>
> Easily repro'd by running stress with a ridiculous rf:
> # tools/bin/cassandra-stress -n 100000 -i 1 -o insert -l123456
> We're supposed to allow creation of a ks where the rf exceeds the amount of
> nodes, but we shouldn't be able to write to it.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
