[
https://issues.apache.org/jira/browse/CASSANDRA-11098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15422197#comment-15422197
]
Boris Simonov commented on CASSANDRA-11098:
-------------------------------------------
We encountered a problem when trying to add new data center to the cluster
(https://docs.datastax.com/en/cassandra/2.2/cassandra/operations/opsAddDCToCluster.html?hl=add,new).
nodetool rebuild <old_datacenter> fails with message:
{code}
$ nodetool rebuild -- SD
nodetool: Unable to find sufficient sources for streaming range
(90751910603388084901079692314015669291,90762843016968867713257115353245790682]
in keyspace system_distributed
$ nodetool ring system_distributed | grep 90762843016968867713257115353245790682
192.168.235.6 r10 Up Normal 915.09 MB 30.70%
$ nodetool flush
$ du -sh system_distributed/
0 system_distributed/
{code}
192.168.235.6 - is in new datacenter and it's empty
I can change them to NetworkTopologyStrategy for this case, but wondering is
these change affect further updates.
> system_distributed and system_traces keyspaces use hard-coded replication
> factors
> ---------------------------------------------------------------------------------
>
> Key: CASSANDRA-11098
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11098
> Project: Cassandra
> Issue Type: Bug
> Reporter: Wei Deng
> Priority: Minor
>
> We introduced system_distributed keyspace in C* 2.2 so that we can save the
> repair histories and ancestors to a system keyspace (due to CASSANDRA-5839).
> However, looks like it's hard-coding the replication factor to 3, according
> to this line:
> https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/repair/SystemDistributedKeyspace.java#L103
> This may confuse some query operations against this keyspace.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
