[jira] [Comment Edited] (CASSANDRA-6551) Rack-aware batchlog replication

["Aleksey Yeschenko (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-6551?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13988870#comment-13988870 ] 

Aleksey Yeschenko edited comment on CASSANDRA-6551 at 5/4/14 12:38 AM:
-----------------------------------------------------------------------

Ninja-d in https://github.com/apache/cassandra/commit/f486193ab5654881986a7f7b655fee3aedabad15
- moved BatchlogEndpointSelector to BatchlogManager.EndpointFilter
- moved the single-node DC special-case logic from StorageProxy to EndpointFilter
- moved the endpoints argument of chooseEndpoints() to the EndpointFilter constuctor + renamed chooseEndpoints() to filter()

Left everything else alone.


was (Author: iamaleksey):
Ninja-d in f486193ab5654881986a7f7b655fee3aedabad15
- moved BatchlogEndpointSelector to BatchlogManager.EndpointFilter
- moved the single-node DC special-case logic from StorageProxy to EndpointFilter
- moved the endpoints argument of chooseEndpoints() to the EndpointFilter constuctor + renamed chooseEndpoints() to filter()

Left everything else alone.

> Rack-aware batchlog replication
> -------------------------------
>
>                 Key: CASSANDRA-6551
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6551
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Rick Branson
>            Assignee: Mikhail Stepura
>            Priority: Minor
>             Fix For: 2.1 rc1
>
>         Attachments: 6551-v2.txt, 6551-v3.patch, cassandra-2.0-6551-2.patch
>
>
> Right now the batchlog replication code just randomly picks 2 other nodes in the same DC, regardless of rack. Ideally we'd pick 2 replicas in other racks to achieve higher fault tolerance.



--
This message was sent by Atlassian JIRA
(v6.2#6252)