cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeff Jirsa (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7306) Support "edge dcs" with more flexible gossip
Date Thu, 14 Jan 2016 19:51:40 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7306?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15098739#comment-15098739
] 

Jeff Jirsa commented on CASSANDRA-7306:
---------------------------------------

[~jbellis] - that was one of my open questions where I asked [~tupshin] for clarification.

As implemented (the way I would expect it to work as an operator), if a DC is not gossipable,
it's not available for requests. For that reason, we remove it from the Map of datacenters
in NTS, and the non-gossipable replicas will not receive the mutation, and will not be counted
towards CL calculations. 

https://github.com/jeffjirsa/cassandra/blob/cassandra-7306/src/java/org/apache/cassandra/locator/NetworkTopologyStrategy.java#L69-L72

If the DC topology provider changes (adds/removes DCs from its white/blacklist), it will re-instantiate
the replication strategy to propagate those changes:

https://github.com/jeffjirsa/cassandra/blob/cassandra-7306/src/java/org/apache/cassandra/db/Keyspace.java#L332-L339



> Support "edge dcs" with more flexible gossip
> --------------------------------------------
>
>                 Key: CASSANDRA-7306
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7306
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Distributed Metadata
>            Reporter: Tupshin Harper
>            Assignee: Jeff Jirsa
>              Labels: ponies
>
> As Cassandra clusters get bigger and bigger, and their topology becomes more complex,
there is more and more need for a notion of "hub" and "spoke" datacenters.
> One of the big obstacles to supporting hundreds (or thousands) of remote dcs, is the
assumption that all dcs need to talk to each other (and be connected all the time).
> This ticket is a vague placeholder with the goals of achieving:
> 1) better behavioral support for occasionally disconnected datacenters
> 2) explicit support for custom dc to dc routing. A simple approach would be an optional
per-dc annotation of which other DCs that DC could gossip with.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message