cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeremy Hanna (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11488) Bug or not?: coordinator using SimpleSnitch may query other nodes for copies of local data
Date Wed, 06 Apr 2016 17:00:30 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11488?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15228643#comment-15228643
] 

Jeremy Hanna commented on CASSANDRA-11488:
------------------------------------------

Giving this some more though, I think we need to document the distinction.  I didn't realize
that the dynamic snitch would route away from local nodes when using TokenAware load balancing
for example.  It does make sense if that local node is really behind on LCS or has some heavier
operation going on.  I just don't think people expect TokenAware would get routed to another
replica necessarily.

In any case I agree that documentation is probably the best outcome for now unless there is
data that shows that this is insufficient.

> Bug or not?: coordinator using SimpleSnitch may query other nodes for copies of local
data 
> -------------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-11488
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11488
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Coordination
>            Reporter: Jim Witschey
>            Assignee: Stefania
>            Priority: Minor
>              Labels: doc-impacting
>
> As [~Stefania] explains [in this JIRA comment|https://issues.apache.org/jira/browse/CASSANDRA-11225?focusedCommentId=15221059&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15221059],
{{SimpleSnitch}} does not implement {{IEndpointSnitch.sortByProximity(localhost, liveendpoints)}},
so a query for data on the coordinator may query other nodes. That seems like unnecessary
work to me, and on that note, Stefania woonders [in this JIRA comment|https://issues.apache.org/jira/browse/CASSANDRA-11225?focusedCommentId=15223598&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15223598]
- should this be considered a bug?
> Stefania, I'm assigning you here -- could you find the right people to involve in this
discussion?



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

Mime
View raw message