cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "T Jake Luciani (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-7063) Raise the streaming phi threshold check in 1.2
Date Mon, 21 Apr 2014 17:19:14 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13975725#comment-13975725
] 

T Jake Luciani edited comment on CASSANDRA-7063 at 4/21/14 5:17 PM:
--------------------------------------------------------------------

You could argue the whole phi*2 is a hack :)

This is for 1.2 branch.  There is no simple/safe way to refactor streaming in 1.2

3569 looks like it would do the same if we removed this check altogether.


was (Author: tjake):
This is for 1.2 branch.  There is no simple/safe way to refactor streaming in 1.2

> Raise the streaming phi threshold check in 1.2
> ----------------------------------------------
>
>                 Key: CASSANDRA-7063
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7063
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: T Jake Luciani
>            Assignee: T Jake Luciani
>            Priority: Minor
>             Fix For: 1.2.17
>
>
> In our production cluster we had quite a problem adding new nodes to the cluster.  One
of the main issues was dealing with spikes in load during the production day causing streaming
of data to fail.
> Changing this value has helped us deal with this: 
> https://github.com/apache/cassandra/blob/cassandra-1.2/src/java/org/apache/cassandra/streaming/AbstractStreamSession.java#L105
> We use 200 vs 2, I think it would be generally useful and low risk change for people
using 1.2



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

Mime
View raw message