cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marcus Eriksson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-5036) Wrong description of 'setstreamthroughput' option
Date Sun, 24 Mar 2013 06:29:18 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-5036?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13612022#comment-13612022
] 

Marcus Eriksson commented on CASSANDRA-5036:
--------------------------------------------

I agree it might be a bit too high

defaulting to 200 sounds good
                
> Wrong description of 'setstreamthroughput' option
> -------------------------------------------------
>
>                 Key: CASSANDRA-5036
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5036
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Config, Documentation & website
>    Affects Versions: 1.1.6
>         Environment: Cassandra 1.1.6 (DataStax distribution)
>            Reporter: Alexey Zotov
>            Assignee: Marcus Eriksson
>            Priority: Trivial
>
> There is a typo in description of 'setstreamthroughput' option. It is measured in megabits
per second. Page with wrong description:
> http://www.datastax.com/docs/1.1/references/nodetool#nodetool-setstreamthroughput
> Page with right description:
> http://www.datastax.com/docs/1.1/configuration/node_configuration#stream-throughput-outbound-megabits-per-sec
> Also I want to discuss possibility to reduce default value for this option. I think that
400 Mbs is too high for common cases. 
> Preface:
> This option is used only in case streams. There are two cases when streams are actual.
They are rebuilding of a node and repair process. Let's skip first case and will talk only
about the second. Let's imagine that we have replication factor 3.
> Cross-datacenter connectivity case: 
> When we start repair process it will borrow all network channel. Let's do some calculations.
You start repair on an one node, e.g. 5 node (3 remote and 2 local) should send us some data.
Note that 3 of them are from remote datacenter. So 400 * 3 = 1,2 Gbs we should receive through
WAN. I'm sure that it's too high.
> I suggest to make it 2 times less.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message