hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gera Shegalov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-6452) ConfiguredFailoverProxyProvider should randomize currentProxyIndex on initialization
Date Wed, 28 May 2014 01:16:03 GMT

    [ https://issues.apache.org/jira/browse/HDFS-6452?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14010631#comment-14010631
] 

Gera Shegalov commented on HDFS-6452:
-------------------------------------

Hi Aaron, the net improvement should be in the overhead smoothness over time. E.g., we will
smooth the storm of {{StandbyException: Operation category READ is not supported in state
standby}}

[~jingzhao], this is targeted for deployments with automatic failover where the emphasis is
on not trying to watch what NN is active all the time. 

> ConfiguredFailoverProxyProvider should randomize currentProxyIndex on initialization
> ------------------------------------------------------------------------------------
>
>                 Key: HDFS-6452
>                 URL: https://issues.apache.org/jira/browse/HDFS-6452
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: ha, hdfs-client
>    Affects Versions: 2.4.0
>            Reporter: Gera Shegalov
>            Assignee: Gera Shegalov
>
> We observe that the clients iterate proxies in the fixed order. Depending on the order
of namenodes in dfs.ha.namenodes.<nameservice> (e.g. 'nn1,nn2') and the current standby
(nn1), all the clients will hit nn1 first, and then failover to nn2.  Chatting with [~lohit]
we think we can simply select the initial value of {{currentProxyIndex}} randomly, and keep
the logic of {{performFailover}} of iterating from left-to-right. This should halve the unnecessary
load on standby NN.



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

Mime
View raw message