[
https://issues.apache.org/jira/browse/HDFS-13779?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16583130#comment-16583130
]
Chen Liang commented on HDFS-13779:
-----------------------------------
Thanks for the correction [~xkrogen]! Makes sense to me. I recall one example
case I was thinking of was say there are 0, 1, 2 proxies, and currentIndex = 2.
If 1 gets removed and the list shrinks to 0, 2. Due to modulo on size, seems
currentIndex would effectively point to 2%size (2) = 0, but it was meant to
point to 2. I guess the whole point I was concerning was that since
{{currentObservers}} is based on modulo of size, but size can be changing at
any point of time, {{currentIndex}} can become not so current over time. But
seems even if {{currentIndex}} is wrong, it should be eventually fixed by the
for loop, and thanks for point out that this should only happen in failover
actions. I agree this should not be an issue.
> Implement performFailover logic for ObserverReadProxyProvider.
> --------------------------------------------------------------
>
> Key: HDFS-13779
> URL: https://issues.apache.org/jira/browse/HDFS-13779
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: hdfs-client
> Reporter: Konstantin Shvachko
> Assignee: Erik Krogen
> Priority: Major
> Attachments: HDFS-13779-HDFS-12943.WIP00.patch
>
>
> Currently {{ObserverReadProxyProvider}} inherits {{performFailover()}} method
> from {{ConfiguredFailoverProxyProvider}}, which simply increments the index
> and switches over to another NameNode. The logic for ORPP should be smart
> enough to choose another observer, otherwise it can switch to a SBN, where
> reads are disallowed, or to an ANN, which defeats the purpose of reads from
> standby.
> This was discussed in HDFS-12976.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
