[
https://issues.apache.org/jira/browse/HDFS-5371?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jing Zhao updated HDFS-5371:
----------------------------
Description: Currently when dfs.client.test.drop.namenode.response.number
is enabled for testing, the client will start failover and try the other NN.
But in most of the testing cases we do not need to trigger the client failover
here since if the drop-response number is >1 the next response received from
the other NN will also be dropped. We can let the client just simply retry the
same NN.
> Let client retry the same NN when
> "dfs.client.test.drop.namenode.response.number" is enabled
> --------------------------------------------------------------------------------------------
>
> Key: HDFS-5371
> URL: https://issues.apache.org/jira/browse/HDFS-5371
> Project: Hadoop HDFS
> Issue Type: Improvement
> Affects Versions: 3.0.0
> Reporter: Jing Zhao
> Assignee: Jing Zhao
> Priority: Minor
> Attachments: HDFS-5371.000.patch
>
>
> Currently when dfs.client.test.drop.namenode.response.number is enabled for
> testing, the client will start failover and try the other NN. But in most of
> the testing cases we do not need to trigger the client failover here since if
> the drop-response number is >1 the next response received from the other NN
> will also be dropped. We can let the client just simply retry the same NN.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
