[
https://issues.apache.org/jira/browse/HDFS-5327?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jing Zhao updated HDFS-5327:
----------------------------
Attachment: HDFS-5327.001.patch
A simple patch that only does the rename. Since currently the standby state is
the only use case in the NameNode side to trigger the client's failover, we can
still directly throw a FailoverException in NN instead of wrapping a specific
exception into FailoverException.
> Rename StandbyException to FailoverException
> --------------------------------------------
>
> Key: HDFS-5327
> URL: https://issues.apache.org/jira/browse/HDFS-5327
> Project: Hadoop HDFS
> Issue Type: Improvement
> Affects Versions: 3.0.0
> Reporter: Jing Zhao
> Assignee: Jing Zhao
> Priority: Minor
> Attachments: HDFS-5327.001.patch
>
>
> After HDFS-5291, client's RPC level can determine whether to retry the same
> server or failover to another server based on the two types of exception:
> # The new introduced RetriableException indicating that the server is
> temporarily unavailable and clients can retry.
> # The old StandbyException indicating that clients can failover and try
> another server.
> In this context, since the term "standby" only tells one scenario (active +
> standby), we may want to rename the StandbyException to FailoverException to
> make it more generic.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
