[
https://issues.apache.org/jira/browse/HDFS-2811?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13196409#comment-13196409
]
Suresh Srinivas commented on HDFS-2811:
---------------------------------------
Using block pool ID which should be opaque to the client (and garbage
in/garbage out) for detecting misconfiguration does not seem correct. The only
reliable identifier a client uses is that of a file URI. If same namespace is
supported on two namenodes with federation (very unlikely) then adding checks
like block pool ID is not fool proof. You might have client failover before
even discovering the block pool ID from the first namenode and hence continues
to use wrong namenode.
> HA: Client should fail if a failover occurs which switches block pool ID
> ------------------------------------------------------------------------
>
> Key: HDFS-2811
> URL: https://issues.apache.org/jira/browse/HDFS-2811
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ha, hdfs client, name-node
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Aaron T. Myers
> Assignee: Brandon Li
>
> Making sure that the client is talking to an NN with the same block pool ID
> as the one it was previously talking to seems like a good sanity check.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
