[
https://issues.apache.org/jira/browse/HDFS-4849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13685826#comment-13685826
]
Suresh Srinivas commented on HDFS-4849:
---------------------------------------
Before commenting on the detailed issues, here are some highlevel questions.
# How does one guarantee the client name is unique? One could think of a
scenario, albeit contrived, where this can result in issues. Assume client 1
and client 2 end up having same client name. Client 1 could create a file and
start writing to block 1. Client 2 can also now create a file and start writing
to block 2 (given we allocate block 2, even though block 1 is in pending
state). This could cause corruption right? This could also happen for
multithreaded clients.
# Should an audit event be logged for retried requests? How does one know this
audit event is for a retried call vs possible bug that might have logged an
audit event twice?
> Idempotent create and append operations.
> ----------------------------------------
>
> Key: HDFS-4849
> URL: https://issues.apache.org/jira/browse/HDFS-4849
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: namenode
> Affects Versions: 2.0.4-alpha
> Reporter: Konstantin Shvachko
> Assignee: Konstantin Shvachko
> Priority: Blocker
> Attachments: idempotentCreate.patch, idempotentCreate.patch
>
>
> create, append and delete operations can be made idempotent. This will reduce
> chances for a job or other app failures when NN fails over.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
