[
https://issues.apache.org/jira/browse/HDFS-3701?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13420689#comment-13420689
]
nkeywal commented on HDFS-3701:
-------------------------------
Hi Uma,
Thank you very much for the feedback. In 1.0.3 ReplicaNotFoundException is not
there for sure. If it's not there on the branch-1, and if the backport is not
planned, a possible workaround would be to at least try on all the datanodes
available, and if they all fail with a socket exception rethrow an exception
instead of continuing. For other cases, we would return 0.
That would not solve all cases, but would decrease the probability of
occurence. And this should not bring false positive, the only case that would
work today but not after this fix is the case with an empty last block and all
datanodes dead. In this case, we would throw an error while previously the
DFSClient would have returned O.
> HDFS may miss the final block when reading a file opened for writing if one
> of the datanode is dead
> ---------------------------------------------------------------------------------------------------
>
> Key: HDFS-3701
> URL: https://issues.apache.org/jira/browse/HDFS-3701
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs client
> Affects Versions: 1.0.3
> Reporter: nkeywal
> Priority: Critical
>
> When the file is opened for writing, the DFSClient calls one of the datanode
> owning the last block to get its size. If this datanode is dead, the socket
> exception is shallowed and the size of this last block is equals to zero.
> This seems to be fixed on trunk, but I didn't find a related Jira. On 1.0.3,
> it's not fixed. It's on the same area as HDFS-1950 or HDFS-3222.
--
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
