[
https://issues.apache.org/jira/browse/HADOOP-1558?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12518653
]
Alejandro Abdelnur commented on HADOOP-1558:
--------------------------------------------
A comment related 1416.
1558 issue/patch is to address the handling of the output directory from
working to final location.
1416 issue is to address the handling of the output of a part from working to
final location with speculative execution considerations.
The OutputHandler (of 1558) provides to the task the base path where to create
the part file.
To solve 1416 a similar pattern (of 1558) could be used.
A way of doing it could be by adding to the OutputHandler interface the
following methods:
Path getUncommittedPartFile(JobConf conf, taskId);
void initializePartFile(JobConf conf, taskId);
void commitPartFile(JobConf conf, taskId);
The speculative execution logic would invoke init/commit methods and this
methods would resolve any cleanup/discard to be done.
As this is handled in the OutputHandler implementation it will work for
non-file base scenarios as well and transparently from the TT.
> changes to OutputFormat to work on temporary directory to enable re-running
> crashed jobs (Issue: 1121)
> ------------------------------------------------------------------------------------------------------
>
> Key: HADOOP-1558
> URL: https://issues.apache.org/jira/browse/HADOOP-1558
> Project: Hadoop
> Issue Type: Improvement
> Components: mapred
> Environment: all
> Reporter: Alejandro Abdelnur
> Fix For: 0.15.0
>
> Attachments: hadoop-1558-JUL2607-1600.txt
>
>
> Add OutputFormat methods like:
> /** Called to initialize output for this job. */
> void initialize(JobConf job) throws IOException;
> /** Called to finalize output for this job. */
> void commit(JobConf job) throws IOException;
> In the base implemenation for FileSystem output, initialize() might then
> create a temporary directory for the job, removing any that already exists,
> and commit could rename the temporary output directory to the final name.
> The existing checkOutputSpecs() would continue to throw an exception if the
> final output already exists.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
