[jira] [Commented] (HDFS-4704) Add a transient flag to file so that transient files won't be included in any snapshot

["Steve Loughran (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HDFS-4704?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13633890#comment-13633890 ] 

Steve Loughran commented on HDFS-4704:
--------------------------------------

What does transient mean here? Is the goal to be able to mark all /tmp & intermediate data, or even anything recreatable?

If so, I could see the flag also being useful in the block replication policies: you'd give underreplicated transient data lower priority than non-transient data with the same #of block copies.

We could also be more aggressive on DN decommission: pull off the non-transient data first, then ramp down the transient data.
                
> Add a transient flag to file so that transient files won't be included in any snapshot
> --------------------------------------------------------------------------------------
>
>                 Key: HDFS-4704
>                 URL: https://issues.apache.org/jira/browse/HDFS-4704
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: namenode
>            Reporter: Tsz Wo (Nicholas), SZE
>            Assignee: Tsz Wo (Nicholas), SZE
>
> See the description HDFS-4529.  We are going to implement (4) shown below:
> (4) mark the files with a new transient flag in HDFS. The files with the new flag will not be included in any snapshot. Then, concat could remove the files as usual.

--
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