[
https://issues.apache.org/jira/browse/HDFS-7717?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14386097#comment-14386097
]
Kai Zheng commented on HDFS-7717:
---------------------------------
bq.In order to reduce complexity of error handling, it may be better to do
convert tasks as MR job.
I do believe this needs a good discussion before we can decide. In your view,
which way would be most lightweight and easy to implement ? We can have easy
one first, then more better one I guess ?
bq.f these information can be obtained from NN, we can implement converter tool
as like Mover
Sure we can get EC schema and policy from NN for target files. The required
interface and API are still under discussion. Before that, perhaps you can
assume you have already got them for your work ? We will update this when we
really get those parts ready.
> Erasure Coding: provide a tool for convert files between replication and
> erasure coding
> ---------------------------------------------------------------------------------------
>
> Key: HDFS-7717
> URL: https://issues.apache.org/jira/browse/HDFS-7717
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Jing Zhao
> Assignee: Kai Sasaki
>
> We need a tool to do offline conversion between replication and erasure
> coding. The tool itself can either utilize MR just like the current distcp,
> or act like the balancer/mover.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
