[
https://issues.apache.org/jira/browse/HDFS-8156?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14498838#comment-14498838
]
Kai Zheng commented on HDFS-8156:
---------------------------------
Thanks [~szetszwo] for the comments!
bq.Adding another schema complicate the testing.
I agree. I thought we're not late and I would take to get more complicated
tests. Maybe you could elaborate a little bit for the concrete job? What
concerned me is if we don't have at least two schemas, it's hard to tell a
schema is actually corrected used across the whole system, client, namenode and
datanode. Two schemas allow to expose some tricky things. On the other hand,
having a schema support in hand, it won't take too much to support the other
one. As the new schema is defined in Java codes as discussed with [~zhz] and
you in HDFS-7859, I thought it removes the overhead or concern related to the
customization XML. I wish this would sound OK for you at this point. Thanks.
Note RS-10-4 is originated from the deployment experience in FB.
bq.It is better to use RS-6-3 instead of SYS-DEFAULT-RS-6-3 so that the default
schema can be changed later on.
Totally agree, will do that. A good point!
Thanks again.
> Define some system schemas in codes
> -----------------------------------
>
> Key: HDFS-8156
> URL: https://issues.apache.org/jira/browse/HDFS-8156
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Kai Zheng
> Assignee: Kai Zheng
> Attachments: HDFS-8156-v1.patch
>
>
> This is to define and add some system schemas in codes, and also resolve some
> TODOs left for HDFS-7859 and HDFS-7866 as they're still subject to further
> discussion.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
