[
https://issues.apache.org/jira/browse/HDFS-7529?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14250698#comment-14250698
]
Haohui Mai commented on HDFS-7529:
----------------------------------
bq. Wouldn't it be better to fail fast in this case? Did you copy the wrong
code to #ensureKeysAreInitialized?
Likewise, I think that the checks for nullness of provider, keyName, and
metadata can be removed from #createEncryptionZoneInt, right?
Duplicating the checks is intentional to define well-formed steps as it is
implied by the name {{ensureKeysAreInitialized()}}.
bq. are now inside the FSN#writeLock(). I suppose that's not the end of the
world, but every little bit of extra code inside the writeLock() hurts.
The performance benefit is minimal as {{getPermissionChecker()}} eventually
synchronized in the {{UserGroupInformation#getCurrentUser()}}. Making it
consistent with other operations allows the further refactoring.
> Consolidate encryption zone related implementation into a single class
> ----------------------------------------------------------------------
>
> Key: HDFS-7529
> URL: https://issues.apache.org/jira/browse/HDFS-7529
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Haohui Mai
> Assignee: Haohui Mai
> Attachments: HDFS-7529.000.patch, HDFS-7529.001.patch
>
>
> This jira proposes to consolidate encryption zone related implementation to a
> single class.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
