[
https://issues.apache.org/jira/browse/HDFS-6474?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14058209#comment-14058209
]
Andrew Wang commented on HDFS-6474:
-----------------------------------
Hey Charles, thanks for taking a look. I incorporated most of your comments, a
few replies:
bq. DEK and EDEK to be proper nouns...
I sprinkled these abbrevs around in addition to the long form, that should be
enough?
bq. LOG.debug and LOG.trace
It's deliberate, trace is basically for devs only, I ask customers to turn on
debug sometimes.
bq. Is it typical to use MS time intervals for config options in HDFS?
Yea, pretty typical, I've turned down monitor refreshes to <1s before for
testing, so would like to keep it as is
bq. would it be useful to assert that the FSD lock is held
did this, also added asserts for the manager lock
bq. In general, is there a reason why you use KeyProvider.KeyVersion instead of
importing KeyProvider.KeyVersion directly?
It's a nested class, so IntelliJ wants me to do a static import to use it
directly. Easy shortcut though, so I did it for EZManager.
I think Yi is also taking a look, so I'll wait for him to post a review before
revving the patch.
> Namenode needs to get the actual keys and iv from the KeyProvider
> -----------------------------------------------------------------
>
> Key: HDFS-6474
> URL: https://issues.apache.org/jira/browse/HDFS-6474
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: namenode, security
> Reporter: Charles Lamb
> Assignee: Andrew Wang
> Attachments: hdfs-6474.001.patch
>
>
> The Namenode has code to connect to the KeyProvider, but it needs to actually
> get the keys and return them to the client at the right time.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
