[ 
https://issues.apache.org/jira/browse/HDFS-11804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16033667#comment-16033667
 ] 

Rushabh S Shah commented on HDFS-11804:
---------------------------------------

bq. Correct. So we should treat AuthenticationException similar as 
AccessControlException and AuthorizationException, right?
Somehow I mixed AuthenticationException with AuthorizationException. my bad.
Yes will incorporate the change in next patch.
bq. Say for example we have 3 KMS's. I think we should retry immediately for 
all 3 at first.
Makes sense..The only issue I can see is the first time we calculate 
{{FailoverOnNetworkExceptionRetry.getFailoverOrRetrySleepTime(int times)}} will 
have {{times == #kms servers}} instead of gradually increasing the sleep time. 
But maybe its ok.
Will attach a new patch shortly incorporating all the comments.
Thanks again [~xiaochen] !

> KMS client needs retry logic
> ----------------------------
>
>                 Key: HDFS-11804
>                 URL: https://issues.apache.org/jira/browse/HDFS-11804
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 2.6.0
>            Reporter: Rushabh S Shah
>            Assignee: Rushabh S Shah
>         Attachments: HDFS-11804-trunk-1.patch, HDFS-11804-trunk.patch
>
>
> The kms client appears to have no retry logic – at all.  It's completely 
> decoupled from the ipc retry logic.  This has major impacts if the KMS is 
> unreachable for any reason, including but not limited to network connection 
> issues, timeouts, the +restart during an upgrade+.
> This has some major ramifications:
> # Jobs may fail to submit, although oozie resubmit logic should mask it
> # Non-oozie launchers may experience higher rates if they do not already have 
> retry logic.
> # Tasks reading EZ files will fail, probably be masked by framework reattempts
> # EZ file creation fails after creating a 0-length file – client receives 
> EDEK in the create response, then fails when decrypting the EDEK
> # Bulk hadoop fs copies, and maybe distcp, will prematurely fail



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to