[
https://issues.apache.org/jira/browse/HADOOP-16857?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17070822#comment-17070822
]
Sneha Vijayarajan commented on HADOOP-16857:
--------------------------------------------
AzureADAuthenticator includes retry logic for all oAuth token providers that
are implemented in ABFS Driver. Where as CustomTokenProvider failures end up in
retry from the ABFS Driver REST request invoking layer which is not expected.
Ideally custom token provider implementation should include retry logic too.
Raising PR to stop retries getting triggered from the REST op layer due to
exceptions from custom token provider.
As custom token providers might have had benefited by the unintentional retry
happening, have retained a linear retry to trigger getAccessToken. The retry
count is made configurable.
> ABFS: Optimize HttpRequest retry triggers
> -----------------------------------------
>
> Key: HADOOP-16857
> URL: https://issues.apache.org/jira/browse/HADOOP-16857
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Affects Versions: 3.3.1
> Reporter: Sneha Vijayarajan
> Assignee: Sneha Vijayarajan
> Priority: Major
>
> Currently retry logic gets triggered when access token fetch fails even with
> irrecoverable errors. Causing a large wait time for the request failure to be
> reported.
>
> Retry logic needs to be optimized to identify such access token fetch
> failures and fail fast.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
