[
https://issues.apache.org/jira/browse/HDFS-13951?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16635643#comment-16635643
]
Daryn Sharp commented on HDFS-13951:
------------------------------------
Please don't combine unrelated patches on another Jira. Glad someone is fixing
this. I forgot hacking this long ago.
Also, there should be no need to conditionalize on hdfs. Just let the tokens
stringify themselves.
> HDFS DelegationTokenFetcher can't print non-HDFS tokens in a tokenfile
> ----------------------------------------------------------------------
>
> Key: HDFS-13951
> URL: https://issues.apache.org/jira/browse/HDFS-13951
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: tools
> Affects Versions: 3.2.0
> Reporter: Steve Loughran
> Assignee: Steve Loughran
> Priority: Minor
>
> the fetchdt command can fetch tokens for filesystems other than hdfs (s3a,
> abfs, etc), but it can't print them, as it assumes all tokens in the file are
> subclasses of
> {{org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenIdentifier}}
> & uses this fact in its decoding. It deserializes the token byte array
> without checking kind and so ends up with invalid data.
> Fix: ask the tokens to decode themselves; only call toStableString() if an
> HDFS token.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
