hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yongjun Zhang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-9732) Remove DelegationTokenIdentifier.toString() —for better logging output
Date Sat, 26 Mar 2016 17:55:25 GMT

    [ https://issues.apache.org/jira/browse/HDFS-9732?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15213129#comment-15213129

Yongjun Zhang commented on HDFS-9732:

Hi [~stevel@apache.org],

Ah, sorry, thanks for your comments, I did not see until now, 

About the method name, {{detailsForCLI()}}, do we have any other occasion that we want a "frozen"
output then CLI? For example, one day we need to call the frozen version for webui? If CLI
is the only situation, I can change the new method name per your suggestion.

One question:
here subclasses would need to know to not call super.toString() and instead call some other
Derived class may have new field to print on top of the base class'. If we don't call {{super.toString()}},
we probably introduce a new method to be called by both the base class'  and child class'

In this case, all the info printed by the base {{toString()}} is applicable to child class.
And in my rev 2, I added {{getKind()}} which is overriden by child classes. Basically we are
in full control of what the output looks like. Would you please explain why calling super.toString()
 is a bad idea (especially we do define our own base class' {{toString()}})?  I can see a
problem when we don't define {{toString()}} class for the base, in which case the java base
{{Object}}'s {{toString()}} would be called.



> Remove DelegationTokenIdentifier.toString() —for better logging output
> ----------------------------------------------------------------------
>                 Key: HDFS-9732
>                 URL: https://issues.apache.org/jira/browse/HDFS-9732
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>    Affects Versions: 2.7.2
>            Reporter: Steve Loughran
>            Assignee: Yongjun Zhang
>         Attachments: HADOOP-12752-001.patch, HDFS-9732.001.patch, HDFS-9732.002.patch
>   Original Estimate: 0.5h
>  Remaining Estimate: 0.5h
> HDFS {{DelegationTokenIdentifier.toString()}} adds some diagnostics info, owner, sequence
number. But its superclass,  {{AbstractDelegationTokenIdentifier}} contains a lot more information,
including token issue and expiry times.
> Because  {{DelegationTokenIdentifier.toString()}} doesn't include this data,
> information that is potentially useful for kerberos diagnostics is lost.

This message was sent by Atlassian JIRA

View raw message