[
https://issues.apache.org/jira/browse/HDFS-4165?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13493644#comment-13493644
]
Binglin Chang commented on HDFS-4165:
-------------------------------------
Thanks Suresh and Brandon for review.
@Eli
Given the current constant value, the current sanity check can't fail
{code}
public static final long QUOTA_DONT_SET = Long.MAX_VALUE;
public static final long QUOTA_RESET = -1L;
{code}
HDFS-46 also reminds me that Long.MAX_VALUE is used both for QUOTA_DONT_SET and
the default namespace quota of root directory, so you can't set quota to
Long.MAX_VALUE for root directory using RPC, a better value for QUOTA_DONT_SET
should be -2 instead.
> Faulty sanity check in FsDirectory.unprotectedSetQuota
> ------------------------------------------------------
>
> Key: HDFS-4165
> URL: https://issues.apache.org/jira/browse/HDFS-4165
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: name-node
> Affects Versions: 3.0.0
> Reporter: Binglin Chang
> Assignee: Binglin Chang
> Priority: Trivial
> Fix For: 3.0.0
>
> Attachments: HDFS-4165.patch
>
>
> According to the documentation:
> The quota can have three types of values : (1) 0 or more will set
> the quota to that value, (2) {@link HdfsConstants#QUOTA_DONT_SET} implies
> the quota will not be changed, and (3) {@link HdfsConstants#QUOTA_RESET}
> implies the quota will be reset. Any other value is a runtime error.
> sanity check in FsDirectory.unprotectedSetQuota should use
> {code}
> nsQuota != HdfsConstants.QUOTA_RESET
> {code}
> rather than
> {code}
> nsQuota < HdfsConstants.QUOTA_RESET
> {code}
> Since HdfsConstants.QUOTA_RESET is defined to be -1, there is not any problem
> for this code, but it is better to do it right.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
