[jira] [Resolved] (HDFS-5706) Should reset Balancer.Parameters.DEFALUT.policy in TestBalancer

["Yu Li (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HDFS-5706?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yu Li resolved HDFS-5706.
-------------------------

    Resolution: Duplicate

After more careful investigation, the issue is already fixed along with the patch of HDFS-4261, so mark the duplication and close it directly

> Should reset Balancer.Parameters.DEFALUT.policy in TestBalancer
> ---------------------------------------------------------------
>
>                 Key: HDFS-5706
>                 URL: https://issues.apache.org/jira/browse/HDFS-5706
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: balancer
>    Affects Versions: 2.2.0
>            Reporter: Yu Li
>            Assignee: Yu Li
>            Priority: Minor
>
> Now in TestBalancer.java, more than one test case will invoke the private method runBalancer, in which it will use Balancer.Parameters.Default, while the policy is never reset thus its totalUsedSpace and totalCapacity will increase continuously.
> We can reveal this issue by simply change
> {code:title=TestBalancer#testBalancer1Internal}
> testUnevenDistribution(conf,
>         new long[] {50*CAPACITY/100, 10*CAPACITY/100},
>         new long[]{CAPACITY, CAPACITY},
>         new String[] {RACK0, RACK1});
> {code}
> to
> {code:title=TestBalancer#testBalancer1Internal}
> testUnevenDistribution(conf,
>         new long[] {70*CAPACITY/100, 40*CAPACITY/100},
>         new long[]{CAPACITY, CAPACITY},
>         new String[] {RACK0, RACK1});
> {code}
> which in current implement, will cause none node under-replication thus cause the test case fail



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)