hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bharat Viswanadham (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-4843) testDeleteBlockPool has a bug which makes it fail occasionally
Date Thu, 08 Feb 2018 20:56:00 GMT

    [ https://issues.apache.org/jira/browse/HDFS-4843?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16357558#comment-16357558

Bharat Viswanadham commented on HDFS-4843:

Hi [~vincent cho]

This has been fixed as part of HDFS-5892

Now, in MiniDFSNNTopology, a method is added simpleFederatedTopology to take DFS_NAMESERVICES
as input. dfs.nameservices set in the test are used in the MiniDFSCluster.

> testDeleteBlockPool has a bug which makes it fail occasionally
> --------------------------------------------------------------
>                 Key: HDFS-4843
>                 URL: https://issues.apache.org/jira/browse/HDFS-4843
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: federation
>    Affects Versions: 2.0.4-alpha
>         Environment: normal enviroment
>            Reporter: vincent cho
>            Priority: Major
> int the test case "testDeleteBlockPool" :
> first we set "DFSConfigKeys.DFS_NAMESERVICES" as "namesServerId1,namesServerId2", but
after a cluster was built, the nameservice are set as "ns1,ns2" because the code set a default
value for them when we new one cluster.
> then we refresh namenode with the nameservice set as "namesServerId2" . so we will add
a new nameservice named "namesServerId2" and remove the two old nameservices. because "ns2"
and "namesServerId2" have the same bpid, so when the add thread run faster than the remove
process, the bpByBlockPoolId in BlockPoolManager will be empty  so that it fails when create
a new file or path.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message