[
https://issues.apache.org/jira/browse/HADOOP-1645?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12514351
]
Raghu Angadi commented on HADOOP-1645:
--------------------------------------
Dhruba and I looked similar output. The problem is that close(DFS_FILE) waits
for dfs.replication.min to be satisfied before returning. But this test
requires all replicas to be present. For this test, may be we could set
dfs.replicaion.min to same as max. This test is done right after closing the
file.
> TestDecommission sometimes fails
> --------------------------------
>
> Key: HADOOP-1645
> URL: https://issues.apache.org/jira/browse/HADOOP-1645
> Project: Hadoop
> Issue Type: Bug
> Reporter: Doug Cutting
> Priority: Blocker
> Fix For: 0.14.0
>
> Attachments: TEST-org.apache.hadoop.dfs.TestDecommission.txt
>
>
> TestDecommission sometimes fails with:
> junit.framework.AssertionFailedError: Number of replicas for block0
> expected:<3> but was:<2>
> at
> org.apache.hadoop.dfs.TestDecommission.checkFile(TestDecommission.java:88)
> at
> org.apache.hadoop.dfs.TestDecommission.testDecommission(TestDecommission.java:285)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
