[ https://issues.apache.org/jira/browse/HDFS-10921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15537074#comment-15537074 ]
Eric Badger commented on HDFS-10921: ------------------------------------ Quick update: Adding the @Before method as [~xkrogen] proposed above does not work. For a reason that isn't immediately obvious to me, {{cluster.getNamesystem().clear()}} breaks all of the tests. Even when run on only a single test in the test class, clearing the namesystem before the test completely breaks it. If I clear the namesystem in tandem with {{cluster.restartNameNodes()}} then all of the tests pass, but they take ~82 seconds to run on my machine. Restarting the cluster after every test, on the other hand, takes under 30 seconds total. However, by adding the else clause to the @Before method, I was able to remove all of the {{restartNameNodes()}} and {{restartDataNode()}} calls in the code and the tests still passed. So the only problem that we have is the issue that [~shahrs87] raised about Namesystem pollution by bad tests. This test suite doesn't have any of those problems, so (unless I can figure out how to effectively clear the namesystem in a small amount of time without corrupting the tests) the question will be whether we want decreased runtime or to be resilient against tests written in the future that use the same file names. > TestDiskspaceQuotaUpdate doesn't wait for NN to get out of safe mode > -------------------------------------------------------------------- > > Key: HDFS-10921 > URL: https://issues.apache.org/jira/browse/HDFS-10921 > Project: Hadoop HDFS > Issue Type: Bug > Reporter: Eric Badger > Assignee: Eric Badger > Attachments: HDFS-10921.001.patch, HDFS-10921.002.patch > > > Test fails intermittently because the NN is still in safe mode. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org