You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Konstantin Shvachko (JIRA)" <ji...@apache.org> on 2009/12/09 02:58:18 UTC
[jira] Created: (HDFS-816) TestLargeDirectoryDelete fails on
windows
TestLargeDirectoryDelete fails on windows
-----------------------------------------
Key: HDFS-816
URL: https://issues.apache.org/jira/browse/HDFS-816
Project: Hadoop HDFS
Issue Type: Bug
Affects Versions: 0.21.0
Environment: Windows
Reporter: Konstantin Shvachko
Fix For: 0.21.0
{{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-816) TestLargeDirectoryDelete fails
Posted by "Steve Loughran (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-816?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12870035#action_12870035 ]
Steve Loughran commented on HDFS-816:
-------------------------------------
Also proposed: threads to catch all exceptions but InterruptedExceptions, store them somewhere, have the test code check them
> TestLargeDirectoryDelete fails
> ------------------------------
>
> Key: HDFS-816
> URL: https://issues.apache.org/jira/browse/HDFS-816
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.21.0
> Environment: Windows
> Reporter: Konstantin Shvachko
> Fix For: 0.21.0
>
> Attachments: TestLargeDirectoryDelete.log
>
>
> {{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-816) TestLargeDirectoryDelete fails
Posted by "Steve Loughran (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-816?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12870065#action_12870065 ]
Steve Loughran commented on HDFS-816:
-------------------------------------
This could be a race condition in shutdown: the DN had shut down while the creator thread was still live, so the creator thread got the cannot reproduce exception. The fact that the Datanode is shutdown message appeared in the log a few milliseconds after is meaningless, the DN shuts down its bits in a sequence, so it may have already dropped out the cluster before the message was printed.
> TestLargeDirectoryDelete fails
> ------------------------------
>
> Key: HDFS-816
> URL: https://issues.apache.org/jira/browse/HDFS-816
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.21.0
> Environment: Windows
> Reporter: Konstantin Shvachko
> Fix For: 0.21.0
>
> Attachments: TestLargeDirectoryDelete.log
>
>
> {{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-816) TestLargeDirectoryDelete fails
Posted by "Konstantin Shvachko (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-816?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Konstantin Shvachko updated HDFS-816:
-------------------------------------
Summary: TestLargeDirectoryDelete fails (was: TestLargeDirectoryDelete fails on windows)
I hear this is not windows specific. Hairong reports the same problem on Mac.
> TestLargeDirectoryDelete fails
> ------------------------------
>
> Key: HDFS-816
> URL: https://issues.apache.org/jira/browse/HDFS-816
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.21.0
> Environment: Windows
> Reporter: Konstantin Shvachko
> Fix For: 0.21.0
>
> Attachments: TestLargeDirectoryDelete.log
>
>
> {{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-816) TestLargeDirectoryDelete fails on
windows
Posted by "Konstantin Shvachko (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-816?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Konstantin Shvachko updated HDFS-816:
-------------------------------------
Attachment: TestLargeDirectoryDelete.log
Here is the log of the failed test.
Before the exception it created a bunch of files, but failed on the last one. Looks like the reason is that {{DFSClient.createBlockOutputStream()}} getst a {{BindException: Address already in use}} from the data-node.
> TestLargeDirectoryDelete fails on windows
> -----------------------------------------
>
> Key: HDFS-816
> URL: https://issues.apache.org/jira/browse/HDFS-816
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.21.0
> Environment: Windows
> Reporter: Konstantin Shvachko
> Fix For: 0.21.0
>
> Attachments: TestLargeDirectoryDelete.log
>
>
> {{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-816) TestLargeDirectoryDelete fails
Posted by "Steve Loughran (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-816?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Steve Loughran updated HDFS-816:
--------------------------------
Comment: was deleted
(was: Also proposed: threads to catch all exceptions but InterruptedExceptions, store them somewhere, have the test code check them)
> TestLargeDirectoryDelete fails
> ------------------------------
>
> Key: HDFS-816
> URL: https://issues.apache.org/jira/browse/HDFS-816
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.21.0
> Environment: Windows
> Reporter: Konstantin Shvachko
> Fix For: 0.21.0
>
> Attachments: TestLargeDirectoryDelete.log
>
>
> {{TestLargeDirectoryDelete}} consistently fails on windows complaining "file3800 could only be replicated to 0 nodes, instead of 1".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.