You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Bill Habermaas (JIRA)" <ji...@apache.org> on 2009/01/23 16:27:59 UTC
[jira] Created: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
---------------------------------------------------------
Key: HADOOP-5116
URL: https://issues.apache.org/jira/browse/HADOOP-5116
Project: Hadoop Core
Issue Type: Test
Components: test
Affects Versions: 0.18.2
Environment: AIX
Reporter: Bill Habermaas
Priority: Minor
This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Bill Habermaas (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12668484#action_12668484 ]
Bill Habermaas commented on HADOOP-5116:
----------------------------------------
When this test was running it was observed to be consuming over 90% of the processor according to the 'topas' command. This is why the report description assumes that a loop was occurring. At the instant that the trace was induced it was probably blocked but not forever. The high cpu utilization suggests that it is not blocking and is repeating the write according to the way the test is written.
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
> Attachments: javacore.20090126.144729.376858.0001.txt
>
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Raghu Angadi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12666687#action_12666687 ]
Raghu Angadi commented on HADOOP-5116:
--------------------------------------
Could attach the jstack of the test when this is stuck?
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Raghu Angadi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12668556#action_12668556 ]
Raghu Angadi commented on HADOOP-5116:
--------------------------------------
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
Is it writing after closing the socket? The stacktrace does not seem to show it.
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
> Attachments: javacore.20090126.144729.376858.0001.txt
>
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Bill Habermaas (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Bill Habermaas updated HADOOP-5116:
-----------------------------------
Attachment: javacore.20090126.144729.376858.0001.txt
This dump was taken after the test was running for several minutes.
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
> Attachments: javacore.20090126.144729.376858.0001.txt
>
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Raghu Angadi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12668553#action_12668553 ]
Raghu Angadi commented on HADOOP-5116:
--------------------------------------
hmm... did you try strace (or its AIX equivalent) on the busy thread? On Linux, I find the pid of the thread doing write() with jstack.
This test depends on the fact that a pipe has a reasonably finite buffer. So when it writes to the pipe in a loop, it expects write to return 0 at some point and polling for 'POLLOUT' would timeout after that since there is no reader on the other side.
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
> Attachments: javacore.20090126.144729.376858.0001.txt
>
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Raghu Angadi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12668261#action_12668261 ]
Raghu Angadi commented on HADOOP-5116:
--------------------------------------
The stacktrace shows that test is blocked inside a write() to a pipe. This write is supposed to be non-blocking. Is it possible that either JVM or AIX ignores {{channel.configureBlocking(false)}} for pipes?
Could you check if this process is busy during this time? It it is then then there could be other possibilities.
{noformat}
4XESTACKTRACE at sun/nio/ch/FileDispatcher.write0(Native Method)
4XESTACKTRACE at sun/nio/ch/FileDispatcher.write(FileDispatcher.java:76(Compiled Code))
4XESTACKTRACE at sun/nio/ch/IOUtil.writeFromNativeBuffer(IOUtil.java:119(Compiled Code))
4XESTACKTRACE at sun/nio/ch/IOUtil.write(IOUtil.java:87(Compiled Code))
4XESTACKTRACE at sun/nio/ch/SinkChannelImpl.write(SinkChannelImpl.java:164(Compiled Code))
4XESTACKTRACE at org/apache/hadoop/net/SocketOutputStream$Writer.performIO(SocketOutputStream.java:55(Compiled Code))
4XESTACKTRACE at org/apache/hadoop/net/SocketIOWithTimeout.doIO(SocketIOWithTimeout.java:140)
4XESTACKTRACE at org/apache/hadoop/net/SocketOutputStream.write(SocketOutputStream.java:146)
4XESTACKTRACE at org/apache/hadoop/net/SocketOutputStream.write(SocketOutputStream.java:107)
4XESTACKTRACE at java/io/OutputStream.write(OutputStream.java:79)
4XESTACKTRACE at org/apache/hadoop/net/TestSocketIOWithTimeout.doIO(TestSocketIOWithTimeout.java:58)
4XESTACKTRACE at org/apache/hadoop/net/TestSocketIOWithTimeout.testSocketIOWithTimeout(TestSocketIOWithTimeout.java:106)
4XESTACKTRACE at sun/reflect/NativeMethodAccessorImpl.invoke0(Native Method)
4XESTACKTRACE at sun/reflect/NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:79)
4XESTACKTRACE at sun/reflect/DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
4XESTACKTRACE at java/lang/reflect/Method.invoke(Method.java:618)
{noformat}
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
> Attachments: javacore.20090126.144729.376858.0001.txt
>
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5116) TestSocketIOWithTimeout fails under
AIX - TIMEOUT error.
Posted by "Steve Loughran (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12666684#action_12666684 ]
Steve Loughran commented on HADOOP-5116:
----------------------------------------
How do you tell if a socket is closed on AIX? {{Socket.isClosed()}} , presumably?
> TestSocketIOWithTimeout fails under AIX - TIMEOUT error.
> ---------------------------------------------------------
>
> Key: HADOOP-5116
> URL: https://issues.apache.org/jira/browse/HADOOP-5116
> Project: Hadoop Core
> Issue Type: Test
> Components: test
> Affects Versions: 0.18.2
> Environment: AIX
> Reporter: Bill Habermaas
> Priority: Minor
>
> This test expects an exception to occur when read/writing a closed socket. Under AIX this does not occur and results in a loop.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.