[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13466205#comment-13466205 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Hdfs-0.23-Build #389 (See [https://builds.apache.org/job/Hadoop-Hdfs-0.23-Build/389/]) HDFS-3373. FileContext HDFS implementation can leak socket caches (John George via bobby) (Revision 1391542) Result = UNSTABLE bobby : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1391542 Files : * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/branches/branch-0.23/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 0.23.3, 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 0.23.4, 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.branch23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13465670#comment-13465670 ] Robert Joseph Evans commented on HDFS-3373: --- The 0.23 patch looks like a fairly straight forward port of the trunk version, but what happened to hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java? FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 0.23.3, 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.branch23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13465718#comment-13465718 ] John George commented on HDFS-3373: --- In 2.0, setting cache capacity to 0 disabled the cache, this (along with some other changes) was not in 0.23. TestSocketCache was a one test file to test if cache disabling worked. Since, 0.23 does not have that, I did not port the test. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 0.23.3, 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.branch23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13465721#comment-13465721 ] Robert Joseph Evans commented on HDFS-3373: --- Makes since. Because it is such a straight forward patch I feel OK checking the code in. Thanks for the work John. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 0.23.3, 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.branch23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13464695#comment-13464695 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Hdfs-trunk #1178 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/1178/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = SUCCESS szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13464749#comment-13464749 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Mapreduce-trunk #1209 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1209/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = SUCCESS szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: [jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
On 9/27/2012 10:00 PM, Hudson (JIRA) wrote: [ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13464749#comment-13464749 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Mapreduce-trunk #1209 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1209/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = SUCCESS szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira unsubscribe
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13464949#comment-13464949 ] John George commented on HDFS-3373: --- Thanks Nicholas. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13465199#comment-13465199 ] Hadoop QA commented on HDFS-3373: - {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12546899/HDFS-3373.branch23.patch against trunk revision . {color:red}-1 patch{color}. The patch command could not apply the patch. Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3246//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 0.23.3, 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.branch23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13463865#comment-13463865 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Common-trunk-Commit #2775 (See [https://builds.apache.org/job/Hadoop-Common-trunk-Commit/2775/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = SUCCESS szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13463869#comment-13463869 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Hdfs-trunk-Commit #2838 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk-Commit/2838/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = SUCCESS szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13463901#comment-13463901 ] Hudson commented on HDFS-3373: -- Integrated in Hadoop-Mapreduce-trunk-Commit #2797 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk-Commit/2797/]) HDFS-3373. Change DFSClient input stream socket cache to global static and add a thread to cleanup expired cache entries. Contributed by John George (Revision 1390466) Result = FAILURE szetszwo : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVNview=revrev=1390466 Files : * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSClient.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/SocketCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestConnCache.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDistributedFileSystem.java * /hadoop/common/trunk/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSocketCache.java FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Fix For: 2.0.3-alpha Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13462787#comment-13462787 ] Tsz Wo (Nicholas), SZE commented on HDFS-3373: -- {code} + private synchronized void evictExpired(long expiryPeriod) + throws IOException { {code} evictExpired(..) still has throws IOException but IOException is never thrown in the code. Patch looks good other than that. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13463385#comment-13463385 ] Hadoop QA commented on HDFS-3373: - {color:green}+1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12546571/HDFS-3373.trunk.patch.4 against trunk revision . {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 3 new or modified test files. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:green}+1 javadoc{color}. The javadoc tool did not generate any warning messages. {color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse. {color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 1.3.9) warnings. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:green}+1 core tests{color}. The patch passed unit tests in hadoop-hdfs-project/hadoop-hdfs. {color:green}+1 contrib tests{color}. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-HDFS-Build/3232//testReport/ Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3232//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.4 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13461872#comment-13461872 ] Hadoop QA commented on HDFS-3373: - +1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12546297/HDFS-3373.trunk.patch.3 against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified test files. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 javadoc. The javadoc tool did not generate any warning messages. +1 eclipse:eclipse. The patch built with eclipse:eclipse. +1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. +1 core tests. The patch passed unit tests in hadoop-hdfs-project/hadoop-hdfs. +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-HDFS-Build/3227//testReport/ Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3227//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2, HDFS-3373.trunk.patch.3, HDFS-3373.trunk.patch.3 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13458012#comment-13458012 ] John George commented on HDFS-3373: --- Thanks for looking into the patch Nicholas. {quote} - In SocketCache.getInstance(..), -* If SocketCache is already initialized (i.e. isInitedOnce == true), then the parameters c and e will be igored silently. Would it be a problem? {quote} This was done on purpose to avoid complications. I thought it would be better to keep it simple and allow only one valid setting of c and e. {quote} -* If it throws IllegalStateException, should isInitedOnce not be set to true? -* typo: initialzied = initialized - In SocketCache.evictExpired(..), -* remove throws IOException since it does not throw it. -* remove synchronized(this) since it is already synchronized. - In SocketCache, add createTime to SocketAndStreams and remove SocketProp. - In SocketCache.put(..), move Preconditions.checkNotNull(sock); to the very beginning. {quote} Thanks for these comments - will upload a patch incorporating these soon. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13458016#comment-13458016 ] John George commented on HDFS-3373: --- {quote} -* If SocketCache is already initialized (i.e. isInitedOnce == true), then the parameters c and e will be igored silently. Would it be a problem? {quote} I suppose you actually meant if we should LOG a message in cases where someone tries to change the value after it has been set once? I agree that it is a good idea. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13458178#comment-13458178 ] Hadoop QA commented on HDFS-3373: - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12545606/HDFS-3373.trunk.patch.2 against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified test files. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 javadoc. The javadoc tool did not generate any warning messages. +1 eclipse:eclipse. The patch built with eclipse:eclipse. -1 findbugs. The patch appears to introduce 1 new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. +1 core tests. The patch passed unit tests in hadoop-hdfs-project/hadoop-hdfs. +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-HDFS-Build/3202//testReport/ Findbugs warnings: https://builds.apache.org/job/PreCommit-HDFS-Build/3202//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-hdfs.html Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3202//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1, HDFS-3373.trunk.patch.2 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13457589#comment-13457589 ] Tsz Wo (Nicholas), SZE commented on HDFS-3373: -- Hi John, thanks for working on this. Some comments on the patch: - In SocketCache.getInstance(..), -* If SocketCache is already initialized (i.e. isInitedOnce == true), then the parameters c and e will be igored silently. Would it be a problem? -* If it throws IllegalStateException, should isInitedOnce not be set to true? -* typo: initialzied = initialized - In SocketCache.evictExpired(..), -* remove throws IOException since it does not throw it. -* remove synchronized(this) since it is already synchronized. - In SocketCache, add createTime to SocketAndStreams and remove SocketProp. - In SocketCache.put(..), move Preconditions.checkNotNull(sock); to the very beginning. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13443524#comment-13443524 ] Hadoop QA commented on HDFS-3373: - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12542795/HDFS-3373.trunk.patch.1 against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified test files. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 javadoc. The javadoc tool did not generate any warning messages. +1 eclipse:eclipse. The patch built with eclipse:eclipse. -1 findbugs. The patch appears to introduce 2 new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. -1 core tests. The patch failed these unit tests in hadoop-hdfs-project/hadoop-hdfs: org.apache.hadoop.hdfs.TestHftpDelegationToken +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-HDFS-Build/3110//testReport/ Findbugs warnings: https://builds.apache.org/job/PreCommit-HDFS-Build/3110//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-hdfs.html Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3110//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch, HDFS-3373.trunk.patch.1 As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13442477#comment-13442477 ] Daryn Sharp commented on HDFS-3373: --- A related issue we need to consider, although tangental is that inability to {{close}} doesn't give a fs a chance to cleanup. Ex. closing streams and deleting temp files. I think there might be another jira for that? FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13442642#comment-13442642 ] John George commented on HDFS-3373: --- Daryn, I am not able to follow - Inability to close fs is not related to this JIRA because this JIRA is related to caching sockets when an input stream is closed. If I misunderstood your comment, could you elaborate a little more on what you meant? FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13442734#comment-13442734 ] Hadoop QA commented on HDFS-3373: - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12542653/HDFS-3373.trunk.patch against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified test files. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 javadoc. The javadoc tool did not generate any warning messages. +1 eclipse:eclipse. The patch built with eclipse:eclipse. -1 findbugs. The patch appears to introduce 2 new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. -1 core tests. The patch failed these unit tests in hadoop-hdfs-project/hadoop-hdfs: org.apache.hadoop.hdfs.TestConnCache org.apache.hadoop.hdfs.TestHftpDelegationToken +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-HDFS-Build/3101//testReport/ Findbugs warnings: https://builds.apache.org/job/PreCommit-HDFS-Build/3101//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-hdfs.html Console output: https://builds.apache.org/job/PreCommit-HDFS-Build/3101//console This message is automatically generated. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha, 3.0.0 Reporter: Todd Lipcon Assignee: John George Attachments: HDFS-3373.branch-23.patch, HDFS-3373.trunk.patch As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13287722#comment-13287722 ] Daryn Sharp commented on HDFS-3373: --- On a related tangent, a similar situation causes the NMs to leak sockets due to unclosed {{DistributedFileSystems}}. Each one holds a {{DFSClient}} that holds a cache with at least 1 CLOSE_WAIT socket. The socket gets closed if/when the {{DFSClient}} looks for a cached socket, but it gets replaced with another CLOSE_WAIT socket. With security enabled, a NM leaks 1 socket/user. With security disabled, a NM will leak *1 socket/job* since the UGI always appears different which causes multiple instances of the same dfs to pile up in the fs cache. The UGI/fs-cache is a separate issue, but there should probably be something like another thread that scans the cached sockets to remove closed ones. Otherwise {{FileContext}} will be ill-suited for a daemon that needs to create more than one context since it too will leak sockets. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0-alpha Reporter: Todd Lipcon Assignee: John George As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13269801#comment-13269801 ] Suresh Srinivas commented on HDFS-3373: --- Is this a blocker for 2.0 given HDFS-3359 has been committed to 2.0? FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0 Reporter: Todd Lipcon As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13269804#comment-13269804 ] Todd Lipcon commented on HDFS-3373: --- No, I don't think so, because we also now have a fix on the DN side which closes the sockets. So it might result in TIME_WAIT or FIN_WAIT sockets on the client, but shouldn't impact DNs anymore. Unless someone has seen this to be a production issue with 2.0, it shouldn't be considered blocker. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0 Reporter: Todd Lipcon As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-3373) FileContext HDFS implementation can leak socket caches
[ https://issues.apache.org/jira/browse/HDFS-3373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13268614#comment-13268614 ] Todd Lipcon commented on HDFS-3373: --- Nicholas had some ideas about this in HDFS-3359. Quoting here: {quote} I think we should either - change it from per DFSClient to per DFSInputStream; or - change it to a global static cache (like what we did for LeaseRenewer.) {quote} Brainstorming: maybe another option would be to have a timer thread which periodically wakes up and scans SocketCaches for old sockets. If any socket has been sitting around longer than a minute or two, we can close it proactively. FileContext HDFS implementation can leak socket caches -- Key: HDFS-3373 URL: https://issues.apache.org/jira/browse/HDFS-3373 Project: Hadoop HDFS Issue Type: Bug Components: hdfs client Affects Versions: 2.0.0 Reporter: Todd Lipcon As noted by Nicholas in HDFS-3359, FileContext doesn't have a close() method, and thus never calls DFSClient.close(). This means that, until finalizers run, DFSClient will hold on to its SocketCache object and potentially have a lot of outstanding sockets/fds held on to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira