[ https://issues.apache.org/jira/browse/HDFS-4387?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13551665#comment-13551665 ]
Colin Patrick McCabe commented on HDFS-4387: -------------------------------------------- even after making this change, JamVM crashes in nmdShutdown. This is a simple function that does nothing but call {{MiniDFSCluster#shutdown}}. {code} 2013-01-11 15:46:06,003 INFO hdfs.MiniDFSCluster (MiniDFSCluster.java:shutdown(1411)) - Shutting down the Mini HDFS Cluster 2013-01-11 15:46:06,004 INFO hdfs.MiniDFSCluster (MiniDFSCluster.java:shutdownDataNodes(1438)) - Shutting down DataNode 0 2013-01-11 15:46:06,004 WARN datanode.DirectoryScanner (DirectoryScanner.java:shutdown(289)) - DirectoryScanner: shutdown has been called 2013-01-11 15:46:06,015 INFO mortbay.log (Slf4jLog.java:info(67)) - Stopped SelectChannelConnector@localhost:0 2013-01-11 15:46:06,016 INFO ipc.Server (Server.java:stop(2096)) - Stopping server on 39373 2013-01-11 15:46:06,018 INFO ipc.Server (Server.java:run(829)) - Stopping IPC Server Responder 2013-01-11 15:46:06,018 INFO ipc.Server (Server.java:run(687)) - Stopping IPC Server listener on 39373 Segmentation fault (core dumped) {code} I really have no idea how to fix this, or even if it can be fixed. It looks like something that shutdown itself is doing, rather than something in the C code. Perhaps JamVM can't handle threads shutting down or the log source being closed? > test_libhdfs_threaded SEGV on OpenJDK 7 > --------------------------------------- > > Key: HDFS-4387 > URL: https://issues.apache.org/jira/browse/HDFS-4387 > Project: Hadoop HDFS > Issue Type: Bug > Components: libhdfs > Affects Versions: 3.0.0 > Reporter: Andy Isaacson > Priority: Minor > > Building and running tests on OpenJDK 7 on Ubuntu 12.10 fails with {{mvn test > -Pnative}}. The output is hard to decipher but the underlying issue is that > {{test_libhdfs_native}} segfaults at startup. > {noformat} > (gdb) run > Starting program: > /mnt/trunk/hadoop-hdfs-project/hadoop-hdfs/target/native/test_libhdfs_threaded > [Thread debugging using libthread_db enabled] > Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". > Program received signal SIGSEGV, Segmentation fault. > 0x00007ffff739a897 in attachJNIThread (name=0x0, is_daemon=is_daemon@entry=0 > '\000', group=0x0) at thread.c:768 > 768 thread.c: No such file or directory. > (gdb) where > #0 0x00007ffff739a897 in attachJNIThread (name=0x0, > is_daemon=is_daemon@entry=0 '\000', group=0x0) at thread.c:768 > #1 0x00007ffff7395020 in attachCurrentThread (is_daemon=0, args=0x0, > penv=0x7fffffffddb8) at jni.c:1454 > #2 Jam_AttachCurrentThread (vm=<optimized out>, penv=0x7fffffffddb8, > args=0x0) at jni.c:1466 > #3 0x00007ffff7bcf979 in getGlobalJNIEnv () at > /mnt/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/native/libhdfs/jni_helper.c:527 > #4 getJNIEnv () at > /mnt/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/native/libhdfs/jni_helper.c:585 > #5 0x0000000000402512 in nmdCreate (conf=conf@entry=0x7fffffffdeb0) at > /mnt/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/native/libhdfs/native_mini_dfs.c:49 > #6 0x00000000004016e1 in main () at > /mnt/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/native/libhdfs/test_libhdfs_threaded.c:283 > {noformat} -- 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