Thanks for your guys' help. It looks like the libhadoop.so didn't link the snappy in it, even it claims to support it in 2.1. I already created a ticket for this. I manually compiled the hadoop native library myself, and use that for now. Thanks Yong
From: [email protected] To: [email protected] Subject: RE: enable snappy on hadoop 1.1.1 Date: Mon, 7 Oct 2013 04:16:47 +0000 Problem might be snappy packages may not bulit(While buliding the hadoop you need to include snappy).. libhadoop.so should contain the snappy packages Please check the following output to check whether snappy is included or not.... Verification : nm {HADOOP_HOME}/lib/native/Linux-amd64-64/libhadoop.so | grep snappy From: bharath vissapragada [[email protected]] Sent: Sunday, October 06, 2013 2:25 AM To: User Subject: Re: enable snappy on hadoop 1.1.1 whats the output of ldd on that lib? Does it link properly? You should compile natives for your platforms as the packaged ones may not link properly. On Sat, Oct 5, 2013 at 2:37 AM, java8964 java8964 <[email protected]> wrote: I kind of read the hadoop 1.1.1 source code for this, it is very strange for me now. >From the error, it looks like runtime JVM cannot find the native method of >org/apache/hadoop/io/compress/snappy/SnappyCompressor.compressBytesDirect()I, >that my guess from the error message, but from the log, it looks like all the >native library, include native-hadoop and native snappy are both loaded, as shown in the failed task log: 2013-10-04 16:33:21,635 INFO org.apache.hadoop.util.NativeCodeLoader: Loaded the native-hadoop library 2013-10-04 16:33:22,006 INFO org.apache.hadoop.util.ProcessTree: setsid exited with exit code 0 2013-10-04 16:33:22,020 INFO org.apache.hadoop.mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@28252825 2013-10-04 16:33:22,111 INFO org.apache.hadoop.mapred.MapTask: numReduceTasks: 1 2013-10-04 16:33:22,116 INFO org.apache.hadoop.mapred.MapTask: io.sort.mb = 256 2013-10-04 16:33:22,168 INFO org.apache.hadoop.mapred.MapTask: data buffer = 204010960/255013696 2013-10-04 16:33:22,168 INFO org.apache.hadoop.mapred.MapTask: record buffer = 671088/838860 2013-10-04 16:33:22,342 WARN org.apache.hadoop.io.compress.snappy.LoadSnappy: Snappy native library is available 2013-10-04 16:33:22,342 INFO org.apache.hadoop.io.compress.snappy.LoadSnappy: Snappy native library loaded 2013-10-04 16:33:44,054 INFO org.apache.hadoop.mapred.MapTask: Starting flush of map output 2013-10-04 16:33:44,872 WARN org.apache.hadoop.io.compress.snappy.SnappyCompressor: java.lang.UnsatisfiedLinkError: org/apache/hadoop/io/compress/snappy/SnappyCompressor.initIDs()V 2013-10-04 16:33:44,872 INFO org.apache.hadoop.io.compress.CodecPool: Got brand-new compressor 2013-10-04 16:33:44,928 INFO org.apache.hadoop.mapred.TaskLogsTruncater: Initializing logs' truncater with mapRetainSize=-1 and reduceRetainSize=-1 2013-10-04 16:33:44,951 INFO org.apache.hadoop.io.nativeio.NativeIO: Initialized cache for UID to User mapping with a cache timeout of 14400 seconds. 2013-10-04 16:33:44,951 INFO org.apache.hadoop.io.nativeio.NativeIO: Got UserName yzhang for UID 1000 from the native implementation 2013-10-04 16:33:44,952 FATAL org.apache.hadoop.mapred.Child: Error running child : java.lang.UnsatisfiedLinkError: org/apache/hadoop/io/compress/snappy/SnappyCompressor.compressBytesDirect()I at org.apache.hadoop.io.compress.snappy.SnappyCompressor.compress(SnappyCompressor.java:229) at org.apache.hadoop.io.compress.BlockCompressorStream.compress(BlockCompressorStream.java:141) at org.apache.hadoop.io.compress.BlockCompressorStream.finish(BlockCompressorStream.java:135) at org.apache.hadoop.mapred.IFile$Writer.close(IFile.java:135) at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1450) at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.flush(MapTask.java:1297) at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:436) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:371) at org.apache.hadoop.mapred.Child$4.run(Child.java:255) at java.security.AccessController.doPrivileged(AccessController.java:310) at javax.security.auth.Subject.doAs(Subject.java:573) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1149) at org.apache.hadoop.mapred.Child.main(Child.java:249) So is there any way I can check if $HADOOP_HOME/lib/native/Linux-amd64-64/libhadoop.so contains the native method as expected above? This hadoop 1.1.0 is not compiled by me, but coming from IBM biginsight 2.1 as we are evaluating it. I will create a ticket for them, but is this kind of strange, as everything loaded shown in the log, but complains later about a native method? Any reason can cause this? Yong From: [email protected] To: [email protected] Subject: enable snappy on hadoop 1.1.1 Date: Fri, 4 Oct 2013 15:44:34 -0400 Hi, I am using hadoop 1.1.1. I want to test to see the snappy compression with hadoop, but I have some problems to make it work on my Linux environment. I am using opensuse 12.3 x86_64. First, when I tried to enable snappy in hadoop 1.1.1 by: conf.setBoolean("mapred.compress.map.output", true); conf.set("mapred.output.compression.type", "RECORD"); conf.set("mapred.map.output.compression.codec", "org.apache.hadoop.io.compress.SnappyCodec"); I got the following error in my test MR job: Exception in thread "main" java.lang.RuntimeException: native snappy library not available So I download the snappy 1.1.0 from https://code.google.com/p/snappy/, compile it and install it successfully under /opt/snappy-1.1.0, and then I link the /opt/snappy-1.1.0/lib64/libsnappy.so to /user/lib64/libsnappy.so Now after I restart the hadoop and tried my test MR job again, this time, it didn't give me the originally error, but a new error like this: Error: java.lang.UnsatisfiedLinkError: org/apache/hadoop/io/compress/snappy/SnappyCompressor.compressBytesDirect()I at org.apache.hadoop.io.compress.snappy.SnappyCompressor.compress(SnappyCompressor.java:229) at org.apache.hadoop.io.compress.BlockCompressorStream.compress(BlockCompressorStream.java:141) at org.apache.hadoop.io.compress.BlockCompressorStream.finish(BlockCompressorStream.java:135) at org.apache.hadoop.mapred.IFile$Writer.close(IFile.java:135) at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.sortAndSpill(MapTask.java:1450) at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.access$1800(MapTask.java:852) at org.apache.hadoop.mapred.MapTask$MapOutputBuffer$SpillThread.run(MapTask.java:1343) I wrote a test problem, like hadoop did load the library: System.loadlibrary("snappy") it works fine in my test program. I don't know why at runtime, the Class SnappyCompressor.compressByteDirect() gave back that kind of error. From the source code, it looks like a native c program from here https://code.google.com/p/hadoop-snappy/source/browse/trunk/src/main/java/org/apache/hadoop/io/compress/snappy/SnappyCompressor.java Any idea why this is happening on my environment? Thanks Yong
