Rajesh Kartha created AMBARI-10573:
--------------------------------------
Summary: During installs set default value of the yarn property
'yarn.app.mapreduce.am.env' to be same as 'mapreduce.admin.user.env'
Key: AMBARI-10573
URL: https://issues.apache.org/jira/browse/AMBARI-10573
Project: Ambari
Issue Type: Bug
Components: ambari-server
Affects Versions: 1.7.0
Reporter: Rajesh Kartha
We recently ran into an issue where after successfully creating HFile or SEQ/RC
files with Snappy Compression in MR, the job fails job commit with:
java.lang.UnsatisfiedLinkError:
org.apache.hadoop.util.NativeCodeLoader.buildSupportsSnappy()Z
at org.apache.hadoop.util.NativeCodeLoader.buildSupportsSnappy(Native Method)
at
org.apache.hadoop.io.compress.SnappyCodec.checkNativeCodeLoaded(SnappyCodec.java:63)
at
org.apache.hadoop.io.compress.SnappyCodec.getDecompressorType(SnappyCodec.java:192)
at org.apache.hadoop.hive.ql.io.CodecPool.getDecompressor(CodecPool.java:122)
at org.apache.hadoop.hive.ql.io.RCFile$Reader.init(RCFile.java:1518)
Incidentally the job commit was run as part of in the MRAppMaster which did not
have the default 'yarn.app.mapreduce.am.env' set.
Here is a simple repro using a uberized task:
$ hadoop jar hadoop-mapreduce-examples.jar wordcount
-Dmapreduce.job.ubertask.enable=true -Dmapred.output.compress=true
-Dmapred.output.compression.codec=org.apache.hadoop.io.compress.SnappyCodec
/tmp/word_count.csv /tmp/wc-error
which would fail with:
2015-04-17 09:40:10,075 FATAL [uber-SubtaskRunner]
org.apache.hadoop.mapred.LocalContainerLauncher: Error running local (uberized)
'child' : java.lang.UnsatisfiedLinkError:
org.apache.hadoop.util.NativeCodeLoader.buildSupportsSnappy()Z
at org.apache.hadoop.util.NativeCodeLoader.buildSupportsSnappy(Native
Method)
at
org.apache.hadoop.io.compress.SnappyCodec.checkNativeCodeLoaded(SnappyCodec.java:63)
at
org.apache.hadoop.io.compress.SnappyCodec.getCompressorType(SnappyCodec.java:133)
at
org.apache.hadoop.io.compress.CodecPool.getCompressor(CodecPool.java:148)
at
org.apache.hadoop.io.compress.CompressionCodec$Util.createOutputStreamWithCodecPool(CompressionCodec.java:131)
at
org.apache.hadoop.io.compress.SnappyCodec.createOutputStream(SnappyCodec.java:98)
at
org.apache.hadoop.mapreduce.lib.output.TextOutputFormat.getRecordWriter(TextOutputFormat.java:136)
It would be desirable to have the yarn property 'yarn.app.mapreduce.am.env' set
to be same as 'mapreduce.admin.user.env' by default during installs so such
tasks would not fail.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
