Dear Kylin developers,
I have a cluster environment (3 zookeeper machine, one yarn RM and one
backup RM, 3 node managers) and install hive-hcatalog on all these nodes
under the same directory. While building the sample cube I got the error
at step 2:
Caused by: java.lang.NoClassDefFoundError:
org/apache/hive/hcatalog/mapreduce/HCatInputFormat
at
org.apache.kylin.job.hadoop.cube.FactDistinctColumnsJob.setupMapper(FactDistinctColumnsJob.java:101)
at
org.apache.kylin.job.hadoop.cube.FactDistinctColumnsJob.run(FactDistinctColumnsJob.java:77)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:84)
at
org.apache.kylin.job.common.MapReduceExecutable.doWork(MapReduceExecutable.java:120)
at
org.apache.kylin.job.execution.AbstractExecutable.execute(AbstractExecutable.java:107)
... 6 more
I see that the mapred and hadoop application classpath both have this
jar: hive-hcatalog
g/share/hcatalog/hive-hcatalog-core-0.13.1.jar. The path is correct. I
do not know why the class FactDistinctColumnsJob.setupMapper cannot find
it. The suggested answers I found from Kylin mails to the users do not
help me.
This actually happened in the setMapper function, so I guess Kylin looks
for the missing class only on the machine where I started Kylin (the
same machine where Hive and Hbase run).
There is a message before the error:
[pool-7-thread-2]:[2016-03-02
17:10:20,450][INFO][org.apache.kylin.job.hadoop.AbstractHadoopJob.cleanupTempConfFile(AbstractHadoopJob.java:401)]
- tempMetaFileString is : null
But I think this is not related.
You said that hcatalog must be available on all machines. Does Kylin
read any configuration files to get the machines?
I worked already two days on this problem but cannot solve. I need your
help.
Thank you in advance.
Best regards,
Jie
