juhwi lee created AIRFLOW-2751:
----------------------------------
Summary: hive to druid library conflicts
Key: AIRFLOW-2751
URL: https://issues.apache.org/jira/browse/AIRFLOW-2751
Project: Apache Airflow
Issue Type: Bug
Components: operators
Affects Versions: 1.10
Reporter: juhwi lee
There is a jackson.datatype.guava library conflict problem in
HiveToDruidTransfer when hive and druid has different library.
In my case
hdp version: 2.6.3.0
druid version : druid-0.10.1( jackson-datatype-guava-2.4.6.jar)
This is the error message I got.
2018-06-22T01:37:11,190 INFO [task-runner-0-priority-0]
org.apache.hadoop.mapreduce.Job - Task Id :
attempt_1525761839652_85792_m_000000_0, Status : FAILED
Error: class com.fasterxml.jackson.datatype.guava.deser.HostAndPortDeserializer
overrides final method
deserialize.(Lcom/fasterxml/jackson/core/JsonParser;Lcom/fasterxml/jackson/databind/DeserializationContext;)Ljava/lang/Object;
2018-06-22T01:37:15,216 INFO [task-runner-0-priority-0]
org.apache.hadoop.mapreduce.Job - Task Id :
attempt_1525761839652_85792_m_000000_1, Status : FAILED
Error: class com.fasterxml.jackson.datatype.guava.deser.HostAndPortDeserializer
overrides final method
deserialize.(Lcom/fasterxml/jackson/core/JsonParser;Lcom/fasterxml/jackson/databind/DeserializationContext;)Ljava/lang/Object;
2018-06-22T01:37:20,231 INFO [task-runner-0-priority-0]
org.apache.hadoop.mapreduce.Job - Task Id :
attempt_1525761839652_85792_m_000000_2, Status : FAILED
Error: class com.fasterxml.jackson.datatype.guava.deser.HostAndPortDeserializer
overrides final method
deserialize.(Lcom/fasterxml/jackson/core/JsonParser;Lcom/fasterxml/jackson/databind/DeserializationContext;)Ljava/lang/Object;
2018-06-22T01:37:26,248 INFO [task-runner-0-priority-0]
org.apache.hadoop.mapreduce.Job - map 100% reduce 100%
2018-06-22T01:37:26,252 INFO [task-runner-0-priority-0]
org.apache.hadoop.mapreduce.Job - Job job_1525761839652_85792 failed with state
FAILED due to: Task failed task_1525761839652_85792_m_000000
Job failed as tasks failed. failedMaps:1 failedReduces:0
We should be able to set this property to solve that problem.
"mapreduce.job.user.classpath.first" : "true"
It means "prefer loading Druid's version of a library when there is a conflict".
reference. http://druid.io/docs/latest/operations/other-hadoop.html (Tip #2)
But Job properties is hard coded in HiveToDruidTransfer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
