01-09-2018 02:51 AM - last edited on 01-09-2018 05:45 AM by cjervis
I am trying to load data from SQL Server into Hive in Hue using Sqoop 1. The following statement is working:
import --connect jdbc:sqlserver://xxx:1433;database=INFA_SOURCE --username infa_source --password B1source --table Personen -m 1
However when I add --hive-import it's not working any more. There are two jobs running and both have status succeeded, but workflow is in status killed. In the log file of first job I can see that there is a failure at the end, but it's not specified what exactly failed:
2018-01-09 10:59:10,544 [main] INFO org.apache.sqoop.manager.SqlManager - Executing SQL statement: SELECT t.* FROM [Personen] AS t WHERE 1=0 2018-01-09 10:59:10,569 [main] INFO org.apache.sqoop.hive.HiveImport - Loading uploaded data into Hive <<< Invocation of Sqoop command completed <<< Hadoop Job IDs executed by Sqoop: job_1515423392261_0009 Intercepting System.exit(1) <<< Invocation of Main class completed <<< Failing Oozie Launcher, Main class [org.apache.oozie.action.hadoop.SqoopMain], exit code  Oozie Launcher failed, finishing Hadoop job gracefully
As the same statement works via command line, I think the issue is with Hue user that doesn't have access to /user/hive/warehouse folder, because of a sticky bit.
I found information about enabling security with Sentry, but I am not sure if that's really the issue and if this could be a solution?.
I have mainly two questions:
- Can I remove sticky bit to test the --hive-import with no issues and then bring it back?
- If the issue is with priviledges, is Sentry a possible solution?
Please let me know if you see here another problem or how else I could check it :)