My jobs have some exception before map-reduce steps, but jobs are not getting killed. How to configure hadoop such that jobs get killed after exception?
Invoking Main class now
Heart beat Heart beat
Invocation of Main class completed
Oozie Launcher ends
stderr logs
org.springframework.jdbc.CannotGetJdbcConnectionException: Could not get JDBC Connection; nested exception is org.apache.commons.dbcp.SQLNestedException: Cannot create PoolableConnectionFactory (Io exception: Unknown host specified )
at org.springframework.jdbc.datasource.DataSourceUtils.getConnection(DataSourceUtils.java:82)
at org.springframework.jdbc.core.JdbcTemplate.execute(JdbcTemplate.java:577)
at org.springframework.jdbc.core.JdbcTemplate.update(JdbcTemplate.java:792)
at org.springframework.jdbc.core.JdbcTemplate.update(JdbcTemplate.java:815)
at com.seven.crcs.export.dao.ReportDAOImpl.recreateReportEntity(ReportDAOImpl.java:151)
at com.seven.crcs.export.dao.ReportDAOImpl.saveActiveUserCount(ReportDAOImpl.java:93)
at com.seven.crcs.export.ReportJdbcExporter.saveActiveUserCount(ReportJdbcExporter.java:55)
at com.seven.dataprocessor.oc.jobs.reports.export.day.ExportDailyUserReducer.exportUserCounts(ExportDailyUserReducer.java:32)
at com.seven.dataprocessor.oc.jobs.reports.export.ExportActiveUser
org.springframework.jdbc.CannotGetJdbcConnectionException: Could not get JDBC Connection; nested exception is org.apache.commons.dbcp.SQLNestedException: Cannot create PoolableConnectionFactory (Io exception: Unknown host specified )
And
2013-02-28 06:06:46,487 INFO org.apache.hadoop.mapred.JobClient: Task Id : attempt_201302270945_0181_r_000000_0, Status : FAILED
2013-02-28 06:07:00,600 INFO org.apache.hadoop.mapred.JobClient: Task Id : attempt_201302270945_0181_r_000000_1, Status : FAILED
2013-02-28 06:07:16,650 INFO org.apache.hadoop.mapred.JobClient: Task Id : attempt_201302270945_0181_r_000000_2, Status : FAILED
2013-02-28 06:07:31,731 INFO org.apache.hadoop.mapred.JobClient: Job complete: job_201302270945_0181
But jobs complete SUCCEEDED
Your job was actually terminated, but only after 3 failed attempts of the map task as the task ids show:
You can limit the number of maximum attempts for each task either by setting the parameter mapred.map.max.attempts to 1 or by using JobConf#setMaxMapAttempts(int)JobConf#setMaxMapAttempts.
This will cause your map task to fail on the first exception and thus terminate your job a little faster.