This is the documentation for CDH 5.0.x. Documentation for other versions is available at Cloudera Documentation.
Apache Oozie Known Issues
— Oozie HA does not work properly with HCatalog integration or SLA notifications
This issue appears when you are using HCatalog as a data dependency in a coordinator; using HCatalog from an action (for example, Pig) works correctly.
Bug: OOZIE-1492
Severity: Medium
Workaround: None
— Oozie can't submit jobs to a secure MRv2 cluster if the Job History Server is down
After trying to submit the job (by default, three times), Oozie will SUSPEND the workflow automatically.
Severity: Low
Workaround:When you bring up the Job History Server, use the resume command to tell Oozie to continue the workflow from where it left off.
— An Oozie server works either with a Hadoop MRv1 cluster or a Hadoop YARN cluster, not both
Severity: Low
Anticipated Resolution: None planned
Workaround: Use two different Oozie servers
When you use Hive Server 2 from Oozie, Oozie won't collect or print out the Hadoop Job IDs of any jobs launched by Hive Server 2
Bug: None
Severity: Low
Workaround: You can get the Hadoop IDs from the Resource Manager or JobTracker.
Log streaming does not work when both Oozie HA and Kerberos are enabled
Trying to get the logs for an Oozie job via the Oozie CLI, REST API, Web UI, or Hue will not work with both Oozie HA and Kerberos enabled; it does work when only one is enabled. Logs for Hadoop jobs and the "launcher job" are still available from Hadoop.
Bug: HADOOP-10158, HADOOP-10322
Severity: Medium
Workaround: None. You can get Oozie logs directly from the local filesystem of the Oozie server(s).
| << Cloudera Impala Known Issues and Workarounds | Apache Pig Known Issues >> | |