David Wannemacher created OOZIE-1399:
----------------------------------------
Summary: Reliability and retry for metastore calls
Key: OOZIE-1399
URL: https://issues.apache.org/jira/browse/OOZIE-1399
Project: Oozie
Issue Type: Improvement
Components: core
Affects Versions: 3.2.0
Reporter: David Wannemacher
Oozie needs to have some mechanism to retry failed metastore queries and config
parameters to govern it. One example is in ActionStartXCommand. An action can
be successfully submitted and completed, which results in an output directory
being created. But when the metastore call to update the job/action fails, the
action gets repeatedly retried and fails due to the output directory already
existing.
These problems are exacerbated when using a metastore (SQL Azure in this case)
that occasionally closes connections due to idleness, throttling, or regular
server maintenance.
Hive has a retry mechanism for metastore calls which can be configured with the
hive.metastore.ds.retry.attempts and hive.metastore.ds.retry.interval
properties. Oozie should have a similar retry mechanism.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
