[OOZIE-3134] Potential inconsistency between the in-memory SLA map and the Oozie database

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 5.1.0
• Assignee: Kinga Marton
• Reporter: Attila Sasvári
• Created: 23/Nov/17 16:42
• Updated: 18/Jan/19 09:02
• Resolved: 05/Apr/18 08:11

Description

Upon SLACalculatorMemory.addRegistration, Oozie puts an entry into an in-memory concurrent hashmap ("slaMap"), and later persists it in the Oozie database.

However, if there is a failure during the database operation, a JPAExecutorException is thrown, and the entry is not removed from the SLA map.
It may introduce inconsistency between the Oozie database and the SLA map.

To prevent this, a rollback mechanism (with proper logging) should be implemented. It would also make sense to do more sanity/consistency check in the Oozie server.

Issue Links

• is related to: OOZIE-3260 [sla] Remove stale item above max retries on JPA related errors from in-memory SLA map
• is related to: OOZIE-1980 Sql error should not fail coord job
• links to: Review board