[FINERACT-859] Optimistic locking errors were detected when flushing to the data store

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Assignee: Unassigned
• Reporter: Michael Vorburger
• Created: 08/Mar/20 21:25
• Updated: 15/Jan/21 20:21

Description

The integration test logs sometimes show:

Optimistic locking errors were detected when flushing to the data store. The following objects may have been concurrently modified in another transaction: ...

FINERACT-857 has one (of many, this happens fairly regularly..) example logs of this happening. (FINERACT-858 will make the real root cause more clearly stand out in future logs.)

I think we need to catch and handle this problem with explicit retries in the code. Perhaps we can draw inspiration from related work I've done in a previous life, from this code:

• https://github.com/opendaylight/genius/blob/master/mdsalutil/mdsalutil-api/src/main/java/org/opendaylight/genius/infra/RetryingManagedNewTransactionRunnerImpl.java
• https://github.com/opendaylight/genius/blob/master/mdsalutil/mdsalutil-api/src/main/java/org/opendaylight/genius/infra/RetryingManagedNewTransactionRunner.java#L38
• NOT https://github.com/opendaylight/genius/blob/master/mdsalutil/mdsalutil-api/src/main/java/org/opendaylight/genius/datastoreutils/TaskRetryLooper.java

BTW: This locking issue is probably (but I'm not 100% sure..) completely unrelated to the other one tracked in FINERACT-856.

Issue Links

• blocks: FINERACT-857 integrationtests.SchedulerJobsTestResults
• blocks: FINERACT-1086 JpaOptimisticLockingFailureException at LoanSchedularServiceImpl.recalculateInterest() - requires retry logic?
• blocks: FINERACT-683 Concurrent withdraw/deposit transactions to savings account fail with locking errors
• is blocked by: FINERACT-858 Errors (JobExecutionException) in Cron Jobs loose exception root cause details
• is related to: FINERACT-924 SchedulerJobsTestResults.testApplyAnnualFeeForSavingsJobOutcome