Application team was executing UPDATE statement in DB through Abinitio graph. When they trigger a job, a session is spawned in DB, in parallel another session is also spawned and executed the same UPDATE statement.
When I checked the locks in DB, I found that both the sessions are updating the same record. My concern is how UPDATE causes locking in DB.
Here is how MVCC works. If one session is updating a record, it should release the lock once it updated the row and other one should be able to acquire the row lock. Maybe I am wrong, please suggest how to handle this situation.
Do the applications use implicit autocommit, or do they use explicit transaction blocks?
