DB Locking mechanism

less than 1 minute read

Pessimistic locking

pessimistic concurrency control - places a lock on a record (row) as soon as one user starts to update it

  • other users have to wait until the lock is released.

“Select for update” works by locking the rows returned

Cons : Deadlock free code is challenge

  • not scalabkle if locked for far too long

Optimistic locking

also called optimistic concurrency control

2 ways to implement : version number and timestamp

version number is more accurate as server clock can be inaccurate

3 DB Constraints

what os change data capture?

2 phase commit - db protocol, used to guarantee atomic transaction commit across multiple nodes.

Saga

You may also enjoy