6 ms·
I understand optimistic concurrency, especially the constraints it places on your code: you must in all cases access the data in the same order, otherwise you d
by stickfigure 9y ago
I understand optimistic concurrency, especially the constraints it places on your code: you must in all cases access the data in the same order, otherwise you do get deadlocks.
The second half of this statement negates the first :-)
You have confused pessimistic concurrency with optimistic concurrency. An optimistic system detects collision at commit time and the actual order of operations in your transaction is irrelevant. You will never experience a deadlock, only aborts/retries.
In a pessimistic concurrency scenario, you take explicit locks as you access resources. Those locks block access by other transactions. This creates the dining philosophers problem and the risk of deadlocks. When locking resources, you must access resources in a consistent order to prevent deadlocks.
Please avoid database locks and let smart databases do what they are supposed to!
- IgorPartola 9y agoLearn something new every day. Yes you are correct, I meant pessimistic concurrency because that's what I'm familiar with. Which databases support optimistic concurrency and full ACID transactions?
- stickfigure 9y agoPostgres, MS SQL Server, Oracle... and more exotic ones like Google's Cloud Datastore and Spanner. Probably many more.
- IgorPartola 9y agoInteresting. I use Postgres and can't find any mention of this in the docs: https://www.postgresql.org/docs/current/static/mvcc.html https://www.postgresql.org/docs/current/static/mvcc.html talks about the usual row and table locks, and https://www.postgresql.org/search/?u=%2Fdocs%2F9.6%2F&q=Optimistic https://www.postgresql.org/search/?u=%2Fdocs%2F9.6%2F&q=Opti... is empty. Is it known by some other name in the Postgres land? Also, I found this and similar on SO: https://stackoverflow.com/questions/17431338/optimistic-locking-in-mysql https://stackoverflow.com/questions/17431338/optimistic-lock... So if this technique actually requires me to write my database access code at this level, it means that using something like an ORM is nearly impossible in a way that makes an ORM useful.
- wbl 9y agoThat's how isolation is implemented. MVCC is not compatible with any standard SQL level, but it is used for the lower ones.
- stickfigure 9y agoThat stackoverflow answer is wrong - or at least, only applies to MySQL. The author is apparently unaware of other databases. If you put Postgres in repeatable read or serializable[1] mode, it manages optimistic concurrency for you. Collisions will produce an error in the second transaction attempting to commit. The only "trick" is that your client code should be prepared to retry transactions that have concurrency failures, and this is very easily abstracted by a library. Your SQL or ORM code does not change (I use Hibernate, fwiw). Some good reading: https://www.postgresql.org/docs/current/static/transaction-iso.html#XACT-SERIALIZABLE https://www.postgresql.org/docs/current/static/transaction-i... [1] Repeatable Read won't detect collisions caused by phantom rows, so pick carefully. Most CRUD apps don't care about phantom rows.
- IgorPartola 9y agoHow does this work if your transactions also do external things? So if have two concurrent transactions that charge a credit card via an external API, as well as mark the order as paid in your DB? Retrying that transaction would be disastrous, no?
- stickfigure 9y agoThis is a problem that affects all distributed systems irrespective of database isolation levels. For example, your call to the external API might timeout; did the charge succeed or not? Database locking doesn't help you. The solution comes in two parts: 1) Ensure your call to the card processor is idempotent and retry it. For example, grep Stripe's documentation for the "Idempotency-Key" header. Every processor should have a similar mechanism. Conveniently, if your database transaction unit-of-work includes the remote call, it will retry automatically in an optimistic scenario. 2) Get a webhook call from the credit card processor. Even though you are retrying your transaction, you can't guarantee that your server didn't crash without a successful commit. This potentially leaves a charged card without recording the purchase; the webhook will sync it up afterwards. Distributed systems are hard if you think past the happy path. I wish every REST API would have an equivalent of Idempotency-Key.