News for hackers. Stuff that matters.

A comment on SQL Maxis: Why We Ditched RabbitMQ and Replaced It with a Postgres Queue. Parent comment

Re:SQL Maxis: Why We Ditched RabbitMQ and Replaced It with a... by foldr on Tuesday April 11, 2023 @01:33PM (#35528366)
A simple implementation of a queue in SQL will need to acquire an exclusive lock on the table anyway. Although it's not necessary to use locking at the level of SQL itself: https://news.ycombinator.com/item?id=27482402
[ Parent | on HN ]
Re:SQL Maxis: Why We Ditched RabbitMQ and Replaced It with a... by TedDoesntTalk on Tuesday April 11, 2023 @05:25PM (#35531792)
The article mentions row-level locking, not full table locking. Big difference in write performance.
[ Parent | on HN ]
Re:SQL Maxis: Why We Ditched RabbitMQ and Replaced It with a... by yawaramin on Wednesday April 12, 2023 @01:03AM (#35535225)
For a client-server RDBMS with writes over the network, sure. For an SQLite database with writes to the local filesystem and an order of magnitude better write performance, not so much.
[ Parent | on HN ]