zlacker

>>andren+(OP)
So why Postgres chooses the worst MVCC design compared to MySQL and Oracle? Is this because of legacy reasons or other factors?

>>Ym9oYW+c4
Legacy reasons. The idea was that you wouldn't need a WAL because the table itself is the log. And then you could support time-travel queries if you never cleaned up the expired tuples.

>>apavlo+w7
Is MVCC actually superior by some other considerations? Less lock contentions, transactional DML.

>>riku_i+1u
The problem is not mvcc but postgres’ implementation details of it.

>>maskli+sP
In what way? I didn't see anything obviously improper when I learned how serialization isolation worked.

>>Max-Ga+TP
TFA covers its issues? It’s specifically compares postgres’ to other implementations’ (oracle and mysql).

>>maskli+PK1
Also the paper linked from TFA goes into the various implementation options in more detail: https://db.cs.cmu.edu/papers/2017/p781-wu.pdf