Hacker News new | ask | show | jobs
by electroly 1523 days ago
They _ultimately_ didn't dump and restore, but it was the first thing they tried. It didn't work; it actually failed catastrophically for them. They describe this under the "Implementing logical replication" section. Their ultimate solution is what they tried after the dump-and-restore based DMS method failed and they took an unplanned outage due to yet more PostgreSQL-specific issues (unvacuumed tuples).

All of this is exactly what I'm talking about. This blog post describes kind of a nightmare process for something that is trivial in SQL Server. They actually needed a third party product from Citus just to successfully upgrade PostgreSQL! Stunning.
2 comments
Beltalowda 1522 days ago
I don't think they "needed" the Citus tool, per se, it was just the easiest option. I don't know much about MS-SQL, but no doubt PostgreSQL has areas that can be improved, or even that outright suck.

The main barrier against adopting MS-SQL is just the pricing and that it's not open source. Another thing that PostgreSQL seems to do a lot better than MS-SQL is in the extensibility department, hence we have things like TimescaleDB, Citus, EdgeDB, and a whole lot more. I can't really find anything like that for MS-SQL, but perhaps I missed it?

link
electroly 1522 days ago
You're absolutely right. A serious part of managing SQL Server is keeping your costs down. RDS for SQL Server is so unbelievably expensive that I can't believe anyone uses it. I'm not aware of any meaningful extensions to MSSQL in the sense of TimescaleDB and friends either. I'll make the claim that we don't need Citus because everything they offer is built into MSSQL, and we don't need TimescaleDB because column stores are built in too, but if you did find some kind of deep extension you wanted to build, you can't do it. Simply not an option with MSSQL. You either build it outside of SQL Server, or you don't build it.
link
VincentEvans 1523 days ago
Postgres dump and restore tooling is very poor performance-wise , easily 10x slower compared to SQL Server. I love Postgres dearly and prefer to use it despite that, but I wish Postgres devs renewed their interest in improving neglected dump/restore tooling.
link