[ruairispain]

Personally, I think MySQL died when Oracle got involved with it. The momentum was sucked right out of the project.

The great think about Postgres is it has huge momentum and the team are delivering new features. Maybe Apache has less momentum because it's matured and doesn't need much momentum.

From a momentum point of view I see Postgres as the biggest FLOSS project. I see no point in choosing another RDBMS right now. Time to short ORA stock

[pizza234]

> Personally, I think MySQL died when Oracle got involved with it. The momentum was sucked right out of the project.

This statement is completely baseless, and you clearly haven't followed the MySQL developments of the last 10 years. It's ironic that around the time Sun was acquired, MyISAM was still a thing; MySQL has gone a long way since then.

Independently of a value judgement of "how good PGSQL/MySQL are" (I'm not making any implication in this sense), MySQL's development has been proceeding steadily, with no particular change (on the engineering side, at least). Oracle has actually put considerable effort both on the engineering and marketing side, of the 8.0 release (but again, I don't make any implication about the value of the products per se).

[jpitz]

As much as I vastly prefer PostgreSQL, I will tell you that MySQL is much more preferred in enterprise settings, probably 8 to 1 in the environments I've seen.

[dragonwriter]

> As much as I vastly prefer PostgreSQL, I will tell you that MySQL is much more preferred in enterprise settings, probably 8 to 1 in the environments I've seen.

That's largely because enterprises often have big investments in SQL server, OracleDB, and/or DB2, and are only using open source engines for more lightweight purposes, and/or as part of cloud transitions where they are just taking vendor default options or whatever options was supported when they came on or longest.

At least, that's my experience working in enterprise and being literally the single voice urging even considering pros and cons before using MySQL-by-default with no particular rationale in a transition effort (which resulted in us using Postgres.)

[GordonS]

I work as a consultant in the enterprise space, and I've seen a shift in recent years towards Postgres.

I believe this has largely been driven by cloud - developers are now more able to choose the components they want to work with, instead of being told what to work with. There are of course plenty enterprises that are strict here, mandating cloud hosted SQL Server for example, but the general relaxation of constraints has struck me as a very pleasant surprise.

I've yet to come across MySQL being used in a cloud-based system, but I'm seeing Postgres more and more. When I do see MySQL, it's part of on-prem services that are considered "legacy".

[jpitz]

One of my contacts is migrating databases between clouds, with MySQL being the most common, followed by sql server, then PostgreSQL.

[brlewis]

That's likely a holdover from when mysql had a better replication story than postgresql. I think enterprises will come around.

[throw0101a]

> That's likely a holdover from when mysql had a better replication story than postgresql.

What is the Galera-equivalent in the Postgres world? There's BDR, but the latest versions are closed source.

We use Galera in a bunch of places because it's fairly straight forward to get an HA cluster going, and with keepalived, we can point a front-end to a vIP that fails over automatically if one system goes sideways.

[cookiecaper]

afaik, MySQL never had a better replication story than PostgreSQL, you've always had to use Percona's add-ons to get something workable. Even then (many years ago), we had multiple data loss incidents that were precipitated by widespread internal confusion over the bizarre intricacies of `binlog_format`.

Like most software that gains adoption, MySQL made some very broad claims about their software's capabilities and never really delivered on them, at least not in a way that would be considered production-ready by Serious Persons(TM).

That's the crux of MySQL v. PgSQL, Linux v. BSD, etc. Good engineers spend their time building good software and are generally too focused on that to spend much time going around and making outlandish promises. Postgres has benefited from Oracle's intentionally-bad stewardship over MySQL, but it doesn't usually work that way.

[dpcx]

I've experienced this as well, and it's almost always because the tooling for MySQL is "better." I don't necessarily agree, but Sequel Ace (formerly Sequel Pro) is hard to beat.

[Izkata]

> "better."

Inertia is what I've noticed as well. The few people I've convinced to actually try out postgres have ended up liking it more than mysql.

[rguldener]

TablePlus is a great alternative for Postgres (and many other DBs) for that on Mac & Linux IMO: https://tableplus.com/

[apazzolini]

Oh how I wish Sequel Pro/Ace worked on Postgres.

[mustardo]

The DB support in Jetbrains IDEA is awesome, they also have it as a stand alone IDE called Data Grip, could be woth a shot? https://www.jetbrains.com/datagrip/

[Shorel]

May be in your small circle.

In AWS, Aurora MySQL is almost always the database of choice, for just about everything SQL.

[cookiecaper]

Do people ever look at how much more they're paying for Aurora? I've saved companies thousands of dollars per month just by switching them to standard RDS instances for systems that see tiny load and were only on Aurora because that's "the super good AWS-optimized version, duh".

The added benefit of this is that you then know what software you're running, instead of Amazon's fly-by-night hackjob. We've hit several real-world bugs that were either incorrect backports of something that had been patched years before upstream, or wholly new issues caused by whatever they're doing to make Aurora auroar.

[petergeoghegan]

Amazon themselves use Aurora PostgreSQL for Amazon retail fulfillment databases:

https://aws.amazon.com/solutions/case-studies/amazon-fulfill...

[makoz]

Source?

[tuatoru]

I first looked at MySQL back in 2000. At the time it couldn't even do row-level locking. I basically wrote it off at that point.

When I looked at Postgres, slightly earlier, I thought the same as you: how can anyone charge money for an RDBMS when this exists?

[njharman]

ORA is a services/consulting company, now. The DB is important but company will easily survive its death (which will never happen) look at IBM big iron for similar story.

[arthurcolle]

*ORCL