Proposed Fedora 19 Feature: Replace MySQL with MariaDB | Hacker News Reader

Proposed Fedora 19 Feature: Replace MySQL with MariaDB

lists.fedoraproject.org

42 pointsIgorPartola27 comments

Kudos to Fedora for making such a bold move. Oracle's commitment to MySQL is questionable. So, even though MariaDB does not have MySQL's 'brand' recognition, it's completely backward compatible with MySQL (at least till now). And I think most people can just switch to MariaDB without any issues.

I run mariadb on my Mac instead of MySQL. It's been fine as a transparent replacement thus far on my contracting gigs.

My apologies. I'm starting to feel like Oracle's becoming to the 33 years old me what Microsoft was to this zeal... - erm - open-source enthusiast 10 years ago.

Even recognizing this similarity and although I know that it is childish and stupid, I cannot help to wear a fat smile when Oracle hits the news in a bad way, recently. And in moments like this I'm proud of the fact that my native language 'invented' Schadenfreude.

They just (subjectively) messed up so much. OpenOffice, Java failures, MySQL - the only news in a long time about that company that weren't about messing up software involved Larry buying a private island..

This has also been considered in Ubuntu: "MySQL's future in Debian and Ubuntu" https://lists.ubuntu.com/archives/ubuntu-server/2012-Februar... and an intermediate conclusion: https://lists.ubuntu.com/archives/ubuntu-server/2012-Februar...

Also see "Oracle Makes More Moves To Kill Open Source MySQL": http://techcrunch.com/2012/08/18/oracle-makes-more-moves-to-...

A similar proposition has been put forward on the openSUSE development mailing list:

http://lists.opensuse.org/opensuse-factory/2013-01/msg00298....

That's good move. OSS software should be written only by passionate engineers for self-realization, not by [outsourced] corporate drones for salary and position.

So , since it seems that MariaDB is the new "real" mysql is there any good reason now to do this?

On a more practical note, if I have production servers running mysql should I in theory at least be able to just drop in MariaDB as a replacement with the same config files?

Speaking of drop in replacements, I've been very happy with Percona Server. The improved monitoring is worth it alone.

Agreed - Percona Community Edition is pretty dang good.

I poked around a bit before picking it. The monitoring is what sold me -- I don't do anything like enough traffic for the performance stuff to affect me.

I looked at MariaDB, Percona and Drizzle. MariaDB is very close to stock MySQL, Percona is stock + lots of performance patches and monitoring.

Drizzle was still in the middle of proving Fredrick Brooks right about second systems.

Understanding the drop-in aspect, I'm curious if you compared those same three against Postgres?

I'm contemplating which DB to move to for a project. I know Mysql and Postgres pretty well.

Postgres is hard to beat right now in terms of the community behind it and the rate of innovation. Everyone is using it in exciting new ways. Hstore is awesome. Postgis is awesome. Async indexing. Heroku Postgres. The list goes on and on. That being said, most of the big projects that I've worked on in my lifetime were on MySQL and it fared very well. Mariadb and Percona are both solid. At some point all of the MySQL variants are still MySQL though... Which means a lot of its inherent flaws are still there. At the end of the day it's not tremendously hard to switch from one to the other. See the recent posts from the Lanyrd dev team on migrating to Postgres.

I greatly prefer PostgreSQL. But it's for a Wordpress installation.

I believe the next version of WP is supposed to move away from a MySQL requirement by way of finishing their database abstraction. If I wasn't half asleep, on cold meds, and typing from my phone I'd validate that and share a link. I'm pretty sure it's in their near future though. Something in regards to PDO is ringing a bell.

Plugins and themes assume MySQL. Not all of them go through the Wordpress interface to the database.

They aren't plugins/themes you want to be using than...

The same could be said about blog engines that rely exclusively on using MySQL.

I've only used MySQL in jobs, but for a personal project I recently changed from MySQL to Postgres just as a learning exercise and because everyone has such great things to say about it. I'm very excited about learning more.

I've often used both for different projects, depending on the wrappers provided.

Most of the PHP CMS world was all about MySQL first and foremost, which isn't a bad thing. But, if you ran into a MySQL limitation, you were pretty much stuck with trying to work around it -- see some of the notes about WP.

Moving to an ORM layer or a level (or two) up from direct DB calls can result in more generic SQL and potentially a performance impact.

My view is that these days, DBMS vendor lockin just isn't a problem if you're using an open source database.

So you might as well take full advantage of the features that are available.

Architecture is a big part of it IMHO.

Products exist like Oracle MySQL cluster or Percona XtraDB which allow you to do scale horizontally e.g. multi-master and have good support as you do grow. Nothing really exists for PostgreSQL on either front. But if you are planning to scale vertically then PostgreSQL seems the clear winner.

MySQL didn't become popular because it had multi-master.

It got multi-master after it was popular.

Who ever said it did ?

I am just saying is it important to consider when evaluating databases.

Sorry, for some reason I was under the impression that you were saying MySQL was popular because of its replication/HA features. On second reading, I'm wrong.

Does MariaDB still use the InnoDB engine for real work?

Yep.

Ermm, doesn't both mariaDB and Percona use Percona's drop in replacement XtraDB to power innoDB instead by default? I think if you want to use innoDB (as in the original code) you have to explicitly enable it as a plugin.

Reply on news.ycombinator.com