Under Review

Add support for MariaDB?

Hi there,

is there any plan to include MariaDB in the list of supported DBs?

As more and more Linux distributions are moving away from MySQL towards MariaDB (RHEL/CentOS 7, OpenSUSE 12.3), it would be nice to have it officially supported. Being binary compatible with MySQL, it runs very well in my test installations, offering even some of the performance options Oracle wants you to pay for (thread pooling anyone?), so any plans on this?

  • *Bump*

    I can understand the implications of supporting more than one database, but the default mysql client on some Linux OS (Debian Stretch for example) is MariaDB. I certainly hope this can still be considered for the future.

  • Hi Angel,

    this is new information for me, thanks for posting. As it seems, MariaDB also needs some income, and this somehow makes it sour again.

    I see now this might be less of a must-have. Please also see my other ieda for supporting Percona Server.



  • That assumes that MariaDB would indeed outperform MySQL, and to a point significant enough to warrant all the testing work that has been mentioned.

    As of today, I would argue that is not true, hence my earlier comments about some scenarios I have personally seen where performance of some queries is worse on MariaDB than on MySQL.

    To your point about Oracle stripping things out, MariaDB is actually the one already doing this, see here: Uproar: MariaDB Corp. veers away from open source | InfoWorld 

    There is no guarantee they wouldn't expand their strategy to encompass other elements of MariaDB either.

  • Hi Joe Mao,

    while I understand your point, you are missing one thing. A better performing database reduces load times and creates a better user experience. Also, as MySQL in the open source version is more and more stripped of features by Oracle, customers on a budget run out of options for a well-performing environment.

    MS SQL is to slow inthe Express variant once you crank up the load, and commercial just shoots big holes in your wallet (same for Oracle 11/12 or DB2) and forces you to host SugarCRM on Windows Servers.

    Postgresql is out of reason following your logic. What is left? Maybe officialy supporting Percona's variant of MySQL is an option (this sometimes comes up when discussing large scale deployments with SugarCRM tech support)

    Any thoughts?



    EDIT: Create the idea to officially support Percona Server:

    Official support for Percona Server for MySQL 

  • Adding more databases only will add addition testing complexity and time on top of an already large matrix. In addition, it doesn't have an end user benefit.

    In the end, the question is what features or bugs are will you give up if it says takes an extra x weeks of testing per release. It's like me asking you to take on another new project without more hours of the day. 

    It does not seem it would accelerate releases, improve sales or improve better experiences for our customers.