Over the last few days, I’ve seen more than a few comments and questions about what the MariaDB Foundation is supposed to do — and what it should do when something in the wider MariaDB ecosystem changes. I am referring here to what I call the “Galera Case“.
Instead of replying to every thread with a link to meeting minutes, I wanted to take a step back and write something more durable: a friendly reset.
Because here’s the honest part: before joining the MariaDB Foundation, I also assumed the boundary between MariaDB plc and the Foundation was… blurry. At times, it even felt inexistent. And from what I’m seeing now, I clearly wasn’t alone.
Where does the confusion come from?
MariaDB is both:
- A server project and an ecosystem, with contributors, distributions, users, and community expectations.
- A commercial company (MariaDB plc), with products, customers, acquisitions, and a business strategy.
When the company makes moves that affect parts of the ecosystem (or appear to), it’s natural that people look at “MariaDB” and ask:
“So what will the Foundation do about this?”
That exact dynamic resurfaced recently, with concerns about Galera/Codership and their potential impact on users, packaging, and long-term availability.
In fact, after acquiring Codership, MariaDB plc has decided that the future of the synchronous replication offering will be dedicated solely to customers. Those having higher requirements.
Examples of what I mean:
- Now it looks like much smaller MariaDB component vendor decided to stop shippingit as Open Source, so why is MariaDB Foundation removing it instead of picking up the slack ?
- Despite years of positioning itself as the open, independent alternative to “Enterprise” open-core, MariaDB is moving Galera Cluster behind a paywall.
- Galera dependencies are being removed even from the binaries, without a commit message or a task description.
- … and many mails or slack messages.
None of these reactions is “stupid” or “bad”. They’re a predictable outcome of people caring about continuity.
So, what is the MariaDB Foundation actually here for?
The MariaDB Foundation is a non-profit with a very explicit mission, built on three principles:
- Openness
- Adoption
- Continuity
These are the Foundation’s operating constraints.
All three of these principles are very important, but today I would like to talk about continuity.
The Foundation exists to provide continuity for MariaDB Server independent of any single commercial entity.
This principle matters especially when the ecosystem is worried about fragmentation, control, or “what happens if…”.
So, the Foundation’s mandate is MariaDB Server and its long-term openness and continuity. Does Galera belong to that scope?
I believe it does — at least the Galera library as it is distributed with the community server packages. A production-grade database needs a credible high-availability story, and I don’t think removing long-standing capabilities from community users is the right direction. That’s also why I consider “part of the server” to include what has historically shipped as part of the community packages.
As you can read in the latest Board meeting minutes, my vote joined the majority in deciding that the Foundation should coordinate, accept community contributions to, and help ensure the distribution of the current Galera (GPL) code going forward.
I’m writing this because:
- The community deserves clarity,
- The Foundation’s role is often misunderstood,
- And when uncertainty spikes, rumors and wrong assumptions fill the gap.
If we want a healthier ecosystem, we need fewer assumptions and more shared understanding of who does what — and where to look when we want facts.
The MariaDB Foundation is here to protect and grow MariaDB Server under the principles of Openness, Adoption, and Continuity—and this is exactly what we do, and we do so with a level of transparency you can verify yourself in the public minutes.
But at the same time, I do think we can and should also credit MariaDB plc for actually acquiring Codership/Galera and not just forking the code. That would have been much cheaper.