MariaDB ServerFest and Contribution With Kaj Arnö, CEO at MariaDB Foundation - Percona Podcast 44

Link to listen and subscribe: PodBean

The CEO of the MariaDB Foundation, Kaj Arnö, joined Matt Yonkovit, The Head of Open Source Strategy at Percona, to talk about all things MariaDB related! Kaj and Matt also talk about the value and process behind getting contributions from the open source community, and why code is not the only way to contribute. Kaj also got to also talk about all the highlights from MariaDB Server Fest including key speakers and new changes to the MariaDB release process. We hope you enjoy the latest episode of the HOSS talks FOSS!

YouTube

Link: https://youtu.be/4BTIbMN81oQ

Kaj Arnö

MariaDB Foundation, CEO

Kaj Arnö is CEO of the MariaDB Foundation. He is a software industry generalist, having served as VP Professional Services, VP Engineering, CIO and VP Community Relations of MySQL AB prior to the acquisition by Sun Microsystems. At Sun, Kaj served as MySQL Ambassador to Sun and Sun VP of Database Community. Initiator of Projekt Fredrika, about Swedish Finland on Wikipedia. Past founder, CEO and 14 year main entrepreneur of Polycon Ab (Finland). Kaj is a co-founder of MariaDB Corporation Ab, and served on its Executive Team in several positions, most recently Chief Evangelist. Kaj lives and works in Munich, Germany, and during summer often in Nagu, Finland. He is a citizen of both countries. Kaj holds a M.Sc. (Eng.) from Helsinki University of Technology (now Aalto University) and an MBA from Hanken School of Economics. He tweets @kajarno.

See all talks by Kaj Arnö »

Matt Yonkovit

The HOSS, Percona

Matt is currently working as the Head of Open Source Strategy (HOSS) for Percona, a leader in open source database software and services. He has over 15 years of experience in the open source industry including over 10 years of executive-level experience leading open source teams. Matt’s experience merges the technical and business aspects of the open source database experience with both a passion for hands on development and management and the leadership of building strong teams. During his time he has created or managed business units responsible for service delivery ( consulting, support, and managed services ), customer success, product management, marketing, and operations. He currently leads efforts around Percona’s OSPO, community, and developer relations efforts. He hosts the HOSS talks FOSS podcast, writes regularly, and shares his MySQL and PostgreSQL knowledge as often as possible.

See all talks by Matt Yonkovit »

Transcript

Matt Yonkovit
Hi, everybody. Welcome to The HOSS talk FOSS. I’m the HOSS Matt Yonkovit. I’m here with Kaj Arnö from the MariaDB Foundation. Hi Kaj, how are you this morning?

Kaj Arnö
Hi Matt! I’m fine. How are you?

Matt Yonkovit
Good! Good. So Kaj is the CEO of the MariaDB Foundation. You’ve been CEO there for several years now. You just had a new addition to the family?

Kaj Arnö
I did. I did. Yes. So there was a new production release spontaneously. As things happen, things happen. So I got the son on the ninth of September. And that’s a great addition to the family. I have had two children before, but that’s more than a quarter century ago. So this is a very new release with a low cadence between the releases.

Matt Yonkovit
Yeah. And so if it’s September, you are now in that phase where you’re probably getting a little sleep.

Kaj Arnö
Right? Yes. But perhaps we should talk about something else than the amount of sleep. But yes, you’re right.

Matt Yonkovit
Well, congratulations on that. I know, it’s always great to have a new addition, a new release to the family. So I just wanted to congratulate you on that. But you also just went through another momentous occasion at Maria, you had a Maria DB server fests recently?

Kaj Arnö
Yes, we did yesterday. So we realized, like a good year ago, that our way of talking to our audience does not work during the pandemics. And last year, we had our first server fest, and then we had a couple of mini fests on some topics. But then every year, we think there should be a proper server fest where the only common denominator between the different talks is just MariaDB, so people can submit anything we just had in the beginning of the month.

Matt Yonkovit
And so how was it? How was attendance? Good how did that go this year.

Kaj Arnö
So I think my observation from this year is that the quality was great. The speakers were fabulous, and we got some good feedback on it. But the amount of attendees was not the same as last year. So last year, it was our first ever service that we had only done if we went to conferences, now those unconferences were preaching to the choir, less than 100 people, and those who contribute code to the server. And none of those had more than 100. But last year, we had more than 10,000. So it was a 100x growth in the amount of attendees. And because we had no experience organizing events for hundreds of people physically and now this has become a virtual event, we had one enormous success last year. And I think that it was part luck and proper planning. And the law, of course, was that people were starting last year for contact for getting interaction for forgetting news. And this year, people are more accustomed to how to do online conferences, and they are a bit more picky. And there wasn’t this in such a world hunger, so our numbers are not as big as they were last year. But the golden we were glad this happened with us whoever the year before

Matt Yonkovit
Yeah, it’s tough with a lot of the open source projects this year with virtual content, I’ve been to a speaker on several, we’ve done a few. And it seems that every time you get better at delivery, your audience kind of gets a little shy about coming to some of these. And so that’s typical, from a, from an open source perspective, because you want to reach as many people but typically, it’s good when you have that content available that people can go get on demand, because I see even for a lot of the content that I do, I might have 10% 20% live, and then I get 80% the next month, right of viewer, so

Kaj Arnö
I think that’s a good observation, this live thing. So why is it important to live? Nobody says that life as such is really more relevant. It feels exciting, and it’s happening just now. And I was there when it happened, and I heard it, first news. But if you really think about it, does it make any sense? It doesn’t, I think I’m more interested in what’s important and what’s relevant and if I get to hear about it a week later or a month later, I probably will be as efficient. And at least if I get, I get to reduce the amount of spammy and things that happen live because of the microphone and all knowing circular will work, we will have to arrange these technical matters, all of those, then you get a much higher signal to noise ratio. So I don’t think that what is happening right now is so important.

Matt Yonkovit
Yeah, and I mean, we’ve just started to do a few in person conferences. And I don’t know if you’ve been able to attend any yet. I’ve done two in the last three weeks. And few virtuals

Kaj Arnö
Physically present again!

Matt Yonkovit
it’s, it’s interesting. So we’ve also had some people go to cubecon. So there was the OSS Summit. So the open source Summit, put on by the Linux Foundation, and the cubecon the week after two weeks after, and then all things open. And so that all three open source focused conferences, interrupts, and attendance was about 20% of what it would normally be in a normal year, they went hybrid. I’m not sure about the hybrid component and how many people tuned in? We didn’t get a lot of questions in the sessions that I did. So I’m not sure if there were many people who have been there, but it was good to see people and there was definitely a thirst for people in person, they wanted to come by and talk to someone, which was good. And I think that a lot of people enjoyed that. Now, it’s interesting, depending on where in the country in the US you are, and it’s probably the same thing worldwide, you’re going to have a very different experience. So I know, high loads coming up in high load in Russia, they’re probably expecting close to probably 80% 90% of what they normally get, which is very, very large crowds. But here, it might be 20%, somewhere else in Europe might be 50%. It just depends on where you are, and how comfortable people are getting together in person

Kaj Arnö
That will be in Moscow for a high load, I would assume that better luck, people don’t have to travel that far that said about traveling inside Moscow, kind of take his time as well.

Matt Yonkovit
Yeah. But now, I’m looking forward to getting in person and being able to see sessions again, and say hello in person. But did you have something that really stuck out in your mind during the server fest that you were like, that was a really good session? Or you learn something or something you would recommend that people see?

Kaj Arnö
Well, yes, there were several. So when I say just three, it’s mostly Peter Zaitsev and Steve show. So, by mentioning singling out these three, I don’t want to offend any of the others. I do think many, many presentations are good. Why am I singling out? Well, because everybody is not among the presentations if you’re into MySQL or MariaDB. Well, this is because his presentation now was a new one. So very often he has a re heated version with some changes, reports of the latest release, giving a bit of an update. But this time, he was triggered by a wedding of people who had nothing to do with IT. And they asked stupid questions but usually stupid questions can be very smart. So we had to explain to these people what is open source? And what is MariaDB? And the end outcome was the curious tale of the disappearing closed source database. So then the talk was about migration. And he managed to tie this into a story that started from why source code initially was open. And then why did it go close? And why was open source invented on top of it, and what’s the relevance for this make comes to databases and why are people now migrating off? closed source databases? So that’s number one out of the number two out of three would be the Peter Zaitsev comparison of Maria dB, the current Maria DB version 10 of six with MySQL8. And has he pointed out that he’s an equal opportunity offender that he wanted to have? I liked the expression but I did not feel offended so usually In any such comparison, you would end up by saying, Well, yes, yes, he sort of did the right thing. But there was this one or two things that he said about MariaDB that really started because they’re just not true. They’re not true and are represented differently. That’s what I had expected my gut reaction to be, but it wasn’t it was he, I think he did throw a couple of punches in everybody’s direction. But sometimes you do deserve punches and the punches that were not there will not be provocative. So in my Q&A session with him afterwards, I asked him Is this the same?’’ I know like the same thing with with very moderate in his presentation, about about when he could represent in a credible way why you should stay with Oracle, it’s not as if he would have caricatured it, and made it obvious that only the truly stupid people go with such a lousy product, there was no no such provocation. And through this balance, also, Peter Zaitsev, I think that the audience will get some understanding of when to use reach out to do two products. And then the third one would be Steve show of Intel, describing some cool hardware, things related to MariaDB and how it’s running on modern hardware. So that’s like a geek attempt to kind of have a presentation that I would recommend, and we are going to have all of these as individual URLs, you don’t have to look for our number two and minute 34. Oh, no, it wasn’t there. Where is it? It will be defined soon. So we’re welcome.

Matt Yonkovit
Great, great! Looking forward to maybe catching up on some of the content. So MariaDB is a really widely adopted technology, you have taken lots of contributions from across the world. It’s not easy to manage such a large, diverse ecosystem. I mean, how do you do it? Do you have tips and tricks, because there’s a lot of other projects out there that are starting to grow in scope, open source is more popular than it’s ever been? Everybody’s looking at it as the next big business model, unfortunately, and I don’t think open source is necessarily a good business model. But everyone is starting to develop their own open source program offices, they’re starting to look at how they manage these massive amounts of contributions in or out of organizations. And I figured you would be a great resource to maybe ask about some of that management of contributions.

Kaj Arnö
So first of all, I would point out that getting a massive amount of contributions is a huge lesson problem to have. And it’s one that most people struggle with, the most projects struggle for a long while to get, I think it’s everybody’s craving for attention, starting from Instagram to open going all the way to open source projects. Why would I contribute to a product? We’re not sure why that contributes? Why is the project going to be successful at all? Anyway, so I’ve separated the system to the problem, how do you get contributions at all? And then once you have that, how do you manage them, but the common denominator between those is to strive to be a good open source citizen. And I’m not saying that we are a model open source citizen, I’m saying that we’re trying to be a good open source citizen by feeding. And this is very important by treating these contributions based on their technical merit. So sometimes, if you’re really, really open source, you think, even if you’re not greedy at all, by giving away your open source, but if you’ve constantly asked him, so what’s in it for me by getting this for free? What where do I want to go with this product in a greedy fashion, you knew that greed will show to your to your, to your community, and there must be a certain level of selflessness, which you balance with a vision for the product, because of course, you can’t take every contribution has to have technical merit, and it has to be technically good. It has to fit into a vision for the product. So I don’t think that there is any overall easy answer to this. The question that you have, I would say that you will not have the issue of plentiful contributions unless you have a great product and a founder for that group or a founding group for that project that understands intrinsically the value of getting contributions. And then there must be this balance also between the vision of those creating the product, and the vision of wanting to get a contribution. And that’s a very, very difficult balance to have. Because I have my vision of where I want to take something, I get this contribution, which I then should judge based on whether it fits into something that is not exactly what I had in mind, but still should serve the customer needs. So my vision is for the customer to need more than the exact architecture of the product or the or the roadmap, this comes next, and then it comes back.

Matt Yonkovit
Yeah, and I think that that’s one of the difficult things from a contributor perspective is sometimes contributors have really good ideas that are super relevant to them. And maybe only then you sometimes see this from I’ve seen this from some large corporations where they’ll develop some very specific feature or enhancement to an open source project, they’ll try and contribute upstream, but they’re the only ones in the world who might need that. And that’s a very difficult contribution. Because, from a maintainer perspective as you get that, it might be interesting, it might be cool, but you try and say, How does this impact the other 99% of users? And what, what’s this look like for maintenance going forward? So there’s always this give and take. And it’s a very difficult balance,

Kaj Arnö
There are and then so some of these contributions are related to the platform. So use cases will this work? Or will this work on some particular integrated in some particular cloud platform? I think that you have to look at the cost side of maintaining things safely. If this is only you wanting it, but you really, really wanted it, and there’s really not much harm done to the others, then why not. But if there’s a distraction, if there really is a maintenance cost to what you’re doing, I literally am the only person who wants it, then probably, this is not a smart contribution to today. But I think this also fits into how you communicate the company’s vision of the product. If there is a coherent definition of the product, and the database is usually a fairly coherent definition, then working into another platform goes along with that. I think security prospects broadly go along with it. So the communication of what the product is meant to be about shouldn’t be helpful.

Matt Yonkovit
And it’s interesting, because we often talk about contributors, and we talk code. I was having a conversation with a few folks on Twitter, this week, someone had posted, and I completely agree that contributions are more than just code in many cases. And I think that, especially for people who are starting out who want to get involved in a project, like MariaDB or other projects, it’s very difficult to jump in and learn at a good enough level to contribute. And so there’s there’s all kinds of other things that projects are often looking to do, whether it’s documentation, whether it’s how tos, providing feedback to the engineering teams, oh, my gosh, every every open source project that I talked to the engineering teams are desperate to hear how these things are working, now they can improve them. All of those things are super valuable contributions to the project and the community.

Kaj Arnö
Absolutely! So the word contribution usually, like refers to code contributions, but there are so many other ways to to contribute, and we definitely value them. In a way it’s sort of a bit of an elitist thing to define contributions as code contributions, but also, those are the most difficult contributions to deal with both to create and to identify the right way of integrating. So it’s sort of the editor of open source to get good contributions that will fit into the product.

Matt Yonkovit
Yeah, yeah. And so for Maria right now for MariaDB. What sort of things Are you looking for? How can the community get involved today? Like, what are the easy paths that you’re like, hey, we could really use help tomorrow, if you can take a look at some of these types of things.

Kaj Arnö
Well, one easy thing is where it’s really easy, then it will happen in a moment. But take a look at our Jira, take a look at the set of ideas or plans, they are always more spectrum between idea and already concrete definition, what you want to do to get done. So you can do that in two ways. One, one is you can implement it. And the other one is you can just vote for it so that others are more likely to implement it and they see the opportunity if you want to grow as a coder that we actually have a backlog of pull requests that we haven’t dealt with. So it’s not over 100. But it’s a large number of pull requests where we have a contribution, but it’s not entered into the goal because it hasn’t been properly reviewed. And in the reviewing process, it can have several layers. So if you integrate something deeply into a database, it’s very hard, very unlikely that that will go into the code on the local machine. But of course, those critical resources don’t have the time to look at every contribution. They’re easy, they need somebody to make that contribution for them before the Sebago particular micro macro market takes a look at it. So a review isn’t just one stamp of approval, it’s about the following step in the process to get the code into the code base. And there’s nothing preventing people who want to contribute from reviewing other contributions. So I think that’s the easiest win. If you want to become a good contributor, then review other existing contributions, you’re, it’s not as if you’d really have to be worried about making a fool out of yourself. And then suddenly, bad code is accepted into the code base, because to get anything, a particular piece of particular release, those who are in charge of that code area needs to approve approval, but they will do so much quicker if there is a free review by somebody else. That was a long, long winded and the answer to your question, but still something if you want to contribute, I do want to underline that reviewing existing contributions is a good thing to do.

Matt Yonkovit
Well, it helps you get familiar with the codebase. Without it it’s easy to go out there, potentially take it, test it, someone else has written it. And if you run into bugs, you can obviously help the bug and see what, what, what’s going on. It probably is a very low barrier of entry for a lot of folks to get home.

Kaj Arnö
That brings me to an answer to your question about how to foster contributions and how to deal with contributions. We just made an innovation that is so simple that you could ask us, Why didn’t you do it before? We just didn’t think of it. So we have a template on GitHub, where one of the new questions is that you’re trying to achieve with your company? And this is so self serving and philosophy? What’s the purpose of your code? That, of course, we should have been asking it all along. But now since we’ve added that question, review, some much easier to make. And it’s the process of getting it accepted, has shrunk quite quite a bit because you don’t have to do any guesswork of it. And of course, you could also form the contributors that they should have told it from the very beginning what their intention is, but now we people don’t think of that now, that would have started with this template, life has become much easier.

Matt Yonkovit
Now I can imagine. So I think that that’s one of the problems that I continually see across contributions and communities is the assumptions, right? Especially if you are in a position like you are in or I am at Percona, where you’ve got internal teams working on patches and working on the software, you make assumptions that just because you might have an open roadmap or open documentation that end users have read it, and that they understand what you’re trying to do. And they make the same assumption that just based on the code, you’ll understand how they’re trying to plug in or help and that assumption is always a problem. And so eliminating as many of those as possible is incredibly useful.

Kaj Arnö
Code is a form of human communication and people’s misunderstanding exists between any human for any type of human interaction and code is no exception to that.

Matt Yonkovit
So I gave a talk specifically about trying to do as much open as possible as a company at the OS summit date for open source program officers. And so, I’m a firm believer in trying to break down some of those silos and get as much information shared as possible. It’s definitely something that’s important. So, Maria DB is at 10.6 right now. And so it is probably approaching a 10.7 release

Kaj Arnö
probably approaching

Matt Yonkovit
We have something called 10, seven preview releases. So there are 10.7.0 out there. And we’ll do something new with 10.7. So there is not one 10.7.0 go out there and test it. No, there are 10.7.0 go out there and try them out. And what we’ve done there is we’ve taken several different branches and made bills out of that. So you would like to have, for instance, a natural sort or is for what being two features or center. So that’s a little sore thing is like, if you have a number series with 10.1 or 10.2 10.3 10.11. So if you thought that literally, then 10.1 or 10.11 will all come before 10.2 natural sources within you sorted as humans do with a 10.9 being before the end of 10. So that was the functionality. And then there’s his format. So that is for those who know, Michael, there’s this new F type string formatting in three dots, I don’t remember which one. And those are two new pieces of functionality earlier like, then on six, they will have all been put into the first dot zero release and go out there please be tested. But that created some internal friction, because by Thursday at 12, everybody needs to commit their code. And if there is such a deadline, well remember that I will get it done. But can I please get a bit of an extension or, or even if I don’t get any extension, there’s a nice review of these changes, not merge conflicts. So this is format, a natural sort will probably not crash, but there might be others who do. But at any rate, there was a lot of congestion at the time where there’s this final deadline. Now there’s one deadline and year. And I absolutely need some code to get into intellectually. So I’m going to fire my tooth and nail to get it done. And I do want your attention for four days to get here. And of course, this creates friction and frustration. Needless stress. So what was then done was that, okay, I can get this into my release, and you can get your feature into your release, and they will they still have the same deadline but there’s no not any congestion, it’s just my tree needs to be coherent, and provide my feature, I go to this forward, you go natural sort, and then both get released. And then everyone has time to sign and then it turns out, of course not your code, oh, my God, but the other guy’s code was so bad. That it shouldn’t be part of 10.7 at all, it doesn’t feel right. Now, if that had not been done up until the end of six will your good mind already. But lousy code wouldn’t be part of 10.7. And it will be the task of management to ensure that the lousy code becomes good, or really, really have a horrible time getting out already committed code, which is not something to do. So now instead, when 10.7.1 comes out, the lousy code won’t get merged. So yes, there was this preview release. And we spoke about this final feature. And some people liked it. But guess what did work. But now, it will contaminate 10.7. And we did not have the stress. So that’s how we’re changing matters with 10.7.

Okay, I noticed that there were also some additional changes to some of the JSON functionality, as well as some histograms. What is interesting is that a lot of developers are exploring and using JSON and relational databases now. So those are also interesting, …

It makes sense. I mean, if Jason always gets the attention, and for good reason people use Json in their applications. And, and, yes, if you want to be fanatic about the normal forms, Json is definitely not something to put into your database. But there’s a reason why people are using JSON. And there’s also a reason why people are using relational databases because you do need the cleanliness of normal forms and you have to have some spectrum leeway on how to apply when you want to have the logical mathematical clean structure of a normalized database. And when you allow for JSON, and if it’s only on off stage, either you go MongoDB and have Json or you go with EF code model and go fascist on on on normalizations, then you’re probably not doing the right thing. So there is a need for a little mixture here.

Yeah, the balance is important. I mean, I think that the number one problem that I continually see with any of the databases, whether it’s Mongo, or MySQL or Maria, it’s actually the design that people put in them. Right. I mean, I think that the more functionality that you can make it easier for users, the better you are, because users do want to use the tools they’re familiar with. But giving them those easy options to do the things that are natural in the development space is very important.

Kaj Arnö
Yeah, the database is a sharp knife. And most good cooks know how to use sharp knives. So we should give them all the different knives that they need in order to make a good database.

Matt Yonkovit
Yes, some cases we don’t like though, like, for some people, they gotta get started, they gotta learn. So what else is coming up? From a roadmap perspective, are there a couple things that you think are really interesting for the community that you’d like to highlight that might not have made it into the preview release yet.

Kaj Arnö
So the intent here is to be fairly strict about those preview releases. So remember, the first day’s three o’clock deadline, if you missed it, you will be there. So it’s in the preview, your preview releases that we have are the only ones that can get into 10.7.1, it still depends on whether we will keep it open for another alpha release. So we have this, of course, it’s always a matter of how conservative you are in your language use, right. But basically, we have definitions whereby Alpha allows for new, new version, sorry, new, new new functionality, and then you gradually lock it up. And what exactly you can do via class cannot air, there’s always room for a bit of a fight around that. So it’s not absolutely clear yet whether there is another opening, so we will have to stay tuned for that. But there’s no hidden gem that I want to highlight.

Matt Yonkovit
That’s fine. It’s always good to put as much out there in the open anyways, so then you don’t have to have those hidden gems that people get surprised by? Yeah, so it’s good that everybody is aware of what’s coming down the pipeline. Now, I wanted to kind of close with this question. Um I’m curious, what are you hearing from the user base from your contributors? Maybe some of the customers are using MariaDB on where they want to see MariaDB evolve? And maybe some of the things that you’re thinking about in the future?

Kaj Arnö
Thanks for a very good question. And I can tell you, it’s a very difficult question, because people are not always pulling in the same direction. So it’s not that there were more platforms, others want to use it more on the in within the cloud. And therefore want to use the same thing in the cloud and on premise, and they want to be able to migrate freely between the two. Yet other people want to do migrations, so they want to make it easier than than before, to migrate of commercial databases, in particular Oracle, Microsoft SQL Server, it’s instead of a fee, we’ve made it fairly easy already to migrate Oracle, and some say that we still have this, this and this feature that is used in commercial database, if you want to add that. And then there’s this pulling in the direction of Json, if you like your Json functionality, you still would like that from that. So I can’t say that there’s one direction in which we are improving. We’re being pulled in several months, and that’s the great part of being a generalist database, not a general database, and we shouldn’t go in alone for most of these directions. So is the data. As I mentioned now, yes, they might be competing for resources, but they are not diluting the product into something that is not coherent. So for the requests that I mentioned and highlighted now can be satisfied, the only difficulty there is to do with relative allocation of all the all the reasons.

Matt Yonkovit
Great. Well, I mean I think that you’re spot on, there is no one thing that anybody is looking at. Everyone’s looking for 50 different things. And even the same company wants vastly different requests from their vendors, right. I want to run on prem, I want to run in the cloud, right? Okay. So you can run both, or I want to run both at the same time. So all kinds of overlap and craziness. Kaj, I wanted to thank you for taking some time today to chat with us. And tell us a little bit about the Maria DB community and how to contribute and give us a bit of a preview into MariaDB. Future and where, where what’s coming out on the preview release there and what you’re seeing out in the market,

Kaj Arnö
Thank you. Thanks for having me. This was quite exciting.

Matt Yonkovit
Thanks, and you’re welcome to come back anytime you want.

Kaj Arnö
Perfect. ∎