Containerization, Multi-Tenant Architecture, and Open Web Collaboration at the Hackathon

Andrew:
Hello, I am Andrew Hutchings. I’m also known as Linux Jedi. I work for the MarieDB Foundation and I am working with my team on a project to integrate our new feature called MarieDB Catalogs with PHP frameworks such as WordPress, Drupal du, et cetera. And general idea of catalogs is containerization inside MariaDB itself. And that means that you have one MariaDB instance running and every customer has their own catalog inside that instance. So the memory is shared. You’re saving a lot of RAM that way. You don’t have to have 50 a thousand different read DB instances running each with using a giga ram. You’ve just got one read DB instance running, but you still have each customer siloed essentially. Now the whole project we’re doing today is to make it easier to administer this, to create new catalogs, to remove catalogs and things like that.

Eventually we’ll also be able to constrain resources per catalog. So you could say that this is a low tier customer, they’re only allowed so much CPU or so many queries per second, et cetera, and higher tier customers can have a lot more. So that’s essentially what we’re doing today as far as the project goes. So we are just after lunchtime on Sunday. This me DB catalogs feature is still very alpha. We don’t actually have an official release even of the alpha yet, so my team are the first people ever to try it and they broke it a lot. I’ve had to file seven bugs so far and there have been people in Finland and places like that trying to fix things for us on the fly. But we have fixed and worked around many of the issues and as of lunchtime today, we have actually something working end to end where we’ve got a test that in PHP we will create a catalog for you, be able to use the catalog, redirect the customer to the correct catalog and things like that. We have made progress. We also have a WordPress integration, which is separate, and now we’re combining the two. Right now we’re working on a lot of L integration and lots of other projects by the end. Today I’m keeping my fingers crossed. We should to have something very good to show off.

Birgit:
Hi, my name is Birgit Olzem. I’m the project lead for the cloud for Hackathon project for the CloudFest hackathon project, inclusive language checkout for open source contributors. And it is scary for me because I’m the first time a project lead from this scope here and we are currently focusing on improving the workflows for open source contributors, currently specific for WordPress contributors who want to publish content on WordPress sites, especially WordPress.org and WordCamp site to make sure that they’re using in their written content is inclusive and using inclusive wording and highlights also, which terms are critical are not really good to use. And the dynamics of the project is really nice and we have a great group setting. We have a diverse team of about 12 people who are really invested into the idea. We have a part who are developing a plugin which is as clean and lean enough to be installed also on a WordPress, but also to be non-destructive and non-invasive when you are publishing content.

But it gives you a nice overview where you may need to improve your text. And currently we were working on an MVP on a basic functionality, but we have also a roadmap which will be published after the CloudFest Hackathon. And the other part of the group is currently working on the documentation for the plugin itself, but also setting up our GitHub project repositories where the documentation will live for the moment. And also we are focusing on the underlying structure, like the administrative tasks so that we can scale the project after the hackathon because the goal is that this kind of plugin will scale and offers an easy and clean opportunity to check your content for inclusive language, but also have basic check of accessibility like hierarchy and yeah, basically helps you to write better content.

Javier:
I’m Javier Casares. I’m one of the team reps from the WordPress.org hosting team and in this CloudFest in the hackathon we have a project to improve some project inside the WordPress community. We have some tools around how hosting companies can test the future releases of WordPress and that’s what we are doing. So the first day I didn’t know how it going to be because it’s funny because I was here last year and this year it was like, okay, what’s going to happen? And we are only three persons in the team, so I think it’s most the little one, but it’s been awesome because we are advancing a lot because we had some ideas in the past, a lot of hosting companies ask about doing some improvements and we didn’t have the time. So we are advancing a lot. The main thing we did the first day, the main goal I think for the whole CloudFest hackathon, it was checking multiple versions of PHP in one hosting company because usually some years ago, probably a hosting company used one PHP version for all the WordPress sites and now hosting companies gives you the option to pick what version is better for you.

So that was some concern in the hosting community and in the developer part on how we can test everything in all the hosting companies. That’s the first step we did. We get that at the end of the day and today it was like, okay, we need another goal and another step. So we are going to create the multi environment system because another thing hosting companies have is different services because they have shared hosting BPS cloud and each of that environments are different from each other and that maybe is different for WordPress, for how WordPress works in each one we are today we are focusing on that. I hope we got there. But yeah, it’s been funny. It’s more funny. I was very nervous before coming here and last week Carole told me it’s like another contributor today and it was like, okay, I can go with that. So yeah, it’s been very funny.

Pawel:
And I’m Pawel, I’m the AlmaLinux evangelist, so I’m obviously with AlmaLinux OS Foundation and the project, of course we deliver the operating system and as such we are maybe not, outliers is a wrong term here, but we are side of this community here, but we’re also the foundation of many businesses around here we produce forever free. That’s the best feature of it. The second, and I say it’s exactly, so it’s the same position that it’s community driven and community owned. So it’s like the system by us for us and as it’s business we need a stable platform, the production grade platform. So we aim to make it robust enough for ourselves and for everybody in the community to use it. So now we aim to be a ABI compatible with re before that we’re cloning it. The recent, let’s say licensing drama made us change the way, but it’s very good for the project.

We can include new threats into it and we already have results. So we have patches that we will really use in our future better releases. We discovered some bugs, but in general it’s been productive. We’re worried about also our strategy depended how many people will get because we had one of the projects more demanding and we attacked things that we could do with these people that we actually got as an open source project. We need something that has a longer life cycle and makes sense to invest into this. For instance, the umbrella name for our two or three projects is Easing OS upgrades. And that seems a nice idea in general because when you find out very late in the night that there is a old forgotten box that you just plan to use tomorrow and need to upgrade it right now possibly or at the best in place, this is very nice to have such an option also when you have a long-term forgotten instance somewhere in the cloud that you need to operate without resources as well.

You can do this with our Elevate tool, which is like a pivotal point for most of the hackathon projects. So the two ideas that we had where one was about sent six to sent seven upgrade, but this is a big project so we didn’t get enough people to it to tackle on this one. But the other one is making something real world specific. We have an WordPress application running on our Linux seven system and we are doing the complete OS upgrade, including the application so that we can cover the limitations of handling the configuration and that stuff. These days when we’re constantly pushing towards the bleeding edge, we want the latest and the greatest, and I completely understand the industry because it’s newer, you just sometimes cannot get support for the older hardware and that is the motivation for people to go further. At the same time, there are people that know their hardware and they know it was reliable enough to use it.

They can assume that the firmware changes are not required hardware to function under newer releases of the operating system. But there are also people that possibly from emerging economies where we can lower the entry barrier for them to be able to just create their own storage systems with lower costs utilizing the hardware that is lying around. But yeah, that’s the reality. During the project, during this hackathon, we were able to produce a patch for Linux eight. We already built it. We already validated with our box available in our labs remotely. So we have an old LSI card. We just discovered that. Yeah, it works with the newer kernel, newly kernel. We’re working on Linux line patch now, which is a bit more tricky. And the thing is we validated that it works, but what about the future? The future is we need to include it into our testing systems to just test the regression on every major release that we’re actually not just delivering the driver, but it’s really working.

It depends of course on contributors. So we need people who have this real installation and are willing to dedicate time and possibly effort of maybe recovering, maybe just playing with it. Time will show, but the idea is that it is for the people and by the people. So as an press object, we need this helping hand coming from the crowd. We need to add it to our rep repos and say, hey, if you want it. So next step will be the betas release, the nine, four and eight 10 respectively. They will have this patched kernel inside so people can really test if it works for somebody, then basing on this feedback will take the decision. What’s the next state for this projects? Yeah, it’s very nice because we also touching el repo and Santos plus patches that were developed to, they have the build switches that in the kernel that govern building or not the extended version with the so-called deprecated. And that is of course a very helpful situation that they have it because we don’t have to go over each line of code and try to add those guards by ourselves. But there are places that we have to do it. We already did this for the patch four eight and in general it’s again leveraging the collaborative effort, the community work that is already done. So the puff turns standing with the shorter we can shout out to these people in our prs. So it stimulates the emotional part that we’re still collaborating.

Anne:
Hey everyone. So this is Anne and if you’ve been following the hackathon, you might guess who I am. I’m the one with the project called “can everyone use” based on the principle of can I use.com where you can check if something will work in a certain browser. We’re trying to create this central resource about components and libraries and see if they function well in regards to accessibility and if they don’t, what you can do to make it more accessible. It will save you time and it will save your boss time. And you could also probably find out very quickly if the framework you are considering to use is really a good idea. So the project lives on GitHub and partially in a website we’ll share it because this is a process. So yesterday in the morning we were pulling out our hair, we’re sitting there with a huge team, everybody good at something.

We have developers with experience and accessibility, some who do not, some who have a lot of testing experience. How do you set up GitHub? Hey, I’m an accessibility advocate, but I can’t set up GitHub for the life of me. It took some time to find direction and who’s going to do what. We have a marketing team, we have a website content team. Our table needed more chairs. It is wonderful and it’s really scary at the same time because you are doing this project with a lot of people and then you have to let go. Things go out of control and I really hate it when it feels like that, but on the other hand it’s like a joyful goosebumps experience when you do your own thing and after an hour people come to you and say, Hey look, we created this. And I’m like, that’s magic. That’s just magic. And that sense, that feeling of togetherness that you get when you create such a project. I can’t describe it. People have to live it. So anyone listening who’s never participated in a hackathon before, sign up next year.

Robert:
Hi Christian, can you explain what you’re doing at the CloudFirst Hackathon?

Chris:
Yeah, sure. Hi, I’m Chris. I’m director of engineering from Syde and I’m leading together with Robert the project managing multilingual content with multi-site, which is quite challenging as a project.

Robert:
I went here five times and saw so many projects. As you are first time ever at the CloudFest hackathon, how is it to organize and lead a project?

Chris:
Yeah, it’s quite exciting and there are many people there from different companies, different agencies, different levels and different sizes. We have a good mix in the team, not only developers on a typical hackathon where you just code basically. That’s what everyone sees from. So we have a good mix of designers and hosts and two, three people and even normal WordPress users. So the input is quite good and it’s a good team. It’s a big team

Robert:
From the feedback, what we already got while we’re doing that, we both noticed like our ness also with us. We noticed that our first step of this topic was a bit too ambitious, I would say. Can you give a little bit of insight where we currently, which routes we already checked while exploring this topic?

Chris:
Yeah, sure. We came with a good preparation because we put some thoughts into that and what we want to achieve and which way want to go, which was not made in stone but was open for discussions and input and basically we changed everything. So we started from scratch, we got input, we took some steps down and started again, and the first day was basically just talking. So no code was written, which was also many people hoped to write code, but it was just starting on a blank sheet of paper and just go through the problems people have from different areas and just simply making it possible that everyone can use multi-site, not only developers. That’s the big downside currently because it’s quite technical, so you need to do many steps and edit files and change constants in WordPress, which is really technical and only for developers.

So a normal user will never do that. And so we decided we need to solve first that multi-site is it’s possible to enable a multi-site for everyone. So every user should do that and should be able to do that. And when that is solved then we can continue. There are many different topics and packages and where we can work in parallel and are not really connected, but going into the final goal and we’re working on that. It’s interesting because it’s not that you start from a end at some point you iterate and go many ways and then you meet at some point.

Robert:
And as we have experienced so much experience with multi-site as a company, that’s why we brought this project. What did you learn something like perception wise from other people with their perception on multi-site, what you didn’t know before because we already know so many things about that. Was it something

Chris:
Not really. We know most of the problems and we are aware of that, but since we are developers and working long time with multi-site, we can get around it because we know the problem. So we know how to solve them in an easy way for us and for the clients. But it’s a common problem and we need to solve those problems not only for us, we need to solve them in the core basically. So the problems shouldn’t be there anymore In the end, that’s the goal. So there are many smaller topics like user session management, bad documentation, which is not only code, it’s documentation needs to be there that people need to be able to read, understand, and get used to the concept of multi-site. And it’s not that complex in the end, it’s just WordPress and everyone knows WordPress. If you have one button or link more or less, that’s not relevant in the end.

Robert:
And as we try to, one of the goals is to make multi-site easier to use. And we had the core philosophy of decision options. We had some interesting conversations on the table, how easy do we do this? Can you give some examples which routes we took to discuss those simplifications of WordPress?

Chris:
Yes, sure, yeah. Decisions not options. It’s nice saying, but it’s not always decisions, right? Because you need to have backwards compatible code and so you cannot always decide because everyone needs to be able to use it right in the end. So yes, you can make decisions, but everyone needs to be on board for that. That’s the biggest problem we face because our first stand point of view was like, yeah, we turn multi-site on by default and that’s it. Then we have multi-site, then we can do it right? But that’s not that easy because we have many single sites. We have many hosting companies which are not supporting multi-sites. And so we need to find a way to convert single sites to multi-site in an easy way for the end user. So non-technical way again. So the easiest way is click a button in the backend and it’s done, which also has some technical details in the background which need to be solved.

And then we also have the installation of WordPress. So should we install WordPress by default in a multi-site? Always for new sites, which sounds easy, but if we imagine that we have the multi-site has new database tables and those are prefilled with some data, and yeah, that’s a few kilobytes, but if we imagine that we have millions of web sites there, which are installed, then that sums up to a lot of data. That’s a decision we could make, but we have to be aware and keep in mind that it’s not that easy. It’s a bigger scope we need to keep in mind. And then in the end, it’s also when you’re already there or if you already have a multisite or you have successfully converted to a multisite, that’s just the first step. You have the multi-site and now you want to manage multilingual content there.

So how do we guide the user? How do we connect stuff to make it easier? The big term here, network admin, which is a multi-site term for managing global settings. So site settings, users, plugins, and that’s just confusing because what is network admin and what is the super admin? What is that role? And it’s quite special. Again, lack of documentation. Yes, that can solve that, but we should make it easier, more intuitive to the user. So what are the next steps after this hackathon here? Yeah, currently for now, currently we focus on having some screen designs via Figma. So we can also click through them and see like a user journey, user story covered. That’s something we want to do on the hackathon. Now we also write bit code for testing stuff like the automatically converting single site to multi-site. So that’s more proof of concept and yeah, let’s say hacky code, but we want to show that it’s possible and all of that will be taken. And then we aim for a proposal on make developers to basically suggest introducing multi-site to the masses. Awesome, thank you. Yeah, thanks a lot, Robert. Thank you. Have a nice day.

Lucas:
Hi, I’m Lucas Radke. I’m one of the project mentors this year’s CloudFest Hackathon. So together with, we were responsible for picking out those excellent projects we currently have. If I don’t work here on the hackathon, I’m a product manager for WordPress VIP. And I’m really happy to be here to see how much stuff actually gets done during the hackathon because there are so many great talents around who really love to contribute to something and bringing them together for 48 hours and enabling them to do stuff they really care about is just amazing to see. And as an example, a few project leads already approached me and said, Hey Lucas, this is so great. You’ve brought so many great projects and members for the teams, we even needed to adapt the scope because otherwise we would be done in 24 hours instead of 48. So they extended the scope, added more stuff, thought about, okay, how can we really use all the skills of everyone joined because not everyone is just a PP developer, a note developer or something.

But also non-developers are around marketeers, people who can write good documentation, people who look at accessibility and inclusion and stuff. And so each team has multiple work streams and at the end they’re going together to a great project. I can only recommend everyone to apply to the cloudfest Hackathon because it’s really one time in a year opportunity to get together with greatly minded people like yourself and simply do something that otherwise won’t be doable. So we have sponsors that covers everything for you. You get nice food, a great hotel, and you can work together with some really great people.