Episode 12

OpenShift, Communities, Culture, and Machetes with Flavio Percoco


June 4th, 2021

37 mins 35 secs

Your Hosts
Special Guest

About this Episode

Sponsored by Reblaze, creators of Curiefense


Justin Dorfman | Richard Littauer | Tzury Bar Yochay


Flavio Percoco

Show Notes

Hello and welcome to Committing to Cloud Native Podcast! It’s the podcast by Reblaze where we talk about open source maintainers, contributors, sustainers, and their experiences in the Cloud Native space. Today, we have a special guest joining us from Italy, Flavio Percoco, who is Senior Principal Software Engineer at Red Hat. Flavio tells us about working at Red Hat, why he uses OpenShift, and how he stays relevant and chooses what projects he’s interested in. We also learn the moment Flavio realized he made the right choice about his profession, where he thinks Cloud Native is going, the project he is most interested in now, and his thoughts on why Curiefense is the best community right now. Download this episode now to learn so much more!

[00:02:30] Flavio tells us about working on OpenStack and Elastic.

[00:04:47] We find out why Flavio uses OpenShift versus another.

[00:06:21] Flavio tells us about the CNCF project Kubernetes.

[00:07:40] Richard asks Flavio since he’s building the tools for people, how does he stay relevant, how does he know what the clients need, and how does he choose what projects he’s interested in.

[00:12:51] Justin wonders what made Flavio want to go back to Red Hat.

[00:15:05] Tzury asks Flavio to share his “moment” in life when he realized he made the right choice with his profession.

[00:17:50] Flavio tells us where he thinks Cloud Native is going and how he thinks we can get there.

[00:27:14] Richard wonders why Flavio is doing the exact thing he is doing now, how did he decide that this was interesting to him, and why Cloud Native.

[00:30:45] We learn what project Flavio is most interested in right now.

[00:31:54] Flavio talks about Curiefense, and why he thinks this is the best community right now.

[00:35:31] Justin talks about what really impressed him about Flavio when the GitHub thing came out when they changed from master to main.

[00:36:25] Find out where you can follow Flavio online.


[00:03:56] “You work with like fewer resources and you’ve got to get creative on how to build your regions and data centers, and how to automate that in a way that you can deploy thousands and thousands of those bare metal nodes.”

[00:04:26] “The real hard problem comes from day-to-day operations, where you have to actually manage the whole cluster, and all of your regions, and all your zones, and all that kind of things.”

[00:08:57] “It’s one of those stories that it works in my laptop, but like, there’s so many gotchas when you actually run yourself, then you put it in production and you actually have people using your stuff that you don’t realize then until you’re actually running it.”

[00:09:11] “And many times people don’t run your software through the way you think they’re doing it, and many times they’re not even using it for the stuff you created it for.”

[00:11:57] “It’s good to have good processes so that engineers know you know what things to work on and what things are important, but it’s also true that it’s also the engineer’s responsibility or the developer’s responsibility to go and try to find this information because at the end of the day you’re building a software that someone else has to use.”

[00:13:02] “To me it was a lot of people, culture, and the fact that Red Hat is an open source company.”

[00:13:26] “I have no time to be fighting with people. I have zero patience for people that are not willing to contribute and cooperate and stuff like that.”

[00:16:33] “That was the thing that really just like made me fall in love with computers is the fact that you’re bringing in this access to technology, you have the power to bring technology to many people in different places and improve people’s lives and have a huge impact in what society looks like today and what the world is going to look like tomorrow.”

[00:19:23] “It’s extremely important for people to have the right infrastructure to be able to develop whatever they’re doing research on and be able to do it also without having to do a massive upfront investment, which is something that tends to be a massive impediment for people to actually move on into the stuff they’re interested in.”

[00:19:47] “And this is the other thing that to me, not only in the Cloud Native world, but like all the open source is extremely important because I didn’t go to college.”

[00:32:20] “It really excites me that from day zero where we’re all being open to making the right calls that will favor community, and it will favor contribution over just pushing features and just like adding stuff to the software.”

[00:33:08] “But, even at the cost of moving slower or taking care of very important things like having proper CI, and I don’t mean CI in terms of we need to test our software, I mean CI in the sense of if someone comes and submits a PR and that PR is not correct, what is the best way for us to communicate that to the person that is contributing that PR.”

[00:34:37] “I’m extremely excited about this and this is something that it’s close to my heart, like in communities and cultures are really close to my heart way more than software is.”



Curiefense Twitter

Cloud Native Community Groups-Curifense



Justin Dorfman Twitter

Richard Littauer Twitter

Tzury Bar Yochay Twitter

Flavio Percoco Website

Flavio Percoco Twitter

Flavio Percoco Linkedin

Red Hat

Red Hat OpenShift

OpenShift Assisted Installer-GitHub


OpenShift Assisted Installer #278-GitHub



Flavio [00:02]: I wasn't into computers until I was probably 18 years old. In my childhood I only played outside in the mountains, in the rivers. I used to run with a machete in my hand, just like cutting grass. Man I'm from South America. That's how we do it. It is like that. I say this because I'm actually happy that I had this childhood when I could just like interact with nature and eventually just find this thing that is going to be the tool and the stuff that I want to.

Richard [00:33]: Hello and welcome to Committing to Cloud Native, the podcast where we talk about the confluence of Cloud Native and open source. Super excited about our guest today, which is kind of what I say for most guests, but this one in particular, just because he is awesome energy and helped us deal with some squabbles in the beginning with technology, which makes me think that open source and cloud native will also be dealt with appropriately. Very excited about that.

Before I introduce Flavio. Oh no, I already did. I want to introduce the other panelists here. We have Justin Dorfman. Justin, how are you?

Justin [01:05]: I'm doing great. How are you, Richard?

Richard [01:07]: I'm good. Thanks for asking and Tzury Bah Yochay. Tzury how are you doing?

Tzury [01:12]: It's bah-yo-hi, not bah-yo-hay. Richard time for you to get it right.

Richard [01:18]: Tzury Bah Yochay

Tzury [01:18]: There you go. Episode number eight or number 10, who knows and you got it correctly. I'm fine by the way. Great day today.

Richard [01:28]: Glad to hear it. Flavio Percoco. Did I get that right?

Flavio [01:33]: You did yeah.

Richard [01:39]: It's so good to have you. Flavio is joining us today from his beautiful mansion in Lake Como, because I can't imagine anyone living in Lake Como and not having a giant awesome house on the lake. He's a software engineer at Red Hat. Flavio, how often do you just sit and look at your lake mansion and think this is it, I've made it?

Flavio [01:58]:Quite often. I'll have to invite you over to show you what a mansion actually means in Italy. It's like very packed. It's not what you're thinking right now. The only people that have mansions here is like [inaudible 02:11] those people. I guess it's fine, but like, it's awesome. I go there quite often when I'm in town. It's great to sit there, eat an ice-cream.

Richard [02:23]: You're not always in town, because you often do a lot of conferences and you're a speaker and your travel the world all the time. Right now, you're working on OpenStack. You talked a bit about that.

Flavio [02:32]: I used to work on OpenStack. I worked in OpenStack for like six years. I don't know. Someone was asking me this the other day, like when did you start? Like, what was the name of the release that you started? I don't know. [inaudible 02:46] one of those. Then worked there for like six years more or less. Then I moved on to working on OpenStack and Kubernetes at the same time because OpenStack was not busy enough, I guess. I started working on integrating and eventually just went on from there for a couple of years, took a break from that and then now I'm back.

I moved to Elastic and I started working on something completely different. It was always related to infrastructure, but I do not have to build the software that provided that infrastructure. I was just able to use it and build stuff on top of that infrastructure and kind of put my hands back into dev ops, which is something that I hadn't done in like several years ever since I started working at Red Hat and then spent a couple of years at Elastic and back at Red Hat.

Full-time there working on OpenShifts, mostly. I had mostly focused on OpenShift deployment workflows on bare metal nodes. Everything that has to do with dealing with their data repair centre or bare metal nodes, and how to make that easier for customers and all the TACOs and edge cases and that kind of thing. It's an excited fill. You work with like pure resources and you gotta get creative on how to build your regions and data centers and how to automate data in a way that you can deploy thousands and thousands of those bare metal nodes without actually bringing everything down, I guess.

Also the hardest part I think, of this whole process is not so much how to automate the whole process, but they're zero operations. To some extent they are easy. There are like a self problem. This is a problem that we're trying to like make easy for everyone. The really hard problem, comes from data operations, where you actually have to manage the whole cluster and all of your regions and all your zones and all that kind of things. That's when the actual real infrastructure management really starts.

If you don't provide the right tools in a way for people to actually monitor infrastructure easily, then it's hard to really scale it at the levels that we want people to scale this thing.

Richard [04:48]: How does OpenShift provide that? You said like it helps you deploy your bare metal nodes. I'm pretty sure I know what that means, but why OpenShift versus any other?

Flavio [04:57]: Well, so this is something that everyone had with deliveries, that is a very important problem to solve. The only reason I say OpenShift is because that's Red Hat's product, really. That's what we go and sell to customers and people that are using it and still open source so we really sell support. But it's all based on Kubernetes. This specific tool that I'm working on is still open source. People can use it. It relies a lot on some of the OpenShift internals, but I will build their pieces around this tool that we call a system installer, assisted service, all the related pieces around it, that all pieces are all part of the obscene community.

They all have an option community, part of the CNCF. They all follow the Kubernetes release cycle and try to be as community friendly and as open source of friendly as possible, which is perhaps full fashion upstream first.

Richard [05:50]: Nice. So it's just an installer is what you're mainly working on at the moment. It sounds like you've touched like everything, which is awesome.

Flavio [05:57]: Yeah the assistant service is just a piece of the system installer. It's mainly the thing that I'm working on right now, spending most of my hours there and trying to get the whole story straight, because a system installer can be not only bare metal, it can also do [inaudible 06:11] from cloud providers and do all that stuff. the part that I'm mostly interested in honestly, it's all there. They're [inaudible 06:17] not use-case of system installer.

Richard [06:24]: You mentioned CNCF there, which one of these is the CNCF project?

Flavio [06:27]: Well, Kubernetes is. I'm just kidding. That is a big part of it. There's another community, it's not so much which of the projects is a hundred percent part of the CNCF, but rather what parts of these projects are already part of the CNCF. For instance, like you see Kubernetes as part of the CNCF, but then there are a bunch of internal APIs, like the cluster API, the machine API, and some other internal interfaces that allow people to be all these tools and services that interact with the Kubernetes cluster and that allowed them to manage all the resources like their metal nodes and like VMs and networking and some other parts of the whole infrastructure environment.

I would say the main project that is part of the CNCF that our [inaudibel 07:16] can work on is really Kubernetes and everything that is around it that will allow us to actually provide these basic install experience that we want people to have when deploying OpenShift.

Richard [07:27]: I have this like existential question, which is, you're talking about all these projects you've been working in this industry for years and you're providing services for people who are setting up their infrastructures, but that's not what you're doing. You're building the tools for them to do that work. How do you stay relevant? How do you know what the clients need? How do you choose what projects you're interested in, seeing is how you're not a person throwing up all these instances every single day.

Flavio [07:53]: Yeah, it's funny you're asking that. That is an interesting question. It's a question that I really posed myself the first time when I joined Red Hat, the first time back in 2013, where I started working OpenStack. The OpenStack story, really for Red Hat when I joined them, we were really few people compared to how big the team is right now. One of the things that we kind of like stress a lot in the OpenStack community when we started building new services and growing the community itself is, how do we get data from the field? Let's look like Red Hat for a second. Let's talk about the OpenStack community.

As a community, assuming we don't have any company behind us, how do we get people to actually tell us how they're using software so that we can make it better?

Something that is, I don't want to generalize, but I believe like many developers miss the fact that if you develop things in a box, if you develop things in your tiny oral without looking at how people are running your software, you're really developing a software that works for you.

It's one of those stories that it works on my laptop, but like there's so glitches when you asked to run it yourself there, and then you put it in production and you actually have people using your stuff, but you don't realize them until you're actually running it. Many times people don't rub your software the way you think they're doing it. Many times they're not even using it for the stuff that you created it for.

How do we get all these people to actually give us that information? What we started doing in the OpenStack community, it was made in this subset of the community, so to speak, that was the operators community. How do we welcome these operators so that they feel that they can be part of the community, like a hundred percent OpenStack community contributors, even if they don't commit any code to the OpenStack database. We pushed for this, established these and try to like make them part of their technical committee and try to make them a central piece of the entire community. And something very similar is what I try to do from within Red Hat so there's these things that you can do in an open source community where you try to like bring more contributors to it. Bring people that are actually using the software and being part of the decision-making process, because at the end of the day, they are the ones actually scanning your software. They bring like some use cases that you're like, I have no idea that you can actually do that with the stuff that I was spreading.

Something similar is this for a company like Red Hat. But like Red Hat is a big company. Then it has like many people taking care of all these product management and all those things. As an engineer, I take care of basically chasing down product managers so that they tell me what customers want. When we were building this OpenStack team at Red Hat, I remember just going to the product managers of this software that I was working on and telling them, I want you to bring me all the customer's experience. I want to know how they're running the software, all the problems that they're running into. There was a lot of that. Also a lot of the tracking with our support team. Red Hat's support team is awesome. They take care of so many issues. We're level two, like engineering is like two or three, I don't know, like you put a number on it, but like many times we don't even know all the issues that they're working on, because they can solve them.

But many times knowing what issues they are solving for the customers will bring more information while we can actually improve it. There's a lot of like chasing that, at least for me at the very beginning. It was a lot of chasing down product managers and making sure that it could be tracked more with support. I would collect all this information and either just build my own proposals, or like bring in up backup streams that people would know how some of their customers were using the software and what other things could be improved.

Many of these conversations just end up being in requests for changes upstream, request for improvement and new proposals of new features or things that are bugs that need to be changed. It's still true. It's still like that. At the end of the day, it's good to have good processes so that engineers know, what things to work on and what things are important, but it's also a true belief, but it's also the engineer's responsibility or the developer's responsibility to go and try to find this information.

Because at the end of the day, like they're building a software that someone else has to use, for various use cases. Some of them are mission critical. Some of them are just someone running OpenShift on the desktop. It's fine, but they're all important. But they are all people that are interested in running the software. Knowing this use cases, you're not going to be able to solve them all, but a least knowing that they exist is definitely gonna make you a better developer, I believe.

Justin [12:37]: Got it. Like on Git Hub, you're part of two really big organizations. Kubernetes, and the Kubernetes special interest group. You could work pretty much anywhere with those in mind. What made you want to go back to Red Hat? What is it exactly that made you say Red Hat's the place to be and I've been there once and I'm going to go back and do the experience again?

Flavio [13:03]: To me it was a lot of people, culture, and the fact that Red Hat is an open source company. To me, culture is something that is extremely important. People say, if you want to really be contributing to open source software, we have to be thick skin, be this super rough human being. I have no time to be fighting with people. I have zero patience for people that are not willing to contribute and cooperate and stuff like that. I want to work on things that excite everyone. I have an impact on digital work. If you look at this software that we normally work on when we're at Red Hat is, something that many times people don't even realize that they're using.

People don't even know that they depend on the software. But then I'm like, why not? It's a software that makes me happy to work on. The people that I work with, they're great human beings. The culture of the company puts people first over many other things, many costs.

That is something that for me, is like super valuable, especially in these times. Especially in this period of like given the tech industry where it's easy to burn out. It's easy to be fed up with many things. I would summarize it in people, culture, and definitely they try and just what we're working.

Tzury [14:24]: Well Flavio, you mentioned that the software you're working on makes you happy. Well, I can tell from the little time I know you that you do what you like and you like what you do, which is a good thing. But everybody I know from our profession, people in our industry, most of them, at least two moments in their life. The first moment is the moment they knew they want to go into computers, software and so on. Then the second one is a moment when they look back and say, well, I think I made the right choice. This is indeed my profession. This is indeed my trait. Would you mind sharing with us those moments?

Flavio [15:07]: Well, the first moment is probably easier for me. I tried many things. I have many ideas. I've always had many things I want to do. It's probably one of my, I wouldn't say my strength, it's actually the opposite. That's one of the things that I tend to spread myself too thin on many things, call it curiosity, call it whatever you want. When I finished high school, I guess I was looking into many things. I even signed up for psychology and medicine and a bunch of other things, but then eventually because of the same curiosity, I ended up taking an online course on it. I was all BHP, my SQL all the way down and some HTML on there and that just clicked with me.

I started like playing with it a little bit, like doing all the changes and all the logic, but it was not that. It was not like love at first sight. The thing that really made me fall in love with this is what I did after. I did my internship, like working on web development and stuff like that. Then eventually I moved into some other company that was building a Linux distribution, mostly focused on assistive technology. It was built by people or people with like vision problems and people with motion problems to actually be able to use a computer.

I was working on that distribution and I was working on the software back in the day. It was called Orca. I created this other software that was called mousetrap, which would allow you to move the cursor just with your webcam in your face. That was the thing that really made me fall in love with computers. The fact that you are bringing this access to technology. You have the power to bring to knowledge to many people in different places and improve people's lives and have a huge impact in what society looks like today and what the world is gonna look like tomorrow.

That is a thing that really made me completely fallen in love with it. I ended up working on that for a while and then eventually ended up moving to a different country side, to switched jobs for obvious reasons, I guess. Then eventually I started doing something else, but it's always been important to be working on something that I believe will have a significant impact in the way we use technology today, in the way the world works.

Richard [17:22]: You talked about going round to the product managers and making sure that they get feedback to you, because it's your job as an engineer to listen to your users, and to make sure that you're able to build tools that they could actually use. You talked about how you want to improve people's lives and improve their access to tech. You must have a very unique position. Being someone who has that mindset and has worked the places you've worked, to know how Cloud Native infrastructure can actually help people's lives in the future. I'm wondering where you think it's going and how you think we can get there.

Flavio [17:55]: I would say it's going in the right direction. That would be my first answer. I think we're making many things right. It's interesting. In some of the cases like when we talked about telcos they're to users, because you end up enabling the infrastructure that eventually enables communication across the world. You're providing the software that these massive companies are actually using. If you think about it, when like I'm not going to say an NSP. But eventually when you're using this carrier and you're calling, they're making goals to use, like, just imagine in your head, it's like right now on this antenna, this is happening because like the software that I built and I worked on, is actually handling all this traffic.

When people can go, I don't know, like emergency number 911. Like even though they're like more less traumatic use cases where you're just going to your loved one and you're enabling communication for people, those are very important. These days, it's Red Hat, Telco, Edge, that kind of things. But if you also look, [inaudible 18:56] which is something that Red Hat is starting to get into, it's pretty awesome to know that you're also enabling and improving the way all the automotive world is actually working and the way it's evolving. They're also like more, I don't want to say simplier, because I don't want this to sound like they're not as important or as impressive.

But even just enabling research for universities, for colleges, like it's extremely important for people to have the right infrastructure, to be able to develop whatever they're doing research on and be able to do it also without having to do a massive upfront investment, which is something that tends to be like a massive impediment for people to actually move on with the stuff that they're interested in. Making this software that people are free to use at any point and that they can learn from. This is the other thing that, to me, I'm not only in the Cloud Native world , but like all the open source is extremely important.

I didn't go to college. Everything I learned it by reading someone else's code. I learned it by working on open source. I learned it by contributing to open source. I learned it by being in RC and talking to people and having people actually explain things to me when I didn't understand them. To me it's extremely important to have all this software that other people can use. Cloud Native in particular is extremely important because that is what we want the world to move into. We want things to be accessible through the internet, which I believe should eventually become a human right.

And just have people be able to just like scale and put everything that they need online and have easy access to all these technology, that for a long time was extremely expensive and really hard to get your hands on. Today you can just click on it and you will have access to all this technology be able to do what you need to do.

Richard [20:51]: It's clear to you how you justify working in Cloud Native and how you tell the story to yourself that this is valuable work. It's clear to me that that's a valid story. That's awesome. It's really great to be able to do that stuff and to enable people to call their loved ones and research possibly at the same time. You answered it just fine. It's really cool to know that you didn't go to college, that you learned everything together because that really shores up this image I have of awesome developers being community-based first. Because you had to learn from other people.

Flavio [21:25]: Totally. I'll tell you even like, I wasn't into computers until I was probably 18 years old. In my childhood, I only played outside in the mountains in the rivers. I used to run with a machete in my hand just like cutting grass. I'm from South America. That's how we do it. It is like that. I say this because I'm actually happy that I had this childhood where I could just like interact with nature, and eventually find this thing that is going to be the tool in the stuff that I want [inaudible 22:01]

Justin [22:03]: I'm really glad that you did bring up the college thing because I don't know about where you were from, but for my parents, it was a very big issue between us because I knew what I wanted to do. I had the same thing with you. Like I saw the source code, I saw the future. So I thought, but it ended up turning out pretty good for me. I'm not going to lie, but I think maybe not so much now, but there's always been this kind of stigma, at least for me, especially like when you have a LinkedIn page and it says, what school did you go to do? Then you put Camino Rio high school. It just looks kind of weird. But at the same time, I think this day and age, that stigma is kind of going away because really the curriculum can't keep up with the technology in a way.

Things are changing so quickly and more and more people are just learning on their own or learning through stack overflow and all these other channels that is not traditional education. Thank you for bringing that up. Because to me, I was like, oh, he definitely, probably has a CS degree and all that other stuff, because you're a principal engineer at Red Hat, that just got acquired for $38 billion. It's like a huge company that makes a huge impact in the world and the open source space.

I hope more people listening to this don't feel bad about not going to college or not being able to afford it or whatever the case might be, because you can't succeed.

Flavio [23:37]: You definitely can. I think people should definitely not feel bad about not going to college. But I also think that if you have gone to college, it's absolutely fine. They are things that you learn in college and there are things that you learn on the field. I think the only events that people didn't go to college have, or there are people that went to college is that they get experience from the field right away from being there. Now it comes with a cost and there are many basic things.

By basic I don't mean like obvious things, but like there's many theories that people that don't go to college to know unless they actually need them. This is how I've been like learning the things that I need. But like you, you face an issue, it sparks your curiosity and you'd go to read about it, learn about it. Then you move on and you keep going and then you go as deep as you have to go to actually be able to do the job that you have to do. Now, these are many things that people that actually go to college and go to their regular curriculum.

They will know from their studies, which is also like, it's all great information. I'll say it's awesome information that maybe eventually when I retire, I'll just go and read all these books because I still think it's very interesting. My fun anecdote about this is that I once tried to interview with Google, they rejected me right off the bat because they asked me a very basic question that you would learn in the first semester of computer science, which is what is the speed of rotation.

I was like, writing this function. That is like, I don't know, like, O N or whatever. I was like, dam I got them. I don't know even know what this is. I was like, can I Google it? In the middle of the Google interview.

Richard [25:16]: I've been rejected for more interviews for that question than any other. I kind of refuse to learn because I'm just so sick of being asked. It's not relevant.

Flavio [25:30]: Right in the middle of the Google interview I was like, can I Google this thing and the guy was like, yeah sure. I did it. I was like, oh, okay. I understand the concept. I understand what it means. But it's clear from our current interaction that I didn't know this basic thing that you believe is extremely important for my job. If you ask me now, I will tell you, like, I would never ask anyone that question on an interview. It has to be a very specific use case or like very specific technology that you need to be working on. Like something that is extremely embedded. It has to be absolutely fully optimized, whatever you want to put it like in whatever way.

But now years and years after that, I have a different view of that interview. That very same day I was like, had I gone to college, I would have known this thing, right? I would would have been able to answer this question.

Richard [26:20]: We could not disagree more. Speaking to someone who has gone to college for more years than is really necessary. College gives you absolutely none of this. In fact, the ability to decide to know what you're interested in, dive down and research it as deep as you want and then move on to something else. It's something they don't teach you in college. They teach you how to do really badly at everything all at the same time. Both of you actually benefit from not having gone. I have two master's degrees. I know what I'm talking about.

That all having been said, you mentioned earlier about diving in. You mentioned that like in the very beginning of the podcast. I forget the exact quote, but I've been wondering for a second now. Since you have this great experience of like being interested in some stuff, which meant you maybe don't have the background other people have. You haven't read code 101. Your code complete two may not be in your shelves, which is probably fine. It's a really big book, but I'm curious. Why are you doing the exact thing you're doing now? How did you decide that this was interesting to you? How is it more interesting than running around with a machete in a river next to the mountains? Why Cloud Native?

Flavio [27:27]: First and foremost, I didn't say this was more interesting that running around with machete. I said, I used to do that. I wish I could go back to doing that, but I also have to eat. Making some money is part of the whole deal. Everyone is attracted to different challenges. I think the main thing under this whole thing that I really enjoy working in is distributed systems. Everything that has to do with multiple computers, multiple software, having to interact with other parts of the architecture, through the network, then having to synchronize and having to solve these issues either to work with concurrency or probabilism, or just distributing the load or whatever.

This really the thing that attracts me the most. There's a lot of that in the stuff that I'm working on right now, then having all these notes, different hosts, and metal nodes coming up and how then all the different agents and tracking the service and making sure that the whole deployment process is working properly. I've always been interested and attracted by distributed systems. Because I guess it aligns a little bit with my obsession of communities and culture and having everyone actually agreeing with each other.

That actually pushes me into trying to solve distributing issues and try to bring all these systems into our agreement and to state that we are all happy with. How did it get here? It's a good question. I actually started working on the search engine back in a different company. It was actually based on the last search before Elastic the company existed. It was very earlier days before Elastic started working on that and trying to build NLP natural language for a system, that goes on top of it.

It was part of like a bigger architecture. I was not so much into writing the algorithm itself. It was not something I was super interested in. I wanted to deal with, what happens to the data for the moment it gets into our services, where it goes, what does it store? How are we going to process it after? how is it going to end in the search engine? How is we're going to use the results from the search engine and how we're going to propose this to the user and how we're going to do this.

Back in 2010, it was like terabytes and terabytes of data. That was like a massive amounts of data. I started working with that. That was a startup like normal, regular startup story eventually ended up failing. But after that, I ended up moving to the Red Hat when I started working on OpenStack and was still attracted to the same issue site.

It's a distributive computer provider and cloud provider.

How are we going to make all these different pieces work together? In this case, it was very, there were all like very separate and different pieces. Like one was providing [inaudible 30:07] The other one was providing images. The other one is providing like [inaudible 30:12] and all those things. It all has to work. Within each of these services that were like different parts I had to travel to. It's always been around distributing systems in general, as the part that really attracts me.

I think the whole theory behind consensus and all these algorithms and how to make systems interact with each other in a way that they are full tolerant is something that I find interesting, which also applies to security defense in great detail, in great extent.

Richard [30:40]: [inaudible] is a huge field. Right now, this week, what are you most interested in? What project is most interesting to you that you've given into?

Flavio [30:50]: I don't think I've had to deal with a lot of tolerance in the last months, I guess. Concurrency I think is definitely a part of what we're doing and the assistance, or because of the need of deploying all this nodes concurrently, and making sure that old data state is properly managed and that all the agents have the information that they need to move on. I guess the hard part comes when we need to make sure that all the network settings are set properly and all the nodes. It doesn't end up like when you're dealing with a list like everyone, you also have to deal with the probationing network, which is a separate network from the one that the actual [inaudible 31:34] end up in, which has its own challenges, especially because then you have to connect all these hydraulics for that.

They have their own problems. I'm probably not going to go too much into that, but there's a different kind of distributive challenges that I'm dealing with right now that don't really have to do a lot with concurrence at this moment.

Richard [31:54]: If you had to choose out of all your different projects, I know you went back to Red Hat because the community is great. If you had to choose a community that you think is doing the best right now, or that really makes you feel good and warm inside, fuzzy, if you will. Which would it be?

Flavio [32:05]: I'm super excited with the community that, [inaudible 32:10] I mean it. It's still small, but it really excites me that from day zero, we're all being open to making the right calls, that will favor community, and that will favor contribution over just pushing features and just like adding stuff to the software bite. Cause being very focused on, we want to build a community like that. The technology is there, the technology's there, the challenges are there, the ideas on where to get it next, they're all there. Now we need to make it happen. How can we make this happen in a way that we're not only going to make it happen, but also gonna make it happen in a way that people will feel comfortable jumping into the project and jumping into slack or the repo and contributing to the project. We want it to make it happen like ultimately fast, which is like not making a community.

We'll just code like crazy and just push everything. But even at the cost of moving slower or taking care of very important things like having proper CI, and I don't mean CI in terms of, we need to test their software, CI in terms of, if someone comes and submits a PR and that PR is not correct, what is the best way for us to communicate that to the person that is contributing that PR. It's not so much about telling them your code is wrong, or protecting that code to going into the repo. But it's about communicating properly to the rest of the community or the rest of the contributor's like, hey, this code is not working properly and this is why, and have that very expressive way to actually say this, because I don't know if that test suite is extensive enough and it's super clear, so

people, when they click on the link and they see the loss, it's obvious to them what's going on and how they can learn more.

The same thing applies for all the docs that are being built. How can we make this feature and document it in a way from day zero, that is clear to people, what it is meant to do not what we're after. The community is not too big, but again, at the cost of moving slower, many things are being done for the community and for the thing that we want to build and that we want to be like this security project, or like all the proxies out there, basically. I'm extremely excited about this. This is something that is close to my heart, like in communities and cultures, are really close to my heart, way more than software is.

To me, this is extremely valuable is something that you would think that you would find pretty much everywhere nowadays, considering how much this has been talked about and social media, like Twitter, everyone is saying like community. You gotta be nice and we're done with all the jerks and we're all above that. Then eventually you go to some repo that is just full of those people. I don't want to rant a lot about it, but I'll see the glass half-full and say that there are good cases, like the pure fense community that are taking all the right steps, where community and people come first. Despite once again, at the cost of moving slower.

Justin [35:30]: Just to add on that real quick, it really comes from the top because here's one thing that really impressed me with Flavio, is when the whole Git Hub thing came out, when they're changing from master to main, the fact that he brought it up first, I was like kind of hesitant. Because I was like, oh, I don't know if this is my place to do it. But he said from the get-go we just switched to main because we want more contributions and we want to be more inclusive to the community.

I'm like, sorry, where did you find this angel? He gets it.

Richard [36:02]: I just opened a drive by issue on a system installer, asking for a contributing guide. I'm really excited about that. Thank you so much, Flavio. I think we're coming up to the end of our time here, because podcasts are mortal things. Also food is very good. Flavio, thank you so much. This was awesome. Where can people find you online? Where can they read more about the things you do? Where can they find your tweets?

Flavio [36:32]: Funnily enough, as much as I complain about Twitter, I'm still on Twitter. That would be flaper87. And yes 87 is the year I was born. Just in case you're asking.

Justin [36:48]: My sister was born in 87.

Flavio [36:52]: There you go. Best year ever. That would be all my links out there. i don't tweet as much. I used to tweet way more than I do these days. I'm still there. I still enjoy reading some of the tweets and blocking the ones that I don't like.

Justin [37:05]: Tzury I think we should get him a mic and then he can be on this podcast as a panelist more often.

Richard [37:17]: Thank you so much. Flaper87, everyone looking forward to our next podcast, Flavio looking forward to seeing you around. Catch you later.