Episode 14: Managing SREs to 90+ Container Talks a Year with Jérôme Petazzoni

Sponsored by Reblaze, creators of Curiefense

Panelists
Justin Dorfman | Richard Littauer | Tzury Bar Yochay
Guest
Jérôme Petazzoni
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. Our amazing guest today is Jérôme Petazzoni, Founder of Tiny Shell Script, LLC, a long-time operator in the Cloud Native space, and was part of the team that built, scaled, and operated dotCloud PaaS before that company became Docker. We learn how Jérôme became involved in Docker, his involvement in Helm, how he loves teaching complicated things to people because he finds it a motivating challenge, and how he has a ReadMe driven development. Jérôme tells us more about what his company Tiny Shell Script does, what he’s doing with Cloud Native Islamabad, and if that’s not enough, he tells us more about how a Launchpad became his travel instrument, and how he can play the theme of Zelda with a Launchpad, Raspberry Pi, and 3000 lines of Python. Download this episode to find out much more from Jérôme!
[00:01:33 (https://podcast.curiefense.io/13?t=93)] Jérôme fills us in how he got involved in Docker, which was dotCloud PaaS when he started, how he ended up doing the job he did there, and what parts of the code base of Docker he’s touched.
[00:09:21 (https://podcast.curiefense.io/14?t=561)] Justin wonders how many talks Jérôme has given in his career at Docker.
[00:13:35 (https://podcast.curiefense.io/14?t=815)] Justin asks Jérôme if he feels that Docker paved that path to the Cloud Native space or if he thinks there’s more to it.
[00:14:58 (https://podcast.curiefense.io/14?t=898)] Richard wonders where Jérôme sees the role of the developer evangelist of people like him, who accidently randomly do all of this marketing and work in building open source communities in the Cloud Native space.
[00:17:48 (https://podcast.curiefense.io/14?t=1068)] We find out how Jérôme is a massive proponent of good documentation.
[00:19:42 (https://podcast.curiefense.io/14?t=1182)] Learn about Jérôme’s involvement in Helm.
[00:22:53 (https://podcast.curiefense.io/14?t=1373)] Jérôme tells us what Tiny Shell Script, LLC does.
[00:24:52 (https://podcast.curiefense.io/14?t=1492)] Justin wonders what made Jérôme just go and do his own thing.
[00:29:51 (https://podcast.curiefense.io/14?t=1791)] Richard asks if Jérôme trains his developers in French as well, and Tzury asks him if he can come do training at Reblaze.
[00:31:37 (https://podcast.curiefense.io/14?t=1897)] Since Jérôme has done a lot of work elsewhere, Richard is curious in what’s going on with Cloud Native Islamabad.
[00:33:18 (https://podcast.curiefense.io/14?t=1998)] Jerome tells us about his launchpad that became his travel instrument and making music.
[00:35:55 (https://podcast.curiefense.io/14?t=2155)] Justin asks Jérôme if Kubernetes be what it is today without Docker, and Justin says no and wonder if he agrees.
[00:37:12 (https://podcast.curiefense.io/14?t=2232)] Find out where you can follow Jérôme online.
Quotes
[00:15:17 (https://podcast.curiefense.io/14?t=917)] “And I would even say for any company who’s going to interact with developers because in a way we have a lower attention span.”
[00:15:29 (https://podcast.curiefense.io/14?t=929)] “So, you know if I have five different containers storage products to choose from in ten minutes, basically I expect all of them to give me a two-minute video that’s going to explain why they’re the best solution.”
[00:18:13 (https://podcast.curiefense.io/14?t=1093)] “Now in a way, I think it may be Lauri Apple who gave me this notion of maybe she doesn’t call it like that, but ReadMe driven development in a way.”
[00:18:34 (https://podcast.curiefense.io/14?t=1114)] “Kind of start with the ReadMe because that’s how you’re going to explain to the world what that is and how it works and how folks should get started.”
[00:21:43 (https://podcast.curiefense.io/14?t=1303)] “Sometimes I say, maybe I stole that from someone, that is kind of the missing package manager for Kubernetes.”
[00:25:11 (https://podcast.curiefense.io/14?t=1511)] “At some point in 2015 when I was kind of questioning myself and what I wanted to do, I went around an interviewed at a few places and that was really interesting and I totally recommend to do that once in a while because it helps to kind of ground your expectations and everything.”
[00:26:10 (https://podcast.curiefense.io/14?t=1570)] “That was part of their greatest course, but I don’t remember a single thing about them.”
[00:34:43 (https://podcast.curiefense.io/14?t=2083)] “Until at some point again, I’m going to say accidentally stumbled upon the right combination while I realized I could very easily process media messages with that thing.”
[00:36:09 (https://podcast.curiefense.io/14?t=2169)] “It’s one of these things where you need a “what if machine.”
Links
Curiefense (https://www.curiefense.io/)
Curiefense Twitter (https://twitter.com/curiefense?lang=en)
Cloud Native Community Groups-Curifense (https://community.cncf.io/curiefense/)
[email protected] (mailto:[email protected])
Reblaze (https://www.reblaze.com/)
Justin Dorfman Twitter (https://twitter.com/jdorfman?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Eauthor)
Richard Littauer Twitter (https://twitter.com/richlitt?lang=en)
Tzury Bar Yochay Twitter (https://twitter.com/tzury?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Eauthor)
Jérôme Petazzoni Twitter (https://twitter.com/jpetazzo)
Jérôme Petazzoni Linkedin (https://www.linkedin.com/in/jpetazzo)
Jérôme Petazzoni Website (http://jpetazzo.github.io/)
Tiny Shell Script, LLC (https://tinyshellscript.com/)
Docker (https://www.docker.com/)
Helm (https://helm.sh/)
Cloud Native Islamabad (https://community.cncf.io/islamabad/)
Cloud Native Islamabad Twitter (https://twitter.com/CloudIslamabad?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Eauthor)
FluidSynth (https://www.fluidsynth.org/)
Container Training with Jérôme Petazzoni (https://container.training/)
Zelda with a Launchpad, a Raspberry Pi, and 3000 lines of Python- Jérôme Petazzoni-YouTube (https://www.youtube.com/watch?v=pzBBoJhk1tY)
Credits
Executive Produced by Tzury Bar Yochay (https://twitter.com/tzury)
Produced by Justin Dorfman (https://www.justindorfman.com/)
Edited by Paul M. Bahr at Peachtree Sound (https://www.peachtreesound.com/)
Show notes by DeAnn Bahr at Peachtree Sound (https://www.peachtreesound.com/)
Transcript by Layten Pryce (https://www.fiverr.com/misstranscript)

Transcript
Jérôme [00:01]: When I joined dotCloud, I was joining with a strong infrastructure background. I knew how to run servers and whisper in their ears when things weren't going well. And I did that for maybe two, three years, moving from managing our EC2 Fleet. Then being part of the big engineering mutation, then managing a small team of SREs and then suddenly Docker happened.
Richard [00:28 ]: Hello, and welcome to Committing to Cloud Native. The podcast where we talk about the confluence of cloud native and open source. We have an amazing guest on today. Before we get around to introducing him. I want to introduce the other panelists. So we have Richard Littauer hello, everyone, that guy. And then we have Justin Dorfman. Justin, how are you?
Justin [00:47]: I'm great, Richard. I'm just so glad you're back. Never leave again.
Richard [00:51]: I'll try not to. And Tzury Bar Yochay.
Tzury [00:55]: How are you today?
Richard [00:56]: And our guest today is Jérôme Petazzoni. Jérôme is calling today from Berlin. Jérôme, how are you?
Jérôme [01:03]: That's great. Yeah. Thank you for having me. I'm doing well today. I just got my first [01:08 inaudible] actually this afternoon. So looking forward to positive things in the future.
Richard [01:14]: Well, that is awesome. Welcome. Jérôme is the founder of Tiny Shell script, LLC. He's been a long time operator in the cloud native space. He was part of the team that built scaled and operated dotCloud Paas and before the company became Docker. So you've been around forever. How did you get involved with that company?
Jérôme [01:36]: How did I get involved with a Docker and more specifically dotCloud since it was not called Docker when I joined. Well, I happened to coincidentally work as a soldier of fortune, a long time ago with another consultant named Solomon Hikes. We both worked for the same company back in France, and I don't it was maybe 2005 ish, something like that. And then we stayed in touch. He was starting to do this container thing. And I had this hosting company and he was looking for servers to host his [02:12 weird chritam] channels to do his container things because back then that was complicated.
[02:18]: And on the other hand, I had servers and I was looking for things to run on, said servers. So that was a good match. So basically every around six months we would get together and talk about containers and hack on things. And until eventually at some point in the summer of 2010, I was in Paris and I got a phone call like, "Hey, this is Solomon, long story short, we just raised some money after YC and we are trying to build a team. Would you like to work in California? And I was like, yeah, sure. Sign me up. And six months later, I was in a big plane on my way to San Francisco. And that's how it started.
Richard [02:57]: So many questions. So how did you end up, where were you in this role? How did you start? Where did you segue to? I know that you've given tons and tons of talks and that you're really big on teaching other people how to scale stuff, but you're also a coder in your own right. So like what parts of the code base of Docker have you touched?
Jérôme [03:12]: So initially when I joined dotCloud, I was joining with a strong infrastructure background. I knew how to run servers and whisper in their ears when things weren't going well. And they did that for maybe two, three years, moving from managing [03:30 inaudible], then being part of the bing engineering mutation, then managing a small team of SRS and then suddenly Docker happened.
[03:40]: And we still had this platform that we wanted to maintain because we didn't want to leave all the customers, you know, like, okay, we're closing the past business and put all their energy on this new open source thing. No, we didn't want to do that. So I kind of stayed behind to keep the lights on with another engineer, Andrew [Name04:00] and we did that for maybe a year or so until we transitioned that dotCloud part of the business to someone else and eventually, it's kind of what I'm doing now, should I stay?
Should they go? Is this new fangled go thing that we're doing? Is this something that I want to be a part of? And I wasn't the only one in that situation, like many folks had been kind of sold on the dotCloud project and now it was Docker and it was a completely different company and future.
[04:32]: And then I accidentally gave a talk at the scale conference in LA cause I was living in Pasadena back then, and that was like maybe one hour drive from my home. And it was like, this seems to be a group, like local community conference. I'm going to give a presentation about containers. And it was pretty funny because back then nobody was really excited about containers at all. But just before me, there was a talk by [Name 04:58]  who back then I believe was working maybe at [05:03 Parallels]. I don't know exactly how the company was called back then, but one of the heavy hitters in the container space back then, and it was pretty interesting because he was really deep in the technical details because here is an extremely badass [05:17 hacker]. And at the end he was like, eh, well folks, if you have ideas of what we could do with this containers, like hit me up because it's cool but we don't know exactly what, what we could do with that technology.
[07:28]: I'm like, okay, what do you want to do? Well, can you come to our office and give this talk that you just gave to our team? I'm like, yeah, but we have Friday 4:00 PM. And my flight is, I don't know, like maybe Saturday morning or Sunday morning, something like that. So that seems complicated. No, that's fine, let's hop in a cab and go to the office. And so we do that in a couple of hours of heavy Beijing traffic later, something that Beijing traffic has nothing on any traffic. I'm in this office doing this container presentation in a room of maybe 20/30 Chinese engineers from Baidu. And that was also a very special experience because I was speaking in English, there were a couple of folks taking turns to translate. And then after each let's say paragraph, they were translating and then folks are asking a ton of questions and translators were asking back a bunch of questions as well.
[08:27]: So, you know, very special experience, but fast forward a few months, and we were able to announce some kind of partnership or at least some kind of joint announcement with Baidu, where they said, oh, we're using Docker. And it's great. So for the company that was awesome. Now you take the same story, but you replace a China with Russia. I ended up being invited to the [Yandex conference] in Moscow a few months later and we did the same thing, go there. I didn't need an interpreter. Not because I speak Russian, but because Russians actually speak good English and we can understand each other very well. And we ended up having a joint announcement with Yandex seeing, oh, we're using Docker. So then the joke started to be, oh, we need to send you to a conference in Mountain view so that Google can say they they're using Docker. So that's how I became kind of Docker's evangelist accidentally, you know, almost of a misunderstanding.
Richard [09:21]: You've done a talk when I was at MaxCDN, you did like a little talk there, but how many talks do you think you've given over your career at dotCloud or was it called Docker at the time?
Jérôme [09:33]: At that point, I think it was Docker. And I think the answer is way too many. So I know how many talks I gave between something like July, 2014 and July, 2015 if I remember correctly, because we counted them because we were with my manager, then we were thinking like, okay, let's set our goals for next year. What kind of goals should we set? I was the only [10:00 inaudible] /advocate, evangelist, whatever there. And I didn't know many other advocates or evangelists back then. So it was like, what should my targets and numbers and whatever be.
[10:12]: And so we count and over one year I had the 90 talks, almost 100. Now to be fair, they were alot of optimization there. So I had a few talks that I would give each time I would go to a different city. So for a while it was just intro to docker. At some point I had the Docker and storage drivers one that also, you know, when I was coming back the second time, in a city then, okay that's going to be the dotdoc. So there was a lot of contents we use, and I'm going to the city for a conference and then I'm going to do as well, like speak in the local meetups. So you kind of do a two for one each time, but still that was a lot. And I think that's kind of, part of the thing that led me to burnout.
[10:59]: In a year about 100 talks and that lasted for maybe three or four years, but not the same rhythm. In 2016, for instance, I was doing a workshop every month, plus a talk almost every month or every two months, maybe, something like that. So over the course of time, I was speaking less and putting my energy into different things. And for instance, by 2017, I was doing very little speaking, like just workshops. And I was helping a lot to shape the technical program of Docker con in particular, which we call the black belt tech track. And that was an amazing experience.
Richard [11:41]: I always say for some reason that you were employee number four at Docker, I don't know why, but you were there from like very early on.
Jérôme [11:51]: Yeah, but that's absolutely correct. Well, maybe plus minus one, but basically there were like the two founders back then. So when Sebastian and I joined in the first pool let's say with Sam Alba, Eric [Name] was, if it's weird to say his name I can the English way [12:07 Name] the marker of man. So definitely single digit number of employees.
Richard [12:13]: Yeah. And, you know, it's just interesting because like anyone remembers when you first started, it was like a rocket ship. It was like this company kind of came out of nowhere and kind of define this new way of doing dev ops and which eventually led into cloud native. How was that on the inside looking out, do you feel that Docker kind of paint that path to the cloud native space? Or do you think there's just more to it?
Jérôme [12:45]: I mean, Docker is definitely part of the story and I don't know which metaphor would be the best one to describe its role. You know, is it one of the mini bricks from which we built the pillows that lead us to cloud native or something like that? Maybe it's one of these bricks, but it's kind of stuck in a really tight place right now and then the comfortable one. But yeah, I think using containers the way we do now and kind of bending them to the will of developers if we want, making them super easy to use seeing potential there. That was a pretty strong defining moment for Docker. And then I don't know why it is in the place where it is today.
[13:28]: So some folks are like, well, Docker should be a huge player now. For many folks it feels like now it's more like a, you know, there's the whole enterprise business that [13:38 inaudible]. And then what remains at Docker, all the developer tooling. And I think it's great in a way, because that's where Docker we shown our Docker itself. I mean, as in the Docker engine and then later Docker desktop. So I think it's great that the company is now putting all its energy into these things on which it executed really well in the past. Of course, as a minor shareholder, I'm disappointed that you didn't do casually and digits exist, but such as life.
Richard [14:09]: That's interesting to me that you started off describing yourself as a soldier of fortune, which is like one of my favorite phrases just cause it sounds awesome.
I mean, it means a consultant or a technologist at some point, unless you're actually a mercenary for some military and it didn't like it.
Jérôme [14:24]: No that was absolutely non- violence.
Richard [14:28]: Well, it's also funny because you talk about, you know, I accidentally stumbled into Pasadena. I accidentally went to China and Russia. And to me what's really amazing about cloud native space and Docker in particular is the strength by which open source and the community has really helped lift those containers. It's just really helped make that entire movement possible. And it's because of people like you going forward and doing 90 talks in one year, which would totally lead to burnout because that is a incredibly fast burning fire. Like that is a lot of work. So I'm curious where you see the role of the developer evangelists of people like you, who accidentally randomly do all of this marketing and work in building open source communities in the cloud native space.
Jérôme [15:15]: I think these kind of roles are super important these days, especially for open source companies. And I would even say for any company, who's going to interact with developers because in a way we have a lower attention span. So, you know, if I have five different container storage products to choose from in 10 minutes, basically I expect all of them to give me a two minutes video that's going to explain why they're the best solution. And I'm kind of exaggerating a little bit here, but you get the idea. There's so many information, so many things to pick from that we have a really hard job to kind of distill that information and making it super short, to the point so that folks can be like, okay, I understand, you know, Docker or [16:01 inaudible] or this or that. And now we understand that it is what I want to use or what I don't want to use.So that's part of the thing.
[16:11]: And to me, that was interesting and exciting because I think I loved teaching from relatively young age. I was a TA back in college and I really enjoyed it, for instance. And then later I realized, well, I like explaining things to people, especially something complicated. To me, it's an extremely motivating challenge to be like, all right, this thing, maybe it's complicated or at least we perceive it as complicated. And I'm going to try and break it down in a way that you can understand it and that you can use it. And I want to empower you to use that complicated thing. You know, whether it's Docker or Kubernetes or tomorrow is something else. I think that's something that I really enjoy doing, almost the same way that, you know, when you build some code and you run it and it runs and you're like, yeah, it works.
[17:07]: Almost the same way. It's like, Hey, I'm building this explanation. I'm building this example, I'm building this tutorial, this training content. And when I see people get it, it's kind of the same, maybe like the dopamine rush that you get when your code actually works and does the thing. So here it's like, okay, the code is the combination of the slides, the good samples, everything we put together and the end result is not like computer goes [17:34 berr] but folks understand it and get it. Then the lights goes up and you're like, okay, I've done it.
Richard [17:41]: I mean it's the wetware. It's like, the software is cool and fun to make, but also it doesn't exist without the other technology that humans are really good at, which is language and sharing and talking about how we do things together. It sounds like you're probably a massive proponent of good documentation. Any thoughts on that?
Jérôme [17:57]: I am, yeah, absolutely and I admit that maybe it was a little bit late to the game in a way, because when I was a younger developer, maybe I didn't hold the same opinions. I used to think, well, you know, well commented code might be enough etc. Now in a way I think it might be a [Lori18:16] who gave me this notion of maybe she doesn't call it like that, but read me driven development in a way, which is the idea like when you're going to, especially for, let's say you're a small library or some code or something like that, like something that you're going to put them on GitHub, let's say and share it to the world, kind of start with the read me, because that's how you're going to explain to the world what that is and how it works and how folks should get started.
[18:42]: And I think that's fairly important. Maybe like some folks insist on fried the tests first and then fried the code so that it passes the tests. And to me that would be the documentation. Maybe don't fry the whole documentation first of course, but at least the read me so that you get an idea of how things are going to go together because if it's just a double of code and some files and some comments, it's going to be pretty difficult for folks to make anything useful out of that. So, yeah. Good documentation is important.
Richard [19:17]: Like you're saying, read me driven development. RD, I like that. I'm stealing.
Jérôme [19:24]: And it might have to be attributed to [Lori Name19:26].
Richard [19:27]: I'm stealing that but I'll give you credit for sure.
Justin [19:30]: So one of the things when you're doing RDD or documentation driven development, or any of them is you're steering the ship, you're saying, this is where we want it to go. How do we do that? Which leads me to my next one. That was a non subtle transition. Can you tell us a bit about your involvement with helm?
Jérôme [19:45]: Initially Helm was part of the contents that I added into my training content, because at first it's like, Hey, we should have an option to install things on our Kubernetes [19:56 inaudible]. And from there you go to, okay, now we want to write helm charts. So it gets to be a bigger and bigger chunk. And then a couple of years ago, helm was maybe one hour chunk of content in my training. And now we have a whole thing, like the whole day just dedicated to helm and CACD pipe line and github and things like that. So like continuation of all that.
[20:24]: And the more I got involved with it, the more I appreciated pretty much everything around it, in the design, the way it works, the community, the way that you have like gazillion of helm charts out there, the artifact hub etc etce. And also sure, maybe it's not perfect. And they are probably a few things that folks maybe would do differently otherwise, but it's still pretty good and I'm glad we have helm to manage stuff in the community specifically.
Richard [20:57]: My only beef with Helm charts is charts. It just always threw me off. It's like when I first got into the space, I was like, oh, so you mean like Grafana stuff. You know, it's just very confusing. However, once I'm in the space, I get it. And I've been told many times over that it is the best package manager for Kubernetes apps. Is that the case, like, I don't think there's any other package manager that stands up to it.
Jérôme [21:28]: I would agree. And I could even, in way tone down that a little bit by seeing, well, anyways, the only one we have, so obviously it has to be the best.
Richard [21:38]: Yeah okay, my bad.
Jérôme [21:41]: But yeah, sometimes I say, maybe a stole that from someone that it's kind of the missing package manager for Kubernetes, and sure there is a bunch of things that we can install with just like cube CTL apply some EML and you'll done, but even very simple things, very quickly it turns out that you want to change one little tiny parameter, one little thing. Then you just really appreciate, you have a helm chart rather than curl this CML pipe said [22:13 gypsy TLO, client Euro]. So I really appreciate that we have Helm and even for stuff that is very easy to install now. Lots of projects are putting the work and providing Helm charts. So yeah, I think it totally deserves the title of package manager for kubernetes. And I don't even know if we need to say the best, because again.
Richard [22:37]: Not so arbitrary.
Jérôme [22:39]: But I'm glad that is the one we have because things could have played out worse.
Justin [22:45]: You mentioned that you have like a whole day dedicated to Helm in your trainings, but we also started this by saying you're the founder of Tiny Shell Script LLC. So what is it, the Tiny Shell script does? You go around and train people how to use Docker?
Jérôme [22:58]: Yeah, Tiny Shell Script LLC, it's a one person operation. So when I quit Docker, after taking something like six months off, I was like, okay, I'm going to try training and see if I can actually live from that. And that's doing great. And at some point we're like, okay, how do we do this? I need to have some kind of entity or whatever, I need the name. So basically it was either kind of a [inaudible] or finding something cute and silly like that.  So basically it was either kind of JP Kudzu or finding something cute and silly like that and I took that as a kind of wink to the, you know, we say that maybe, the best [23:36 inaudible] operator from hell was supposed to be like this old grumpy unique, and super silly. When it's super angry it threatens to replace folks with a Tiny Shell Script.
[23:47]: Okay I'm a Tiny Shell Script. So hopefully not in the sense that I want to replace what people do, but in the sense that hopefully what I do is going to eventually be fully automated. And, even if I don't know exactly how we're going to automate training, even though I spend a lot of time thinking about it.
Tzury [24:09]: I would say you will always need eventually a Tiny Shell Script to glue components, which for some reason I know with all the helm in place and all the CIC, you always need to go. You always have these little bash snippets that you're actually using.
Right.
Richard [24:26]: It start with the shell script. Always starts with a shell script, not to cut you off, but I totally agree with you, sir. It just starts with the shell script and it never leaves, you know, it just works sometimes. And I just got to ask you something. So I know from doing my three or four talks, anytime I did a talk, like I got a recruiter saying, Hey, saw you at X. Just want to ask you if you like this position? You probably have offers from every type of company, all in the fangs, Facebook, Amazon. Why go Indy? What made you just say, you know what, I just want to do my own thing. What was it?
Jérôme [25:09]: First as a really quick sidebar. At some point in 2015 when I was kind of questioning myself and what I wanted to do, I went around and interviewed at a few places and that was really interesting. And I totally recommend that folks do that once in a while, because it helps to kind of ground your expectations and everything. So some of these were amazing because you get along super well with everyone and you're like, that's awesome. And sometimes you completely like aced the interviews and sometime you completely [25:37 inaudible], but you know, it's life. And actually the Google interviews were really interesting because everybody's said, okay, it's going to be green, it's going to be like stupid white boarding, when they ask you to put, you know, like crinkly sideways, et.
[25:54]: And I actually found them pretty interesting because yes, there was a lot of white boarding, not going to deny it. But for instance, somebody asked me something about [26:03 inaudible], which is a thing they hadn't touched in 20 plus years. And so I was super honest, I was like, well, I think I've done [26:08 inaudible] when I was in college a little bit, that was part of their grading scores, but I don't remember a single thing about them. They ate Like, okay, fine. [26:16 inaudible] is like this and that. Now can you write functions to the lesson, whatever. And that was fine.
[26:22]: And very interestingly it was, I don't know, half a dozen of interviews. And the only one that went really bad was with somebody who asked me, okay. It was basically something about a scribbling algorithm, but they use containers as an example and then started to ask me, okay, do you know about containers? And I'm like, yeah a little bit. And I was kind of resisting.
Richard [26:44]: Google me,he didn't google your name.Yeah I mean come on.
Jérôme [26:49]: But I'm like, yeah, I know container as a little bit. I worked at Docker. I don't know if you know, Docker. He was like yes. So I was like, okay, good. And then the question was completely unrelated to containers and it was something about the skidding algorithm and I could not come up with the answer we wanted, but what was kind of interesting to me that even though, you know, I tanked one of these, they still gave me an offer, which I didn't get, but it was closed initially. They told me, oh yeah, we can hire you in our office in San Francisco. And I was like, yeah, that's going to be a five minutes walk from my place. And then the offer was something in Mountain view. And I was like, no, that's a one hour commute each way each day.
[27:27]: So no, that's not worth all the money in the world. But that was interesting because that thing helped you kind of, you know, lots of misconceptions I had about the process about the company, about, you walking in. I was effectively a nobody, because even that guy didn't know about, I mean, the link between what I was doing at Docker and then [27:51 inaudible]. So that, was very interesting. Now to go back to answer your question, like, why go Indy and why not join a big company like that? Part of it was because I wanted, at least for a while to work at my own pace after the depression and burnout that I went through in 2017, in the beginning of 2018. I was like, okay, now I want to have the possibility. You know, if at some point I want to kind of take two months off, no questions asked. If I'm independent I can do that. I mean, depending of commitments, etc, of course, but it's easier to do than if I'm in a company.
[28:27]: I also wanted freedom to work both with US-based customers and Europe based customers, which in a way, when I was evangelist [28:36 dev] advocated to the raffle Docker, that's kind of what I was doing because I was very often spending like a few months in Europe and doing round of conferences and workshops, etc. And that was like an incredible opportunity and I wanted to keep that, plus the fact that they have so many family and friends in France, of course. So I wanted the freedom to spend time with them without having to negotiate with my manager etc. Then things worked out really well since 2019. I've been booked, maybe not like solid back to back, but a lot, like more than I was hoping to be.
[29:16]: And I wouldn't even say that since the beginning of the pandemic it's like training business has increased because I mean, my personal theory is that I help folks scale things. And with so many things going online with the pandemic, so many of these companies actually need to scale more and Kubernetes and containers are a huge opportunity for that. So more training needs. Since the beginning of this year, well I took January of, but otherwise I've been training almost nonstop since then.
Justin [29:51]: Do you train your developers in French as well?
Jérôme [29:54]: Both, actually the training I've wrapped up this month. Well, last month in May was in French. We've done a thing with some of my French partners, where we do a kind of training, where we start with, you know, container basics. And we ended up with building Kubernetes cluster from French, and folks can take if they want like 1, 2, 3, 4, or 5 things and some folks get the whole, how do we say the grand slam maybe, which is pretty ambitious. And we do that one in French. But it's the same content, like the slides are in English. It's the same slide as I deliver usually except it's in French now.
Richard [30:29]: So I want you to come to do your training for our company [foriblaze 30:33].  What will be the earliest time slot that you can do it, assuming you go fully booked ahead. Right?
This is an exclusive.
Richard [30:43]: This is a first, we've never had a business offer. This is pretty cool. Go ahead.
Jérôme [30:51]: I believe the earliest would be September. I think I have a week available in September.
Richard [30:57]: Lock it up for us. Guys, listeners, you can't do anything before November. We took the September one.
Jérôme [31:06]: And then October is actually, I have one week free in October because we aligned the next edition of the French one. And we avoided the vacation period. French folks take alot of vacation. So I know we have a week free in October because that's the French vacation. And then things open up again in November.
Richard [31:27]: This is awesome. Did you think you would get like a deal? Like I didn't that [31:32 inaudible] that's pretty cool.
Jérôme [31:33]: We never know, it's the first time it happens like that but you never know.
Justin [31:38]: I know you've done a lot of work around elsewhere. I'm curious.
What's going on with cloud native Islamabab?
Jérôme [31:44]: Yeah, I think it's Patricia from Traffic Labs, connected me with the organizer of the [31:50 inaudible] Islamabad and asked me, Hey, would you like to go in and split [31:54 inaudible] We're like, sure. Why not? And that's it.
Justin [31:58]: Cool. We'll have to have the link for that. I mean you give so many talks that you must have so much, like out there to talk about but we'll drop the link in the show notes, for those of you who want to see that podcast, I think it's really cool that you do stuff globally, that you're not just stuck in San Francisco and you're not just stuck in France. I mean, even though you're calling from Germany. How man languages do you speak?
Jérôme [32:17]: I'm only fluent in French and English. I'm trying to be better with German, but it's a little bit hard and I can say bad things in Italian, Russian, Spanish, not actually communicate.
Justin [32:32]: I was wondering if your Mandarin got any better since that visit to Baidu?
Jérôme [32:37]: Unfortunately. No. So when I was at Docker, I was doing a lot of intentional speaking and I kind of the rule of beginning each doc by saying hello, I'm Jérôme and unfortunately I don't speak the local language, but saying that in the local language. And unfortunately Beijing was when I broke that rule because even though like practicing for hours, I was never able to say it in a way that would be remotely acceptable. And that was weird for some of my friends, because they're like, well, you play music. You have a really good ear for chords and melodies and stuff. So you should be able to pick up languages easily. I'm like apparently this is using different wires in my brain.
Justin [33:19]: We should talk about music. What are we talking about, Docker? I just thought his photo of yours with, is it a drum machine, it's a beat. What is it?
Jérôme [33:28]: It's a launch pad and that became my travel instrument when I was in that like depression, burnout phase. And you know, it's kind of, how do we stay in a kind of hanging to things, et. One of my friends from France come and visit and he had that launchpad and we just tinkered a little bit with it. And I was like, oh, this is really cool. And I feel like I could do some music with it. When I drove him back to the airport on the way back, I stopped by, like guitar center or whatever. And I bought one and I started to use it and make music with it. At some point when I was traveling again for conferences, that was a way to recharge, you know, like you want the conference and you're stemming the bar, it's kind of getting low. And I would kind of dash to my hotel room and plugged the thing to the twig, which is a [34:20 inaudible] do some music and recharge.
[34:24]: And at some point I started thinking, wow, it would be nice if I didn't need the computer to do that. Technically, maybe I could have like a raspberry PI and we would connect the thing to the raspberry PI and the raspberry PI will make the sounds. But that really seemed like science fiction because I couldn't see any combination of software to make that happen. Until at some point again, I'm going to say accidentally stumbled upon the right combination while I realized I could very easily process media messages with that thing. So that was the really easy part. But making sounds, I found something called fluid synth, which takes sound fonts, which are like sound bings, something that I knew like in the late nineties, on the soon blaster [35:10 inaudible] 32 records, that's super old, but that's still around and it works great.
[35:15]: And that way I had my sound making module, so to speak. So then it was just a matter of writing a bunch of code to turn that into a self-standing solution. And yeah, there we go. We have like this little instrument. I spent a lot of time in 2018, like building it mostly by code and then in 2019 kind of playing it around. And now that I'm traveling less, I prefer to play, in like bigger instruments, like a piano, guitar etc. But I still have it handy for when going out in the park or hopefully when travel will be a thing again.
Richard [35:53]: I need to make a quick U-turn and go back to when we were talking about like Kubernetes and Docker. Would Kubernetes be what it is today without Docker. I say no, what do you say?
Jérôme [36:07]: I think I would agree. Yeah, because it's a really good question. It's one of these things where you need a, what if machine, because they have no idea. I mean, would we end up with another kind of container engine or whatever, or, you know, maybe we would have kind of jumped to Kubernetes without the containers or something. I mean, we had Measus, we would Nomad have happened. Things like that. I don't know, like we need a what if machine to answer this question.
Richard [36:39]: Like what Rick and Morty have, like the green gun, where they can go and see what happened in the different dimension.
Tzury [36:45]: I like it. Thank you for asking that question, Justin, because during that time I was able to ascertain that yes Jérôme does have a YouTube channel and yes, there is a video of him playing Zelda with his launch pad, using a raspberry PI and 3000 lines of Python. And it's totally awesome. So everyone should check it out. I just drop it in the show notes and we are running up on time. So before we close up everything, and that is so awesome that you did that by the way. So awesome. Where can people find you online besides this YouTube channel? Where can they hear your words?
Jérôme [37:20]: Well, I have container.training, where I usually announce when I'm going to do new, like conferences, workshops and things like that, except conferences and workshops aren't actually a thing for me anymore this year. So I rant and complain on Twitter, on JP kudzu and I'm also JP kudzu on every other social media out there, basically.
Tzury [37:43]: Awesome. Thank you so much, everyone. Go check that out. Thank you for talking about your training and your work and also for all the work you did in the early days. And like until now, right. It's because of dedicated people like you that cloud native exist at all and because of you, that people learned about it and kept going and could ask you really dumb questions. Like have you heard about containers before, doing interviews? So thank you so much.
Jérôme [38:09]: Thanks for having me.
Richard [38:10]: Thanks, it was great seeing you again.
Special Guest: Jérôme Petazzoni.

Sponsored by Reblaze, creators of Curiefense

Panelists

Justin Dorfman | Richard Littauer | Tzury Bar Yochay

Guest

Jérôme Petazzoni

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. Our amazing guest today is Jérôme Petazzoni, Founder of Tiny Shell Script, LLC, a long-time operator in the Cloud Native space, and was part of the team that built, scaled, and operated dotCloud PaaS before that company became Docker. We learn how Jérôme became involved in Docker, his involvement in Helm, how he loves teaching complicated things to people because he finds it a motivating challenge, and how he has a ReadMe driven development. Jérôme tells us more about what his company Tiny Shell Script does, what he’s doing with Cloud Native Islamabad, and if that’s not enough, he tells us more about how a Launchpad became his travel instrument, and how he can play the theme of Zelda with a Launchpad, Raspberry Pi, and 3000 lines of Python. Download this episode to find out much more from Jérôme!

[00:01:33] Jérôme fills us in how he got involved in Docker, which was dotCloud PaaS when he started, how he ended up doing the job he did there, and what parts of the code base of Docker he’s touched.

[00:09:21] Justin wonders how many talks Jérôme has given in his career at Docker.

[00:13:35] Justin asks Jérôme if he feels that Docker paved that path to the Cloud Native space or if he thinks there’s more to it.

[00:14:58] Richard wonders where Jérôme sees the role of the developer evangelist of people like him, who accidently randomly do all of this marketing and work in building open source communities in the Cloud Native space.

[00:17:48] We find out how Jérôme is a massive proponent of good documentation.

[00:19:42] Learn about Jérôme’s involvement in Helm.

[00:22:53] Jérôme tells us what Tiny Shell Script, LLC does.

[00:24:52] Justin wonders what made Jérôme just go and do his own thing.

[00:29:51] Richard asks if Jérôme trains his developers in French as well, and Tzury asks him if he can come do training at Reblaze.

[00:31:37] Since Jérôme has done a lot of work elsewhere, Richard is curious in what’s going on with Cloud Native Islamabad.

[00:33:18] Jerome tells us about his launchpad that became his travel instrument and making music.

[00:35:55] Justin asks Jérôme if Kubernetes be what it is today without Docker, and Justin says no and wonder if he agrees.

[00:37:12] Find out where you can follow Jérôme online.

Quotes

[00:15:17] “And I would even say for any company who’s going to interact with developers because in a way we have a lower attention span.”

[00:15:29] “So, you know if I have five different containers storage products to choose from in ten minutes, basically I expect all of them to give me a two-minute video that’s going to explain why they’re the best solution.”

[00:18:13] “Now in a way, I think it may be Lauri Apple who gave me this notion of maybe she doesn’t call it like that, but ReadMe driven development in a way.”

[00:18:34] “Kind of start with the ReadMe because that’s how you’re going to explain to the world what that is and how it works and how folks should get started.”

[00:21:43] “Sometimes I say, maybe I stole that from someone, that is kind of the missing package manager for Kubernetes.”

[00:25:11] “At some point in 2015 when I was kind of questioning myself and what I wanted to do, I went around an interviewed at a few places and that was really interesting and I totally recommend to do that once in a while because it helps to kind of ground your expectations and everything.”

[00:26:10] “That was part of their greatest course, but I don’t remember a single thing about them.”

[00:34:43] “Until at some point again, I’m going to say accidentally stumbled upon the right combination while I realized I could very easily process media messages with that thing.”

[00:36:09] “It’s one of these things where you need a “what if machine.”

Links

Curiefense

Curiefense Twitter

Cloud Native Community Groups-Curifense

[email protected]

Reblaze

Justin Dorfman Twitter

Richard Littauer Twitter

Tzury Bar Yochay Twitter

Jérôme Petazzoni Twitter

Jérôme Petazzoni Linkedin

Jérôme Petazzoni Website

Tiny Shell Script, LLC

Docker

Helm

Cloud Native Islamabad

Cloud Native Islamabad Twitter

FluidSynth

Container Training with Jérôme Petazzoni

Zelda with a Launchpad, a Raspberry Pi, and 3000 lines of Python- Jérôme Petazzoni-YouTube

Credits

Executive Produced by Tzury Bar Yochay
Produced by Justin Dorfman
Edited by Paul M. Bahr at Peachtree Sound
Show notes by DeAnn Bahr at Peachtree Sound
Transcript by Layten Pryce

Transcript

Jérôme [00:01]: When I joined dotCloud, I was joining with a strong infrastructure background. I knew how to run servers and whisper in their ears when things weren't going well. And I did that for maybe two, three years, moving from managing our EC2 Fleet. Then being part of the big engineering mutation, then managing a small team of SREs and then suddenly Docker happened.

Richard [00:28 ]: Hello, and welcome to Committing to Cloud Native. The podcast where we talk about the confluence of cloud native and open source. We have an amazing guest on today. Before we get around to introducing him. I want to introduce the other panelists. So we have Richard Littauer hello, everyone, that guy. And then we have Justin Dorfman. Justin, how are you?

Justin [00:47]: I'm great, Richard. I'm just so glad you're back. Never leave again.

Richard [00:51]: I'll try not to. And Tzury Bar Yochay.

Tzury [00:55]: How are you today?

Richard [00:56]: And our guest today is Jérôme Petazzoni. Jérôme is calling today from Berlin. Jérôme, how are you?

Jérôme [01:03]: That's great. Yeah. Thank you for having me. I'm doing well today. I just got my first [01:08 inaudible] actually this afternoon. So looking forward to positive things in the future.

Richard [01:14]: Well, that is awesome. Welcome. Jérôme is the founder of Tiny Shell script, LLC. He's been a long time operator in the cloud native space. He was part of the team that built scaled and operated dotCloud Paas and before the company became Docker. So you've been around forever. How did you get involved with that company?

Jérôme [01:36]: How did I get involved with a Docker and more specifically dotCloud since it was not called Docker when I joined. Well, I happened to coincidentally work as a soldier of fortune, a long time ago with another consultant named Solomon Hikes. We both worked for the same company back in France, and I don't it was maybe 2005 ish, something like that. And then we stayed in touch. He was starting to do this container thing. And I had this hosting company and he was looking for servers to host his [02:12 weird chritam] channels to do his container things because back then that was complicated.

[02:18]: And on the other hand, I had servers and I was looking for things to run on, said servers. So that was a good match. So basically every around six months we would get together and talk about containers and hack on things. And until eventually at some point in the summer of 2010, I was in Paris and I got a phone call like, "Hey, this is Solomon, long story short, we just raised some money after YC and we are trying to build a team. Would you like to work in California? And I was like, yeah, sure. Sign me up. And six months later, I was in a big plane on my way to San Francisco. And that's how it started.

Richard [02:57]: So many questions. So how did you end up, where were you in this role? How did you start? Where did you segue to? I know that you've given tons and tons of talks and that you're really big on teaching other people how to scale stuff, but you're also a coder in your own right. So like what parts of the code base of Docker have you touched?

Jérôme [03:12]: So initially when I joined dotCloud, I was joining with a strong infrastructure background. I knew how to run servers and whisper in their ears when things weren't going well. And they did that for maybe two, three years, moving from managing [03:30 inaudible], then being part of the bing engineering mutation, then managing a small team of SRS and then suddenly Docker happened.

[03:40]: And we still had this platform that we wanted to maintain because we didn't want to leave all the customers, you know, like, okay, we're closing the past business and put all their energy on this new open source thing. No, we didn't want to do that. So I kind of stayed behind to keep the lights on with another engineer, Andrew [Name04:00] and we did that for maybe a year or so until we transitioned that dotCloud part of the business to someone else and eventually, it's kind of what I'm doing now, should I stay?

Should they go? Is this new fangled go thing that we're doing? Is this something that I want to be a part of? And I wasn't the only one in that situation, like many folks had been kind of sold on the dotCloud project and now it was Docker and it was a completely different company and future.

[04:32]: And then I accidentally gave a talk at the scale conference in LA cause I was living in Pasadena back then, and that was like maybe one hour drive from my home. And it was like, this seems to be a group, like local community conference. I'm going to give a presentation about containers. And it was pretty funny because back then nobody was really excited about containers at all. But just before me, there was a talk by [Name 04:58]  who back then I believe was working maybe at [05:03 Parallels]. I don't know exactly how the company was called back then, but one of the heavy hitters in the container space back then, and it was pretty interesting because he was really deep in the technical details because here is an extremely badass [05:17 hacker]. And at the end he was like, eh, well folks, if you have ideas of what we could do with this containers, like hit me up because it's cool but we don't know exactly what, what we could do with that technology.

[07:28]: I'm like, okay, what do you want to do? Well, can you come to our office and give this talk that you just gave to our team? I'm like, yeah, but we have Friday 4:00 PM. And my flight is, I don't know, like maybe Saturday morning or Sunday morning, something like that. So that seems complicated. No, that's fine, let's hop in a cab and go to the office. And so we do that in a couple of hours of heavy Beijing traffic later, something that Beijing traffic has nothing on any traffic. I'm in this office doing this container presentation in a room of maybe 20/30 Chinese engineers from Baidu. And that was also a very special experience because I was speaking in English, there were a couple of folks taking turns to translate. And then after each let's say paragraph, they were translating and then folks are asking a ton of questions and translators were asking back a bunch of questions as well.

[08:27]: So, you know, very special experience, but fast forward a few months, and we were able to announce some kind of partnership or at least some kind of joint announcement with Baidu, where they said, oh, we're using Docker. And it's great. So for the company that was awesome. Now you take the same story, but you replace a China with Russia. I ended up being invited to the [Yandex conference] in Moscow a few months later and we did the same thing, go there. I didn't need an interpreter. Not because I speak Russian, but because Russians actually speak good English and we can understand each other very well. And we ended up having a joint announcement with Yandex seeing, oh, we're using Docker. So then the joke started to be, oh, we need to send you to a conference in Mountain view so that Google can say they they're using Docker. So that's how I became kind of Docker's evangelist accidentally, you know, almost of a misunderstanding.

Richard [09:21]: You've done a talk when I was at MaxCDN, you did like a little talk there, but how many talks do you think you've given over your career at dotCloud or was it called Docker at the time?

Jérôme [09:33]: At that point, I think it was Docker. And I think the answer is way too many. So I know how many talks I gave between something like July, 2014 and July, 2015 if I remember correctly, because we counted them because we were with my manager, then we were thinking like, okay, let's set our goals for next year. What kind of goals should we set? I was the only [10:00 inaudible] /advocate, evangelist, whatever there. And I didn't know many other advocates or evangelists back then. So it was like, what should my targets and numbers and whatever be.

[10:12]: And so we count and over one year I had the 90 talks, almost 100. Now to be fair, they were alot of optimization there. So I had a few talks that I would give each time I would go to a different city. So for a while it was just intro to docker. At some point I had the Docker and storage drivers one that also, you know, when I was coming back the second time, in a city then, okay that's going to be the dotdoc. So there was a lot of contents we use, and I'm going to the city for a conference and then I'm going to do as well, like speak in the local meetups. So you kind of do a two for one each time, but still that was a lot. And I think that's kind of, part of the thing that led me to burnout.

[10:59]: In a year about 100 talks and that lasted for maybe three or four years, but not the same rhythm. In 2016, for instance, I was doing a workshop every month, plus a talk almost every month or every two months, maybe, something like that. So over the course of time, I was speaking less and putting my energy into different things. And for instance, by 2017, I was doing very little speaking, like just workshops. And I was helping a lot to shape the technical program of Docker con in particular, which we call the black belt tech track. And that was an amazing experience.

Richard [11:41]: I always say for some reason that you were employee number four at Docker, I don't know why, but you were there from like very early on.

Jérôme [11:51]: Yeah, but that's absolutely correct. Well, maybe plus minus one, but basically there were like the two founders back then. So when Sebastian and I joined in the first pool let's say with Sam Alba, Eric [Name] was, if it's weird to say his name I can the English way [12:07 Name] the marker of man. So definitely single digit number of employees.

Richard [12:13]: Yeah. And, you know, it's just interesting because like anyone remembers when you first started, it was like a rocket ship. It was like this company kind of came out of nowhere and kind of define this new way of doing dev ops and which eventually led into cloud native. How was that on the inside looking out, do you feel that Docker kind of paint that path to the cloud native space? Or do you think there's just more to it?

Jérôme [12:45]: I mean, Docker is definitely part of the story and I don't know which metaphor would be the best one to describe its role. You know, is it one of the mini bricks from which we built the pillows that lead us to cloud native or something like that? Maybe it's one of these bricks, but it's kind of stuck in a really tight place right now and then the comfortable one. But yeah, I think using containers the way we do now and kind of bending them to the will of developers if we want, making them super easy to use seeing potential there. That was a pretty strong defining moment for Docker. And then I don't know why it is in the place where it is today.

[13:28]: So some folks are like, well, Docker should be a huge player now. For many folks it feels like now it's more like a, you know, there's the whole enterprise business that [13:38 inaudible]. And then what remains at Docker, all the developer tooling. And I think it's great in a way, because that's where Docker we shown our Docker itself. I mean, as in the Docker engine and then later Docker desktop. So I think it's great that the company is now putting all its energy into these things on which it executed really well in the past. Of course, as a minor shareholder, I'm disappointed that you didn't do casually and digits exist, but such as life.

Richard [14:09]: That's interesting to me that you started off describing yourself as a soldier of fortune, which is like one of my favorite phrases just cause it sounds awesome.

I mean, it means a consultant or a technologist at some point, unless you're actually a mercenary for some military and it didn't like it.

Jérôme [14:24]: No that was absolutely non- violence.

Richard [14:28]: Well, it's also funny because you talk about, you know, I accidentally stumbled into Pasadena. I accidentally went to China and Russia. And to me what's really amazing about cloud native space and Docker in particular is the strength by which open source and the community has really helped lift those containers. It's just really helped make that entire movement possible. And it's because of people like you going forward and doing 90 talks in one year, which would totally lead to burnout because that is a incredibly fast burning fire. Like that is a lot of work. So I'm curious where you see the role of the developer evangelists of people like you, who accidentally randomly do all of this marketing and work in building open source communities in the cloud native space.

Jérôme [15:15]: I think these kind of roles are super important these days, especially for open source companies. And I would even say for any company, who's going to interact with developers because in a way we have a lower attention span. So, you know, if I have five different container storage products to choose from in 10 minutes, basically I expect all of them to give me a two minutes video that's going to explain why they're the best solution. And I'm kind of exaggerating a little bit here, but you get the idea. There's so many information, so many things to pick from that we have a really hard job to kind of distill that information and making it super short, to the point so that folks can be like, okay, I understand, you know, Docker or [16:01 inaudible] or this or that. And now we understand that it is what I want to use or what I don't want to use.So that's part of the thing.

[16:11]: And to me, that was interesting and exciting because I think I loved teaching from relatively young age. I was a TA back in college and I really enjoyed it, for instance. And then later I realized, well, I like explaining things to people, especially something complicated. To me, it's an extremely motivating challenge to be like, all right, this thing, maybe it's complicated or at least we perceive it as complicated. And I'm going to try and break it down in a way that you can understand it and that you can use it. And I want to empower you to use that complicated thing. You know, whether it's Docker or Kubernetes or tomorrow is something else. I think that's something that I really enjoy doing, almost the same way that, you know, when you build some code and you run it and it runs and you're like, yeah, it works.

[17:07]: Almost the same way. It's like, Hey, I'm building this explanation. I'm building this example, I'm building this tutorial, this training content. And when I see people get it, it's kind of the same, maybe like the dopamine rush that you get when your code actually works and does the thing. So here it's like, okay, the code is the combination of the slides, the good samples, everything we put together and the end result is not like computer goes [17:34 berr] but folks understand it and get it. Then the lights goes up and you're like, okay, I've done it.

Richard [17:41]: I mean it's the wetware. It's like, the software is cool and fun to make, but also it doesn't exist without the other technology that humans are really good at, which is language and sharing and talking about how we do things together. It sounds like you're probably a massive proponent of good documentation. Any thoughts on that?

Jérôme [17:57]: I am, yeah, absolutely and I admit that maybe it was a little bit late to the game in a way, because when I was a younger developer, maybe I didn't hold the same opinions. I used to think, well, you know, well commented code might be enough etc. Now in a way I think it might be a [Lori18:16] who gave me this notion of maybe she doesn't call it like that, but read me driven development in a way, which is the idea like when you're going to, especially for, let's say you're a small library or some code or something like that, like something that you're going to put them on GitHub, let's say and share it to the world, kind of start with the read me, because that's how you're going to explain to the world what that is and how it works and how folks should get started.

[18:42]: And I think that's fairly important. Maybe like some folks insist on fried the tests first and then fried the code so that it passes the tests. And to me that would be the documentation. Maybe don't fry the whole documentation first of course, but at least the read me so that you get an idea of how things are going to go together because if it's just a double of code and some files and some comments, it's going to be pretty difficult for folks to make anything useful out of that. So, yeah. Good documentation is important.

Richard [19:17]: Like you're saying, read me driven development. RD, I like that. I'm stealing.

Jérôme [19:24]: And it might have to be attributed to [Lori Name19:26].

Richard [19:27]: I'm stealing that but I'll give you credit for sure.

Justin [19:30]: So one of the things when you're doing RDD or documentation driven development, or any of them is you're steering the ship, you're saying, this is where we want it to go. How do we do that? Which leads me to my next one. That was a non subtle transition. Can you tell us a bit about your involvement with helm?

Jérôme [19:45]: Initially Helm was part of the contents that I added into my training content, because at first it's like, Hey, we should have an option to install things on our Kubernetes [19:56 inaudible]. And from there you go to, okay, now we want to write helm charts. So it gets to be a bigger and bigger chunk. And then a couple of years ago, helm was maybe one hour chunk of content in my training. And now we have a whole thing, like the whole day just dedicated to helm and CACD pipe line and github and things like that. So like continuation of all that.

[20:24]: And the more I got involved with it, the more I appreciated pretty much everything around it, in the design, the way it works, the community, the way that you have like gazillion of helm charts out there, the artifact hub etc etce. And also sure, maybe it's not perfect. And they are probably a few things that folks maybe would do differently otherwise, but it's still pretty good and I'm glad we have helm to manage stuff in the community specifically.

Richard [20:57]: My only beef with Helm charts is charts. It just always threw me off. It's like when I first got into the space, I was like, oh, so you mean like Grafana stuff. You know, it's just very confusing. However, once I'm in the space, I get it. And I've been told many times over that it is the best package manager for Kubernetes apps. Is that the case, like, I don't think there's any other package manager that stands up to it.

Jérôme [21:28]: I would agree. And I could even, in way tone down that a little bit by seeing, well, anyways, the only one we have, so obviously it has to be the best.

Richard [21:38]: Yeah okay, my bad.

Jérôme [21:41]: But yeah, sometimes I say, maybe a stole that from someone that it's kind of the missing package manager for Kubernetes, and sure there is a bunch of things that we can install with just like cube CTL apply some EML and you'll done, but even very simple things, very quickly it turns out that you want to change one little tiny parameter, one little thing. Then you just really appreciate, you have a helm chart rather than curl this CML pipe said [22:13 gypsy TLO, client Euro]. So I really appreciate that we have Helm and even for stuff that is very easy to install now. Lots of projects are putting the work and providing Helm charts. So yeah, I think it totally deserves the title of package manager for kubernetes. And I don't even know if we need to say the best, because again.

Richard [22:37]: Not so arbitrary.

Jérôme [22:39]: But I'm glad that is the one we have because things could have played out worse.

Justin [22:45]: You mentioned that you have like a whole day dedicated to Helm in your trainings, but we also started this by saying you're the founder of Tiny Shell Script LLC. So what is it, the Tiny Shell script does? You go around and train people how to use Docker?

Jérôme [22:58]: Yeah, Tiny Shell Script LLC, it's a one person operation. So when I quit Docker, after taking something like six months off, I was like, okay, I'm going to try training and see if I can actually live from that. And that's doing great. And at some point we're like, okay, how do we do this? I need to have some kind of entity or whatever, I need the name. So basically it was either kind of a [inaudible] or finding something cute and silly like that.  So basically it was either kind of JP Kudzu or finding something cute and silly like that and I took that as a kind of wink to the, you know, we say that maybe, the best [23:36 inaudible] operator from hell was supposed to be like this old grumpy unique, and super silly. When it's super angry it threatens to replace folks with a Tiny Shell Script.

[23:47]: Okay I'm a Tiny Shell Script. So hopefully not in the sense that I want to replace what people do, but in the sense that hopefully what I do is going to eventually be fully automated. And, even if I don't know exactly how we're going to automate training, even though I spend a lot of time thinking about it.

Tzury [24:09]: I would say you will always need eventually a Tiny Shell Script to glue components, which for some reason I know with all the helm in place and all the CIC, you always need to go. You always have these little bash snippets that you're actually using.

Right.

Richard [24:26]: It start with the shell script. Always starts with a shell script, not to cut you off, but I totally agree with you, sir. It just starts with the shell script and it never leaves, you know, it just works sometimes. And I just got to ask you something. So I know from doing my three or four talks, anytime I did a talk, like I got a recruiter saying, Hey, saw you at X. Just want to ask you if you like this position? You probably have offers from every type of company, all in the fangs, Facebook, Amazon. Why go Indy? What made you just say, you know what, I just want to do my own thing. What was it?

Jérôme [25:09]: First as a really quick sidebar. At some point in 2015 when I was kind of questioning myself and what I wanted to do, I went around and interviewed at a few places and that was really interesting. And I totally recommend that folks do that once in a while, because it helps to kind of ground your expectations and everything. So some of these were amazing because you get along super well with everyone and you're like, that's awesome. And sometimes you completely like aced the interviews and sometime you completely [25:37 inaudible], but you know, it's life. And actually the Google interviews were really interesting because everybody's said, okay, it's going to be green, it's going to be like stupid white boarding, when they ask you to put, you know, like crinkly sideways, et.

[25:54]: And I actually found them pretty interesting because yes, there was a lot of white boarding, not going to deny it. But for instance, somebody asked me something about [26:03 inaudible], which is a thing they hadn't touched in 20 plus years. And so I was super honest, I was like, well, I think I've done [26:08 inaudible] when I was in college a little bit, that was part of their grading scores, but I don't remember a single thing about them. They ate Like, okay, fine. [26:16 inaudible] is like this and that. Now can you write functions to the lesson, whatever. And that was fine.

[26:22]: And very interestingly it was, I don't know, half a dozen of interviews. And the only one that went really bad was with somebody who asked me, okay. It was basically something about a scribbling algorithm, but they use containers as an example and then started to ask me, okay, do you know about containers? And I'm like, yeah a little bit. And I was kind of resisting.

Richard [26:44]: Google me,he didn't google your name.Yeah I mean come on.

Jérôme [26:49]: But I'm like, yeah, I know container as a little bit. I worked at Docker. I don't know if you know, Docker. He was like yes. So I was like, okay, good. And then the question was completely unrelated to containers and it was something about the skidding algorithm and I could not come up with the answer we wanted, but what was kind of interesting to me that even though, you know, I tanked one of these, they still gave me an offer, which I didn't get, but it was closed initially. They told me, oh yeah, we can hire you in our office in San Francisco. And I was like, yeah, that's going to be a five minutes walk from my place. And then the offer was something in Mountain view. And I was like, no, that's a one hour commute each way each day.

[27:27]: So no, that's not worth all the money in the world. But that was interesting because that thing helped you kind of, you know, lots of misconceptions I had about the process about the company, about, you walking in. I was effectively a nobody, because even that guy didn't know about, I mean, the link between what I was doing at Docker and then [27:51 inaudible]. So that, was very interesting. Now to go back to answer your question, like, why go Indy and why not join a big company like that? Part of it was because I wanted, at least for a while to work at my own pace after the depression and burnout that I went through in 2017, in the beginning of 2018. I was like, okay, now I want to have the possibility. You know, if at some point I want to kind of take two months off, no questions asked. If I'm independent I can do that. I mean, depending of commitments, etc, of course, but it's easier to do than if I'm in a company.

[28:27]: I also wanted freedom to work both with US-based customers and Europe based customers, which in a way, when I was evangelist [28:36 dev] advocated to the raffle Docker, that's kind of what I was doing because I was very often spending like a few months in Europe and doing round of conferences and workshops, etc. And that was like an incredible opportunity and I wanted to keep that, plus the fact that they have so many family and friends in France, of course. So I wanted the freedom to spend time with them without having to negotiate with my manager etc. Then things worked out really well since 2019. I've been booked, maybe not like solid back to back, but a lot, like more than I was hoping to be.

[29:16]: And I wouldn't even say that since the beginning of the pandemic it's like training business has increased because I mean, my personal theory is that I help folks scale things. And with so many things going online with the pandemic, so many of these companies actually need to scale more and Kubernetes and containers are a huge opportunity for that. So more training needs. Since the beginning of this year, well I took January of, but otherwise I've been training almost nonstop since then.

Justin [29:51]: Do you train your developers in French as well?

Jérôme [29:54]: Both, actually the training I've wrapped up this month. Well, last month in May was in French. We've done a thing with some of my French partners, where we do a kind of training, where we start with, you know, container basics. And we ended up with building Kubernetes cluster from French, and folks can take if they want like 1, 2, 3, 4, or 5 things and some folks get the whole, how do we say the grand slam maybe, which is pretty ambitious. And we do that one in French. But it's the same content, like the slides are in English. It's the same slide as I deliver usually except it's in French now.

Richard [30:29]: So I want you to come to do your training for our company [foriblaze 30:33].  What will be the earliest time slot that you can do it, assuming you go fully booked ahead. Right?

This is an exclusive.

Richard [30:43]: This is a first, we've never had a business offer. This is pretty cool. Go ahead.

Jérôme [30:51]: I believe the earliest would be September. I think I have a week available in September.

Richard [30:57]: Lock it up for us. Guys, listeners, you can't do anything before November. We took the September one.

Jérôme [31:06]: And then October is actually, I have one week free in October because we aligned the next edition of the French one. And we avoided the vacation period. French folks take alot of vacation. So I know we have a week free in October because that's the French vacation. And then things open up again in November.

Richard [31:27]: This is awesome. Did you think you would get like a deal? Like I didn't that [31:32 inaudible] that's pretty cool.

Jérôme [31:33]: We never know, it's the first time it happens like that but you never know.

Justin [31:38]: I know you've done a lot of work around elsewhere. I'm curious.

What's going on with cloud native Islamabab?

Jérôme [31:44]: Yeah, I think it's Patricia from Traffic Labs, connected me with the organizer of the [31:50 inaudible] Islamabad and asked me, Hey, would you like to go in and split [31:54 inaudible] We're like, sure. Why not? And that's it.

Justin [31:58]: Cool. We'll have to have the link for that. I mean you give so many talks that you must have so much, like out there to talk about but we'll drop the link in the show notes, for those of you who want to see that podcast, I think it's really cool that you do stuff globally, that you're not just stuck in San Francisco and you're not just stuck in France. I mean, even though you're calling from Germany. How man languages do you speak?

Jérôme [32:17]: I'm only fluent in French and English. I'm trying to be better with German, but it's a little bit hard and I can say bad things in Italian, Russian, Spanish, not actually communicate.

Justin [32:32]: I was wondering if your Mandarin got any better since that visit to Baidu?

Jérôme [32:37]: Unfortunately. No. So when I was at Docker, I was doing a lot of intentional speaking and I kind of the rule of beginning each doc by saying hello, I'm Jérôme and unfortunately I don't speak the local language, but saying that in the local language. And unfortunately Beijing was when I broke that rule because even though like practicing for hours, I was never able to say it in a way that would be remotely acceptable. And that was weird for some of my friends, because they're like, well, you play music. You have a really good ear for chords and melodies and stuff. So you should be able to pick up languages easily. I'm like apparently this is using different wires in my brain.

Justin [33:19]: We should talk about music. What are we talking about, Docker? I just thought his photo of yours with, is it a drum machine, it's a beat. What is it?

Jérôme [33:28]: It's a launch pad and that became my travel instrument when I was in that like depression, burnout phase. And you know, it's kind of, how do we stay in a kind of hanging to things, et. One of my friends from France come and visit and he had that launchpad and we just tinkered a little bit with it. And I was like, oh, this is really cool. And I feel like I could do some music with it. When I drove him back to the airport on the way back, I stopped by, like guitar center or whatever. And I bought one and I started to use it and make music with it. At some point when I was traveling again for conferences, that was a way to recharge, you know, like you want the conference and you're stemming the bar, it's kind of getting low. And I would kind of dash to my hotel room and plugged the thing to the twig, which is a [34:20 inaudible] do some music and recharge.

[34:24]: And at some point I started thinking, wow, it would be nice if I didn't need the computer to do that. Technically, maybe I could have like a raspberry PI and we would connect the thing to the raspberry PI and the raspberry PI will make the sounds. But that really seemed like science fiction because I couldn't see any combination of software to make that happen. Until at some point again, I'm going to say accidentally stumbled upon the right combination while I realized I could very easily process media messages with that thing. So that was the really easy part. But making sounds, I found something called fluid synth, which takes sound fonts, which are like sound bings, something that I knew like in the late nineties, on the soon blaster [35:10 inaudible] 32 records, that's super old, but that's still around and it works great.

[35:15]: And that way I had my sound making module, so to speak. So then it was just a matter of writing a bunch of code to turn that into a self-standing solution. And yeah, there we go. We have like this little instrument. I spent a lot of time in 2018, like building it mostly by code and then in 2019 kind of playing it around. And now that I'm traveling less, I prefer to play, in like bigger instruments, like a piano, guitar etc. But I still have it handy for when going out in the park or hopefully when travel will be a thing again.

Richard [35:53]: I need to make a quick U-turn and go back to when we were talking about like Kubernetes and Docker. Would Kubernetes be what it is today without Docker. I say no, what do you say?

Jérôme [36:07]: I think I would agree. Yeah, because it's a really good question. It's one of these things where you need a, what if machine, because they have no idea. I mean, would we end up with another kind of container engine or whatever, or, you know, maybe we would have kind of jumped to Kubernetes without the containers or something. I mean, we had Measus, we would Nomad have happened. Things like that. I don't know, like we need a what if machine to answer this question.

Richard [36:39]: Like what Rick and Morty have, like the green gun, where they can go and see what happened in the different dimension.

Tzury [36:45]: I like it. Thank you for asking that question, Justin, because during that time I was able to ascertain that yes Jérôme does have a YouTube channel and yes, there is a video of him playing Zelda with his launch pad, using a raspberry PI and 3000 lines of Python. And it's totally awesome. So everyone should check it out. I just drop it in the show notes and we are running up on time. So before we close up everything, and that is so awesome that you did that by the way. So awesome. Where can people find you online besides this YouTube channel? Where can they hear your words?

Jérôme [37:20]: Well, I have container.training, where I usually announce when I'm going to do new, like conferences, workshops and things like that, except conferences and workshops aren't actually a thing for me anymore this year. So I rant and complain on Twitter, on JP kudzu and I'm also JP kudzu on every other social media out there, basically.

Tzury [37:43]: Awesome. Thank you so much, everyone. Go check that out. Thank you for talking about your training and your work and also for all the work you did in the early days. And like until now, right. It's because of dedicated people like you that cloud native exist at all and because of you, that people learned about it and kept going and could ask you really dumb questions. Like have you heard about containers before, doing interviews? So thank you so much.

Jérôme [38:09]: Thanks for having me.

Richard [38:10]: Thanks, it was great seeing you again.

Special Guest: Jérôme Petazzoni.

Sponsored By:

Curiefense: Protect Your Cloud Native Applications 👉 Try Curiefense today!