Phil Estes "The King of containerd" joins us as we discuss his journey in the cloud native and the success of the containerd project from inception to its stance as an amazing container runtime for kubernetes. He discusses how being open source made him an better developer and much much more! Phil Estes is cloud native royalty and i had a great time catching up with him
Timeline/Topic
00:00 Season 2 Opener /goraft.tech
01:47 - Show Opener
02:09 - Phil Estes - King of containerd
02:50 - Phil's Journey
07:58 - Linux then containers..
11:10 - The king goes deep with Containers
14:32 - Containerd (History and what is it?)
21:43 - Open Container Initiative
25:45 - Phil likes Biking and how it helps him solve problems
28:23 - Drones....
30:48 - King Estes... 5 star chef
33:49 - What work if King Phil Estes most proud of?
Episode Links
https://containerd.io/
https://opencontainers.org/
https://kubernetes.io/blog/2018/05/24/kubernetes-containerd-integration-goes-ga/
POPCAST SHOW DETAILS
Watch (YouTube): https://www.youtube.com/c/thepopcastpop
Listen (Apple PODCAST and others): https://popcast-d9f7b6dc.simplecast.com
Follow us on (Twitter): https://twitter.com/PopcastPop
Follow us on (Linkedin): https://www.linkedin.com/company/the-popcast-with-danpop
SPONSORED BY RAFT - Raft is a full stack digital consulting firm for delivery of human-centered and highly available data-intensive software solutions. Raft partners with government agencies like HHS and the DoD to accelerate innovation at the national level. As “non-traditional” as you can get, their team’s experience ranges from Walmart and the NSA to Microsoft and Buffalo Wild Wings. They brought Kubernetes into the federal space before it was cool, and are pushing the limits of Kafka every day. Combining open source and agile delivery with HCD and DevSecOps, Raft is breaking the status quo in GovTech for good. Check them out: https://goraft.tech/
Podchaser: 3iCM9Pu0eQmzYxvWF7dZ
POP (00:00):
Hi everyone. Welcome to season two. So excited to be back with you all and doing this. This is, you know, again, this is community led. I've said it a billion times, but, you know, that I do this for you. So, um, this year's goal, we're just trying to shoot for $10,000 to charity, and so, you know, definitely if you're in a cloud native space and wanna reach out to me via DM, it's @PopcastPop so that's, uh, on Twitter. It's @-P-O-P-C-A-S-T P-O-P, if you're interested in the cloud native space and you're interested in sponsoring the show. Uh, we have some really big, uh, sponsors we're gonna announce then going forward.
POP (00:34):
Um, one of those sponsors is Raft. Raft is really cool. Yup.
POP (00:38):
Got a nice little coffee cup from you all. Um, teams been in Bara-Barak. Those teams have just been fantastic. So, they're like a full stack um, digital consultancy company, like, more human centered like, data intensive applications and-and, you know, they do government agencies like, you know, DOD and, uh, that type of thing. So they're trying to do that in-in the government and um, and also it's, it's pretty non-traditional. They have, like, experience with folks from Microsoft and like, what, you know, Buffalo Wild Wings as an example. So it's, you know, they were bringing, um, Kubernetes into the federal space really early. Very excited to be partnering for them. Uh, they'll begin sponsoring the whole month of uh, February, so it's definitely cool to do that. So, uh, definitely reach out to them. It's GoRaft.tech, so it's, uh, G-O-R-A-F-T dot tech, and they'll be in the liner notes of the episode as well.
POP (01:26):
So, uh, this is the first episode I could not have picked a better guest. Uh, this is Phil Estes. He's in IAWS now. He was at IBM. Uh, he's the..I call him the king of containerd. Just a fantastic dude, and uh, I'm so excited. Let's-let's, stop me ramblin'. Let's get on with this. Let's go. So, now on with the Popcast, you all. Thank you so much for, for enjoying it, and please like and subscribe.
POP (01:46):
(intro music)
POP (02:09):
Hello everyone and welcome to the Popcast. Today this is my old friend. This is Phil Estes. He's the distinguished engineer at IBM, he's a container, uh, he's the king of Containerd. Let's just call him what he is. He's one of the maintainers over there. It's a great project, um. He's the OCI, uh, uh, top chairman. Great dude and we're gonna talk about, you know, his journey. We're gonna talk about containerd and a lot of other things.
POP (02:32):
Welcome to the show Phil Estes.
Phil Estes (02:35):
Awesome. Great to be here. I, I feel like I've joined some pretty amazing company that have been on the Popcast. So, uh, yeah, great to be here.
POP (02:44):
You were always on my mind and in my heart, man. You know that, so it's all good.
Phil Estes (02:47):
(laughs)
POP (02:48):
So, let's, let's talk about, listen, before we had containers in life, right? Which, was there a before time? No.
Phil Estes (02:53):
Yeah, I think so.
POP (02:54):
Uh, let's talk about the journey, then. Let's talk about, like, how you started on computing and all the way to here.
Phil Estes (03:01):
Yeah. So um, you know, I'm old enough to-to be from that land before time, before Windows and PC's. I mean, okay, PC's were starting to show up when I was in, like junior high, and um, I-I went to the, uh, I-I remember you askin' Liz Rice about RadioShack, whether she knew what that was, you know, in the UK, but ... You know, uh, I remember, literally remember walking to the mall after school to like, play with like, the TRS-80, and then later the Tandy 1000, and...
POP (03:33):
I loved the Tandy 1000. Like, you know, it was like, you all don't know the pain of like, we're playing a game at RG-..like really below level RGB. What was it like, you know, like, with these pinks and blues. Oh, man, it's such bad color pallets. Sorry, y'all.
Phil Estes (03:48):
Yeah. Yeah, that was, it was pretty crazy. So, um, yeah. The cool thing was, like, obviously my family knew I'd love to have one and it wasn't something we could really, you know, figure out. You know, computers were expensive back then. It was like buying a MacBook Pro today. Top of the line, uh, (laughs) which I guess, you know, today people fork that kind of money out easily. But back then, it was like ugh, I don't know if we can do that. But, uh, Christmastime, one year my grandpa showed up with a TRS-80 and uh, I guess I've never been that far from a keyboard since then. I think, you know I-I just kinda got hooked on oh, I can learn basic and I can make things show up on the screen. I can play, you know, goofy noises, um, and of course play games and all that. So, uh..
POP (04:34):
Games played such a, I feel like it's this tail that just weaves its way, the string that weaves its way through all of the guests I've had. It's like, you know early folks again, like you and like me, like we basically had like these, you know, Byte Magazine, where you could actually do the, you know, the codes and that thing..
Phil Estes (04:50):
(laughs)
POP (04:52):
You know I had like, Joe Beda and those, and-and-and-and Brenda Burns and it's the same narrative. It's like you try to play games, try to hack it into making it work and that led to the, a career. So..
Phil Estes (05:01):
Yeah, yeah, yeah. So, yeah, I-I've never been too far from computers since then I, you know, wanted to do, I actually had an interest in hardware. I went to college for computer engineering, which again, had more of the electrical and the design classes. So, I kinda came out, um, came out of-of college having worked some semesters at IBM as a Co-Op and all I'd ever done is software, but what I really wanted to do was hardware, but it-the mid 90's were kind of a, you know, crazy time to try and jump into the hardware industry 'cause everybody's like, you know, oh, it's all going to be software in the future.
Phil Estes (05:40):
You know the hardware engineers were guys in their 40's and 50's at that time, you know, they were the experts and they were, you know, but, I think there's, you know, there's maybe been a-a-ah-revival you know, with Edge, and IOT, and you know, even in the hobbiest space with Raspberry Pie, but anyway, in that era, it was like, oh, rejection letter, rejection letter, um, and so I ended up working in software, because I'd really, I'd literally, couldn't find a job doing my degree again, focused, on the hardware design aspect. So, it was interesting to come into IBM, so I came, you know, there the one company that said, "Hey, you worked here a couple of summers. People said you weren't that bad. (laughs) We'll take you back." Uh, so I joined, you know, that fall after-after college, um, but I fell kind of behind the eight ball because I hadn't taken a bunch of CS classes, I mean, I, again, I monkeyed around computers for 10 years by this time. Um...
POP (06:45):
You think that made you a better engineer? Like, or, a better actual developer you think? 'Cause you understand the bits and bytes, like the underlying...
Phil Estes (06:53):
Yeah, it's interesting you mention that because, while I did struggle early on with kind of some of the metaphors of-of code and, you know I didn't have a lot of the data structures background. I couldn't right you a sort algorithm on the white board, probably still can't but, uh, but..When it came to debugging and at that time, like you know there weren't a lot of kind of user space debuggers like, you know, us two, which again blast from the past for anyone, that shout out to IBM's OS2 Operating System you know pre Windows3.
Phil Estes (07:28):
Um, you know, we had a kernel debugger and like, the fact that I knew all of our registers and how the program counter worked because, that's what I focused on in college. Um, I ended up being the go-to guy for debugging like, I loved debugging problems and, so yeah, I think in some sense maybe that-that was the relationship, you know, why that came easy to me but, yeah.
POP (07:52):
Got it. And so, let's.. We're at IBM, right so like talk-talk to me about containers, like how did you to that next step there? I'm sure there was a little step in between there, maybe we need to talk about that, but..
Phil Estes (08:07):
Yeah, I mean, I think the step right before that's interesting, just because containers came via Linux. Um, I got really excited Linux as a user, um..
POP (08:19):
Which was your first distro?
Phil Estes (08:20):
Uhhh...I am pretty sure I downloaded Slackware first.
POP (08:27):
Mm-hmm (affirmative)
Phil Estes (08:27):
On diskette, like 12 diskettes if I remember correctly.
POP (08:31):
Wow!
Phil Estes (08:32):
(laughs) And so yeah, I had a-a PC that IBM had given me for home use, and I was actually doing a Master's Degree while I worked at the time, and we had all these programming assignments and it was like, oh, do you want to go buy some compiler you know, install windows, and its like, oh, Linux, I can install Linux, I got GCC. I can do the programming assignments. So, that yeah, that was my entry path. It was like, oh, this is awesome. So then I installed Linux at work and I was one of a handful of people using Linux and it was like my PC at IBM and so the-the...
POP (09:08):
Can I ask you a question on that?
Phil Estes (09:09):
Yeah.
POP (09:09):
I'm sorry to interrupt you.
Phil Estes (09:10):
Yeah.
POP (09:10):
Do you think the electrical engineering kind of hardware background made you much more inclined to be integrated with like, Linux, because you had to, like compile things, to get drivers to work and all of that fun stuff. Do you think that's probably the reason why. Is that?
Phil Estes (09:25):
There may have been a sense that it was attractive to me for those reasons. I mean, I, the funny thing is like, I-I don't know how often these two things go together today, but like, when I downloaded Slackware, I joined the Linux Kernel mailing list (laughs) and like read, like read daily, like, all the traffic and also, you know, found you know, at that time it- it was pretty usual to hit some problem like oh well, this PC has this card and oh, there's a patch out here, and I need to grab that and re-compile this driver. And so, yeah, my familiarity with those things made it, you know, it didn't seem like a big deal to do that... yeah..
POP (10:05):
Got it. Got it. And so..
Phil Estes (10:07):
So.
POP (10:07):
Go ahead, yup go ahead.
Phil Estes (10:08):
Oh no, so I was just going to say, you know, Linux use in IBM was growing and around this 99-2000, IBM said, hey we're going to go all in on Linux we'll enable it for all our platforms, and so we created this Linux technology center, and although I didn't join that team initially, uh, people there knew I was a Linux aficionado, and we had a JVM IBM had become a son licensee we were producing it for OS2 and Windows, WebSheres coming out DB2 for, um, with the job of drivers coming out, and we were like, hey, we need a JVM for Linux and so I ended up porting a job into Linux because again, I had the interest and I had, I was on the Java team at the time so, and that was kind of my gateway into like, making Linux part of my job and from there on I worked in the LTC for about a decade.
POP (11:06):
Very cool. Then the lightening struck and containers showed up. Talk to me about that.
Phil Estes (11:13):
Yeah.
POP (11:14):
Or am I skipping a step, I mean.
Phil Estes (11:15):
No, no, no. I mean I yeah, so I did a few other things in there, but by that time Linux was, kind of, my main skill set. IBM was working with distro partners, working with Red Hat and SUSA and then Canonical, and of course containers, were you know, born out of a bunch of these Linux kernel features, so when you know, Docker exploded, um, I was working on some other things for the cloud unit.. but..
POP (11:41):
Can we talk about that real quick?
Phil Estes (11:43):
Yeah.
POP (11:43):
I'm sorry to interrupt you again.
Phil Estes (11:44):
Yeah.
POP (11:44):
I think.. and I'm going to say this out loud and I don't care 'cause it's my show.
Phil Estes (11:49):
(laughs)
POP (11:49):
Solomon does not get the respect he deserves.
Phil Estes (11:51):
Yeah.
POP (11:51):
Solomon without a doubt, what Docker did it terms of enabling even the-even the Kubernetes folks that are talking about distributions in terms of, you know, run time capability if that didn't, they didn't have that they weren't going to have the, let me container it for you externally for folks to be able to like, absorb, and use Kubernetes so props to you Solomon. Hope to have you on the show someday.
Phil Estes (12:13):
Yeah. Awesome. Yeah, I-I agree, so, yeah, uh, a friend and I were told hey, this community is exploding like we want some people invested, you know, go get involved in the open source project. And, it's funny, because I'd done Linux-y stuff for over a decade by this time but I-I can't say I would've called myself an open source contributor. I mean, We were doing a lot of packaging. We were working with the distros and I'd open some bugs and I think I'd send maybe one patch to a mailing list at this point, um, but diving into the Docker open source project, especially at this time, you know, late 2014-2015, it was like stepping into a new world. I mean I-I remember bumping into Kelsey Hightower on an IRC channel about Docker and he was poking around looking for someone if they had looked into enabling IPV6 and the Docker engine and then on that shows up for Microsoft, and hey I want to put this thing into Windows and now he's at Google and he's written some of the Kube cuddle plug in features.
Phil Estes (13:19):
So, you know all these folks like back then, Jessie Frazelle was there and Michael Crosby, and uh, yeah, I mean that next year, I would say is probably, the most fun year of my career, like, working with this amazing set of folks on a technology that was, like, everyone was using it and everyone was excited about it and every vendor in the world wanted to have their hands on it, and hands in it, and again, that brought challenges that brought, you know, some slug it out, slug fests and pull requests and issues, uh, but, just an amazing time, where I feel like...It's -it's weird to say this, but after 20 years in software, like, I grew so much in that year of, you know, really learning from amazing people and having Mic go public for the first time there...
Phil Estes (14:11):
It's amazing to think, you written all this amazing code in your life, but like, you're sweating thinking about opening this PR, like, what's Solomon going to say or what's Michael going to say or Jess or, you know, whoever comes across this thing...So..Yeah, it was an awesome time.
POP (14:27):
That's great. It's awesome. And again, it's like, you know, all great stuff that you contributed and I want to talk about obviously the obvious one and it's containerd. Right. So let's talk to me about. What is containerd? So, for the uninitiated.
Phil Estes (14:40):
Yeah, so you know, it's hard to talk about containerd without talking about what Docker is, especially, if you're an early Docker user. You love the fact that this thing was a Swiss Army Knife, all-in-one tool. Like you had a registry client, you may have not thought about that, but when you did Docker push or pull there was a registry client that talked to Docker Hub or whatever registry. There was a build tool. You know, you just typed Docker build, but a whole separate part of the engine, you know, uh, processed your Docker file and did those build steps, um, and then obviously as Docker grew, new things came out. Docker Inc, you know, wanted to create an enterprise product, which they did and Swarm showed up, and-and networking plug-ins and storage plug-ins. Underneath all that, and well in the early days, right in with all that, was the actual code that talked to the operating system and ran your container, and when we formed the open container initiative, so the OCI, um, to kind of standardized that layer, uh, most of you have heard of RunC, you know, this binder that actually does that, it was, a subdirectory of code in Docker called lib container and, you know, that actually had the interactions with the operating system to actually start them you know, start your process within these name spaces and controlled by these C groups.
Phil Estes (16:11):
So, RunC was-was way down there at the operating system. Docker had all these growing, you know, list of features and capabilities. Containerd was created to sit in between those two layers, and so really, you know, runC if you think of it's just, kind of, a transient process, like it starts up your container and then it's done. So containerd could be that demon that sits above that handled the life cycle. When you say pause my container, or kill it, or start it, containerd would manage, wo-working with runc beneath it to actually do those steps for you.
Phil Estes (16:46):
Now, if you fast forward to, you know, 2017, actually late 2016, you know, there was growing contention, especially, you know, people writing these Kubnernetes vs Docker blog posts, um, and one-one of the big calls during that era, you know, you can go back and find blog posts and tweets about this. There's like, can we just have a stable run time that allows Docker to do what they want to do, you know build enterprise products, build Swarm, uh, Kubernetes could be allowed to do what they want to do, expand their scope and their feature set, but all built on this stable run time, and so containerd had already come to life Michael Crosby and others had kind of, already written that core code, but at that time, uh, you could almost think we expanded the scope to say, okay, containerd is going to be this full core run time that Docker can use, that Kubernetes can use. It'll have that registry interaction piece. It'll be able to build containers that can be build kit or somebody else's tool. Um, and so it kinda had this narrow scope but it was enough of a container run time that you could use it isolated. You could use it under Docker. You could use it under Kubernetes, um, and with a nice clean API, that means you could build another tool on top of it, so you know, Alex[inaudible 00:18:12] building FaaSd or Darren Shepard building K3S.
Phil Estes (18:17):
You know, containerd could be this embedable run time in some other project that you want to build and so that's really where we ended up and then of course, this idea of being stable and vendor neutral, it was donated to this CNCF, um, and you know I think if anything, um, you know, I'm proud of the fact that we-we a lot of us helped kind of manage a lot of those tensions and relationships to get to that point, where Solomon was happy to donate and the Kubernetes community was happy to use it, and, um, you know, we dealt I-I hope diffused a lot of those, you know, mid-2016 tensions that existed around the communities... so...
POP (18:59):
You were like the Ombudsman of...
Phil Estes (19:01):
(laughs)
POP (19:01):
.. Containers pretty much. That's great man. Um, so uh, the other thing is like. I've so experienced this and I think I tagged you in a post about this but like basically and so no secret so I work at [inaudible 00:19:16] We do telemetry for IBM cloud and all that fun stuff and there was a thing where we're like, hey, we're moving over to you know containerd as the sole run time on Kubernetes at the point of IKS and other folks were asking for it and I was so proud that they, when we actually had our agent work on containerd I actually tagged him and like look at this we did it you know and something like this and then the customer moved some of their workload over to it and the performance for containerd,it was incredible like it was incredible almost a 30-40% increase in their, um, their performance and resource utilization was less and again man, props to you, that's really good stuff.
Phil Estes (19:59):
Yeah, I mean it was awesome to be able to have that narrow scope of the entire kind of Docker ecosystem, because it really allowed us to focus on just a really clean API to do those simple life cycle tasks of a container which takes away a lot of, you know, the-the run time overhead of-of managing a bunch of extra metadata and information about your container, um, so, you know, I wish I could say I was the author of all that greatness.
Phil Estes (20:28):
I mean, there-there are a handful of containerd folks that really made it what it is today. But yeah, we're-we're super excited for all the people that have used it, been really happy with it and like you said I you know there's some blog posts out there I think Lantow, from Google, you know he was a big part of the CRI plug in, which is how Kubernetes actually uses containerd..
POP (20:54):
Right.
Phil Estes (20:54):
And, so there's a blog post on Kubernetes blog when we announced containerd plus the CRI going GA that talks the exact points you made, like, how much less memory does it use, what's the performance characteristics, and so, you know people can go find that information on I-I think, um, you know there's been no monkeying with the numbers or playing games that's the, that's the end result of really being able to write a minimal core run time that doesn't have to have a lot of other baggage around it.
POP (21:28):
And again, you can kick the tires on it on a live cluster with IKS. You can you it in GKE as well and I think you can do it with others, and even it you spin your own if you look at the Kubernetes documentation, it's one of the run time, uh, choices...
Phil Estes (21:40):
Yeah.
POP (21:40):
there
Phil Estes (21:40):
Right.
POP (21:41):
Definitely good. So let's talk a little bit about, again, you're the you know the OCI chairman of the talk right, in terms of technical oversight committee. Talk to me about what the OCI is and what you do in the talk.
Phil Estes (21:53):
Yeah, so, um, yeah, the funny thing is, every foundation picks a slightly different name so talk is-is correct in-in general but OCI chose TOB-Technical Oversight Board, but committee...
POP (22:07):
Apologies. Apologies, you're all takin' care of business. So you're TCB'n . See what I did there?
Phil Estes (22:12):
Yeah. I like that. So, anyway, in case anyone is searching, and hey, where is that OCI talk, well it's the TOB.
POP (22:18):
TOB!
Phil Estes (22:19):
All the same thing.
POP (22:20):
TOB! Yup, yup.
Phil Estes (22:21):
Um, so yeah, you know when I gave my hopefully, not too long-winded history of things, you that-that layer of standardization that we try to add to diffuse some of the Rocker Docker, etc. complexities of that early container era, you know, one of the things we wanted to standardize was this idea of like, when I run a container, let's all agree on what that means. What are the things we want to specify, the name spaces which command to run what the environment variables are.
Phil Estes (22:57):
And so you can think of those CI really, kind of, covering three big topics: Run time, like how do I define what a container is? Image: How do I actually package all that stuff, its configuration and the layers into an image and then the latest thing that's an expansion of the OCI idea is distribution, which is that API how do I talk to a registry? So, you put all three pieces together and now that the distribution spec is release candidate for 1-0, hopefully in a few months, you'll be able to say, hey, we've standardized this entire path of how I run a container, how I package it as an image and how everyone talks to registries, and so, that's awesome for the vendor communities that's awesome for users that, you know, a lot of other technologies over the years have gotten stuck in this model, of well, I've used tool x to make something and now tool y won't run it and OCI again, imperfectly, I'm sure, in some ways, but, has really brought standardization I'm sure in some ways, but is really cool that you can use red hat tools to build containers, you can containerd to run it.
Phil Estes (24:12):
You can use Harbor registry, you can use Docker Hub, you can use Qway and you're going to get the same experience as we standardize those pieces, so yeah, it's pretty awesome.
POP (24:23):
Yeah, yet again, you're the Ombudsman. Right. It's hard though, like, it's all these folks that have obviously, you know, vendors have them own somewhat agendas. They have things they want to push through. Open source projects have things they want to push through and being able to corral that, I just give props to you all. That's really a tough thing to do and now you're doing it twice, man, that's crazy.
Phil Estes (24:47):
Yeah, so it's been another great group of people that, you know, have crossed all the major players that everyone would recognize, um, and it's also cool that the OCI like started with this run time community because that's what was important then and now a lot of those people have kind of filtered off to other things.
Phil Estes (25:07):
And now, four years later you know there's a different set of ... all the registry teams are here. We've got Azure and Microsoft and IBM and Google and Harbor and all those folks show up in those meetings today because we're standardizing distribution, so it's kinda cool that even though that group of people has kind of migrated over the years, um, we've had great collaboration across the clouds, and across the vendor space, so yeah, it's been cool.
POP (25:34):
Nice, man. Nice. So, I wanna shift gears a little bit. Alright. Let's talk, you know in terms of what do you do in downtime, 'cause you're not workin' on containers and container spec and OCI all the time. I hope that, look, you know you have got that fun couch behind you....
Phil Estes (25:47):
Yeah.
POP (25:47):
You know we've got a bicycle shirt. Let's talk about biking first. So, you know, Liz Rice was on, she talked about that and did a lot of that. Talk to me about that, your passion and about biking.
Phil Estes (25:58):
Yeah, um, so it's one of those things that has come later in life, you know, hopefully, you know, a lot of people learned to ride a bike, so it wasn't that I need to learn to ride a bike but, the, I got to that age where, like, if I didn't do something, I was going to get way bigger than I wanted to get, (laughs) so cycling just became this thing that intrigued me.
Phil Estes (26:21):
You know, almost anywhere you live in the world you can find a local cycling group or club that does road rides and of course, you know, if you like mountain biking there's mountain biking clubs, so I've gotten into road cycling, probably-probably seriously into it about six, seven years ago. Had some friends that really wanted to get into it. We started doing longer, and longer rides, did some centuries, did some sprint triathlons, um, but it, yeah, it's been my way to be, active and to, you know, I think almost any technologist you talk to, will say, yeah when I went and rode my bike or ran or whatever my thing is that's when I figured out this problem I was, you know, kept fighting with so it's my way of clearing my head and ...
POP (27:10):
It's just your balance.... pretty much, this is your.. that that is the balance aspect that helps you kind of th-th-think through things and again.... talk to me can you give me an example, give me an example, man, where, like, you took a ride were like, holy shit, I figured something out.
Phil Estes (27:29):
(laughing) Yeah, I mean, I'm trying to think of a recent one where I can give real practical specifics, but I know for certain, there have been.. well I've done a lot of conference speaking and sometimes I get stuck thinking, what's the flow I want for this talk and sometimes just going out on a ride, you know it all just comes to my head and like oh, I'll go down this path and explain this thing. Um, and the, in the past where I was doing more coding, I know for a fact there've been times on rides it's like, that's the thing because I forgot about checking this. So, you know it's usually that out-of-the-box, you know, thing. I couldn't think when I was just staring at the screen but you know getting out helped, kind of, free my mind to solve it in some other way.
POP (28:21):
That's cool. That's cool. Now, let's talk about the other passions. There's two passions that you told me about before, but I kinda wanna, I wanna talk about them because they're pretty you know... we've got drones, right. Let me talk, talk to me about drones. Drones is a recent passion of yours. Is there a specific one that, like got you involved, 'cause like I have friends that like are hard-core with it right and ... you know..
Phil Estes (28:42):
Yeah, so you know. I've always been interested. I've actually been interested in flight for a long time. I've had friends that were private pilots and, you know, obviously in this job I've gotten to travel the world which is not necessarily the fun kind of flying, but it's- it's fun to go somewhere in the world, uh, but um, yeah, so couple friends got seriously into drones and we were producing these amazing videos over the last you know, four or five years um, and you know, if-if you look at the prices points you can either buy a piece of junk that doesn't do hardly anything or you can spend thousands of dollars and do amazing things and I was like I don't know if I want to pull the trigger on that, but DGI came out with a Maverick Mini in the last couple of years. So finally this year, I was like that's reasonable, even if I don't do much with it, it not like I spent several thousand dollars. So...
POP (29:39):
Right.
Phil Estes (29:40):
Um, but anyway, I've always loved taking photos when I travel and you know kind of seeing what I can capture, how I can frame things. I wouldn't claim to be anything like a great photographer, but it's a great fun side passion, um, and so I've just had a blast you know just being by a river side or finding a place to craft like a flight and okay I can take this video and then come across here so It's just another creative outlet to create something cool, um...so
POP (30:18):
Do you think DGI somewhere is using some containerd. Wouldn't that be like, awesome like stuff going on. I mean it, wouldn't it be good, some converged stuff.
Phil Estes (30:27):
Yeah, let's hope so, you know you fly off an app on your phone which is pretty cool in general and you know there's a back end and these days you know there's an 80% chance that it's somewhere close to Kubernetes and containers so...
POP (30:42):
Come on DJI, throw my boy Phil, like, an endorsement here, come on (laughs) let's go
Phil Estes (30:45):
Right, right.
POP (30:48):
So, uh, let me ask you about this, because again, everybody during the pandemic has been probably or is now a 4-star chef. So talk to me about, like, I guess you're cooking a lot more for the family. What's on the menu man, because you're making me hungry already.
Phil Estes (31:01):
Ohh, yeah, no, I -I have almost always loved to cook, which you know it's funny because it does seem like there's a pretty good cross section of tech folks that are, you know, Paul Zarkowski, I don't know if anybody, you know, surely some people out there know him. He's like master level posting stuff on Instagram, these amazing meals. I'm not quite at that level, but um, yeah, I've always, you know, even growing up, I've always even growing up, I always helped my mom and cook a lot and my grandma was this amazing cook uh, Upstate New York, kind of old world cooking, the big roast on Sundays and ...
POP (31:40):
Let's talk about that. I'm from Update New York too. What part of Upstate New York are you from? Well, I'm from [inaudible 00:31:46]
Phil Estes (31:45):
Well, my mom's family is all from the Binghamton area. Endicott, Binghamton
POP (31:49):
Oh. I'm from Oneonta.
Phil Estes (31:50):
Oh yeah..yeah
POP (31:51):
Yeah.
Phil Estes (31:52):
Yeah not too far. So, but yeah, so you know, food has always been a big part of that cultural background, you know, I love to try to cook Indian food, Thai food, you know, a bunch of staples of course, but yeah, it's fun, it's a fun thing, just like any other creative, uh, outlet just to like figure out how sauces come together. Like why does that come together, into a perfect sauce and how can I do better next time and so, yeah, you know again, some of the artistic stuff, presentation, I like the family to ooh and ahh and wow that looks amazing.
POP (32:35):
So you do the micro-greens and you know do the, like, baby carrots and all that? That's funny. Are you like would you say you go through phases where sometimes you go do French cuisine and sometimes you do like or it's mostly like, here's my staples and if I want to really wow it up I do something else?
Phil Estes (32:53):
Yeah, I mean, you know, with kids and families there's a bunch of staples everybody wants to have regularly whether it's Italian stuff or stuff we've found over the years. But, it, the funny thing is my wife and I are a great team, because she's the one looking out for recipes and finding new stuff and then she's like, hey, why don't you do..I got this recipe, why don't you do dinner tonight. So, it's kind of tag team approach, you know, we both love to cook.
Phil Estes (33:20):
We used to, before we had kids, we used to get together with this other couple and spend like all day Saturdays, like, cooking meals, because everybody worked. We were, all of us were doing full-time jobs and so we'd freeze up, like, meals for a month and we'd spend the whole day cooking, and so, yeah, we've been doing it a long time and it's a fun outlet too.
POP (33:41):
That's fantastic. Hopefully when this gets done, I'm going to invite myself over to your place and I wanna have a-a fantastic meal man.
Phil Estes (33:47):
Yeah, awesome.
POP (33:47):
(laughs) So, um my last question for you. What work are you most proud of in your career?
Phil Estes (33:54):
Yeah, so, you know, I think I've already tipped my hat a little bit you know, those first couple of years working on the Docker engine and working in that community of-of superstars really, um, you know, that really pushed me forward, I think in a lot of ways to do some things that-that, you know, were a little out of my comfort zone.
Phil Estes (34:15):
So, early on I remember sitting with Michael Crosby, like, hey, I can grab issues off the issue tracker all day but, is there like, a bigger chunk of something that like, no ones attacked that everyone's asked for in Docker? And so, um, he's like user name spaces, people have always, you know, couple people have tried to throw together an implementation, but it'd be awesome, you know, if you wanted to take that, um, and, so, you know, long story short, I ended up writing an implementation that ended up getting merged and is still there today in the engine.
Phil Estes (34:52):
And, I think there's a lot of cool stuff that's been built since then, like rootless containers, rootless Podman, Rootless Docker, Rootless Kubernetes, and [inaudible 00:35:04], Sudad, and a bunch of other folks from red hat, have, kind of, taken, not even taken what I did, because it really, Rootless is coming at it from a different angle. Uh, anyway the user name support initially in Docker was an awesome achievement.
Phil Estes (35:22):
I mean, I'd never received the kind of accolades I received. I'd never wrote a blog post that had so many views as my first blog post about user name spaces. So, it was definitely a pinnacle for me to feel like, oh wow, people wanted and I was able to pull it off and so, yeah, it's awesome to have that kind of experience especially in open source and-and getting kind of recognized for achieving, you know, that feature, so, it's awesome, yeah.
POP (35:57):
Great stuff man again. So, we did it, we went from containerd to cuisine to drones, I mean, we did it all today.
Phil Estes (36:04):
Yeah, yeah.
POP (36:06):
I want to thank you so much for being on the Popcast, man. Again, you know I'm a big fan of yours and I hope you all enjoy this, but thank you so much.
Phil Estes (36:13):
Yeah, thanks so much for having me. It was awesome to be here.
POP (36:21):
(music)