Fedora as a Server? Would You Do It? | Fedora Podcast 063
Recorded October 20, 2026

Eric Hendricks, host
Noah Chelliah, host

----------------------------------------

[00:00:00] Noah Chelliah: Welcome to the Fedora podcast. Plenty of system administrators would tell you that Fedora is a great choice for the desktop, a great choice to build professional development, and a poor choice for a server. Is that still true? Or is that the reputation from a few years ago that Fedora just hasn't outgrown? Today, we're going to put that question to the test. Welcome to the Fedora podcast. This is episode number 63. This is the week where it's just the two of us, Eric the IT guy and myself. On today's episode, we're going to ask the question that likely splits the viewers right down the middle. Some system administrators are going to be in the yes camp. Others are going to be in the no camp. And that question is, would you run Fedora on a server? Joining me on the show, as always, it is Eric the IT guy. Welcome in, sir.

[00:01:03] Eric Hendricks: Hey Noah, glad to be with you today.

[00:01:05] Noah Chelliah: Okay. So this is a really timely question because you just got done rebuilding the Hendricks data center inside of the Hendricks data center. Isn't that accurate?

[00:01:17] Eric Hendricks: Yeah. So for those of you that haven't heard my ridiculous ramblings, I've got a Dell PowerEdge R720 that lives in a co-located data center in the great city of Kansas City. And it's running Proxmox. And over the summer, I spent hours and hours rebuilding and consolidating and re-rebuilding as those projects go in my home lab. And so now sitting on top of Proxmox, I've got about a dozen or so servers whose operating system I'll keep close to the chest for a second. But it's all unified. It's all automated. Nightly rebuilds. It's really, really great. So this conversation was suggested by our amazing audience in the Matrix room. And so Noah and I want to kind of dive in and kind of pick each other's brains on whether or not this makes sense or not.

[00:02:04] Noah Chelliah: Yeah, 100%. So let's start with this, Eric. When you go to look for a server platform, are you thinking about it in terms of what's right for the home lab, what's right for business, and do you draw a distinction between the two?

[00:02:17] Eric Hendricks: So I spent a lot of my early career as a systems administrator. So I developed that instinct of, I want something that is predictable. I want something that is reliable, something that I'm not going to wake up one day and it's just broken. Now, take that very, very same mindset and instincts and everything and throw them out the window for my home lab. I want the latest and greatest. I want to be trying out the new features. I want to know, in my sysadmin days, I want to know what was coming. I want to be ready so when that feature dropped in the enterprise version that I could say, hey, I've got this thing in my home lab, it's working great. You know, containers are a great example. I started running containers at home long before companies ever decided that containers were a good idea. And so having had some of that experience, I could go to my boss and say, hey, IT operations manager, this is a great idea, we should use containers. So it depends on the use case. Are we talking home lab? Are we talking enterprise? Because my answer is going to vary pretty greatly.

[00:03:20] Noah Chelliah: So from that, I will choose to interpret that you do draw a distinction between what you're running in your home lab and what you'd run in production for a client or for somebody who asked you inside of a business someplace where we'll just say we'll put the line at where money's on the line.

[00:03:36] Eric Hendricks: Yeah, I would. I tried to come up with something rhetorical or pithy to come back to that, but I've got nothing. But yeah, really, if money's on the line, and that's something that's actually shifted this year with my HomeLab rebuild, was originally my HomeLab was built on top of Red Hat Enterprise Linux. I was on the latest major version at all times, partly because I worked at Red Hat and I wanted to have the ability to demo that to customers or use it to record videos, that kind of thing. But partly because having crazy family life and busy at work and that kind of thing, I wanted my home lab to be stable. I didn't want to play sysadmin. So it's kind of funny that we're having this conversation today because the summer that changed.

[00:04:23] Noah Chelliah: I said it was family. Okay. So from my point of view, Eric, I draw a couple of distinctions. The first is, I think it is an axiom or a self-evident truth that everybody has a development environment smart people separate that a development environment from their production environment okay so that is to say i personally draw a distinction between the kind of home lab for the stuff that runs the media server that my family's relying on the thing that runs the photo server that my wife is using to back up her photos there's a distinction in my mind's eye between that home lab and the oh it's shiny what's coming down the pipe and then And, you know, that's a separate home lab, right? So that's the first thing.

[00:05:04] Eric Hendricks: So what you're saying is in your home lab, you even have production and development environments.

[00:05:09] Noah Chelliah: Yeah, I do. And I'd be lying to you if I said I always had a clear delineation between those two. But here's the way I alleviate the guilt of that, right? So I very much have a virtual host that its function in life is to run the quote-unquote production stuff that's in my house. And then I have all manner of different concocted machines that are doing other stuff but the way that this stuff ventures its way into my house is very much not I sit down at a production thing and I start tweaking it or monkeying with it and then it turns into this development mess. That's really not the way it goes in my life. In my life, the way that these things come about is that I start with something in a development environment and it's very much in the development sandbox and it's never intended to be production and it is purely just a toy. And oh my gosh, this toy is so cool. It works so well. I can't believe how well this toy works. I can't believe all the things I can use this toy for. Look at how great this thing is going. It's doing so much. Wait a minute. I have just crossed the line and now I'm relying on this thing that was started over here in development world and has slowly made its way into like, I can't imagine living without it. Right. And that has been the story of volume. That's been the story of home assistant. That's been the, there's been a whole bunch of things that have worked their way into my life from that direction. When it comes to the development environment or when it comes to playing around, I don't have enough good things to say about the fedoras of the world, the arches of the world, the things that if they're not bleeding edge, leading edge type software, because to your earlier point, I want to see what was what's coming down the down the down the stream. And that's even more relevant now today with the way that Red Hat and CentOS have kind of repositioned themselves, whereas before, I would say Fedora fit that bill really, really well. Today, I would say if you're trying to scratch that itch, that professional development, I want to see what's coming down the pipe sort of a thing. That really is, in my mind's eye, is parked more in the CentOS world. And today, Fedora is a completely different animal. And so whereas before you might have said, oh, it was so leading edge and it was, you know, people incorrectly, I would say, use this concept of, oh, it's the beta thing that is being tested before production. These days, I would say there are plenty of things that run in production that run on Fedora to include some very popular POS appliances that are built. And we talk about money. I know of one particular organization that in any given weekend, they'll clear about $100,000 in 48 hours. And the point of sale service, the point of sale machines that they use are running Fedora underneath the hood. So to some degree, I think that is a reputation that Fedora has kind of drug with it into the cruft of 2026, but isn't really a fair summation of where Fedora stands today.

[00:08:05] Eric Hendricks: So, that line has slowly slid backwards for me. So, way back in the 20-teens, I was running CentOS Linux, which was a downstream of RHEL. When I started working for Red Hat, I had this beautiful employee subscription where I got unlimited everything. It was great. I miss it. But if, to your point, you're wanting to do some kind of business development, or if you've got a small business, disclaimer, I don't work for Red Hat and they're not paying me to say this, but they do have developer subscriptions where you can get up to 16 instances of RHEL for free. You renew it once a year. It's one email. It's really easy. I still have my developer subscription, 16 instead of infinity. There's that. But I mentioned that my home lab used to run. Let's be honest.

[00:08:53] Noah Chelliah: You need another email if you need another 16. It's not like anybody from Red Hat is in terror. Let's be honest. I mean, it's not like anybody from Red Hat is showing up and doing audits like the, you know, what's in the world somewhere in Redmond. It's a real thing, right? They come and look and look for license compliance. It's a bad day if they show up.

[00:09:12] Eric Hendricks: Well, I won't promote loopholes or anything because, to your point, CentOS Stream is now repositioned. And we talked to the CentOS group a few episodes ago and talked about how CentOS Stream fits into the development lifecycle. So if you don't want to deal with Subscription Manager, if you don't want to deal with signing up for subscriptions to RHEL, like Noah was alluding to, you can use CentOS Stream. And it's incredibly stable. I know of very, very high-level organizations, very leading-edge organizations, companies that you would have heard of that use CentOS Stream as their server for their for-profit business. And so, actually, when I was thinking about rebuilding my home lab, I was like, I don't want to deal with Subscription Manager. I don't want to sign up and go through Red Hat's process, also run CentOS Stream. And it's really kind of funny because, The Fedora podcast community actually challenged me in July. Well, you're the Fedora podcast host. Why not rebuild on Fedora? I was like, that'll never work. It changes too fast. It's too bleeding edge. I was in July. We're recording the last week of October. And I couldn't be happier with my Fedora server environment.

[00:10:27] Noah Chelliah: Rule number one, the audience is always smarter than you are.

[00:10:30] Eric Hendricks: Yes, I admit that on air on the regular.

[00:10:37] Noah Chelliah: So if we take that kind of to the extreme, or if we dig into that a little bit, so you rebuilt your home, Are you kind of letting the cat out of the bag? Did it eventually wind up in a Fedora way?

[00:10:50] Eric Hendricks: I am, in fact, running Fedora Server 44, which at the time of recording is the latest and greatest. But there's an added twist. At the same time, people were pushing me to try out Fedora Server for my home lab. I was also drug pushed and...

[00:11:09] Noah Chelliah: Encouraged.

[00:11:10] Eric Hendricks: Sure.

[00:11:12] Noah Chelliah: Persuaded.

[00:11:14] Eric Hendricks: Forced into trying out fedora bootsea and noah it is a glorious life on this side of the boot see hurdle i am loving my home lab i spent about four weeks working a warehouse, right in august in kansas city which can be sweltering hot and needless to say when i got done at the warehouse i wanted to do nothing but vegetate in front of my pluck server and so for almost four weeks after this rebuild was done, I didn't touch my home lab, which a year ago would have meant there's no security updates. There's probably some logs that had filled up a disk somewhere. None of that happened for four weeks because I was using Forgeo, which is the same SCM and CICD pipeline building tools that Fedora proper uses. So I'm using Forgeo or Forgeo or different people say it different ways, but I'm actually running a Forge server here in my home lab with a couple of runners that do all the building. And with bootc, it's basically turning a container file into a server image or a desktop image, whichever your use case. I use them to build server images. So I've got a dozen Fedora server, latest and greatest bootc-based images. And guess what? They get updates every single night. At about 2 in the morning, my Forge server goes out to Fedora's registry. Pulls down the latest image and then around three in the morning another job kicks off that rebuilds my base image so if if you think about a container file, you start building layers and so, the first layer of my image is just the standard fedora server image that gets rebuilt anytime there's an update so once a night my server goes out and pulls that down that becomes layer one layer two i add in my user accounts layer three i add in ssh keys layer four and so forth to get, everything that i want in every one of my server images and then i mentioned i've got about a dozen different virtual machines, running everything from minecraft to plex to my ford server itself and they build each night if there's an update they'll they'll look at they'll look at the previous step and go oh look there's an update and so they pull and if and and just just like our mobile devices they pull the full image and then, to to update all you do is all you do is reboot, and you reboot it to the new image. It works so well. Almost every single night, most if not all of my servers pull down some sort of update, whether that's a container update or base image update, and they reboot around four in the morning. So by the time I come back and sit down at my desk, everything's still up and running. But I still, within about a 24-hour period, I've got the latest and greatest that Fedora Boot C server has to offer.

[00:14:09] Noah Chelliah: Now, would you say the magic sauce in there is fedora, bootsy, or the combination of both?

[00:14:14] Eric Hendricks: Fedora, bootc, and having automated CI pie points.

[00:14:18] Noah Chelliah: Okay.

[00:14:20] Eric Hendricks: Because otherwise, what are you doing? You have to log into your system. You have to run DNF update. And then you have to check and see, does this impact something like GLIPC, something like the kernel? And then you have to manually decide that, yes, it pulled down a new kernel. I should reboot so that my system's actually running the latest kernel. With all this, that is automated. I don't have to decide whether or not I need to reboot because there's logic built into that, into bootc itself, coupled with my CI pipelines that say the base image or some kind of configuration changed, reboot the system because everything is baked into that image. So it really eliminates a lot of decision-making process. It eliminates my need to go in once or twice a week and run DNF update. I mean, a year ago, I would have been having, I would have had Tmux session with like 12 panes and go into all of my servers and do a DNF update on all of them and then one by one check to make sure, okay, updates are done. Do I need to reboot? No. Okay, we're good to go. Now do the same process with, if I'm running a dozen servers, 30-some-odd containers maybe?

[00:15:27] Noah Chelliah: I just have this visual in my head of copy, pasta, pasta, pasta, pasta, pasta, pasta, pasta, pasta. Yes, yes, yes, yes, yes, yes, yes, yes, yes, yes, yes. Reboot, reboot, reboot, reboot, reboot.

[00:15:39] Eric Hendricks: And that's how it was for a long time until I realized that that's how it was for a long time until I realized Tmux can actually synchronize your pain. So if you type DNF once it replicates across all the pains, but it was still still pain in the rear.

[00:15:54] Noah Chelliah: So we divide development and production. We kind of separate those two things, whether we're in our house, whether or not we're in a business, whether or not there's money on the line. Then we separate with what I'll call a critical path. OK, so something that sits inside of a business or an organization or something that would cause them to lose money, which is going to be different. And, you know, and so your house, there may be a critical path, right? There's a spousal approval factor, those sorts of things.

[00:16:21] Eric Hendricks: So the way that I would define production in a home lab is, is a friend or relative going to text me at an inopportune time, which for me was usually when I was halfway across the country at a conference, and say, hey, your XYZ service is down. That's how I define home lab production.

[00:16:37] Noah Chelliah: There you go. So for you, that's a critical path. So I think you start to create these boxes and you start to put things that helps define things. Now, here's something interesting. You use the term in a server we want stability and predictability okay now let me give you two examples of how i will define stability and then you tell me which one resonates with you more okay here's the first one the first one is stability means it doesn't ever change so i think debian And I think free BSD. At the same time, I'm going to tell you, stability means as soon as a fix to a problem is available, it is applied. The developer knows what he or she is doing. So when there's a problem, stability is allowing that developer to fix the problem as quickly as possible and being able to implement that as quickly as possible. Which one of those two resonates with you more when you think stability?

[00:17:39] Eric Hendricks: Mm-hmm.

[00:17:41] Noah Chelliah: It's difficult, isn't it? Because you got two different problems and they can both cause instability. And it isn't a, it isn't a all one or the other. It depends on what the problem is. You get a problem that is because two libraries are incompatible. You have to have the developer come in and change it. But wait, that's confusing because if we let developers constantly monkey, then maybe something else unpredictable breaks. And so what does stability really mean? And so here that, that question and, and your, uh, and thankfully it did confuse you. Well i would ruin my example but that to me is where fedora fits the the bill so uniquely well on one hand you have this stable base you have these developers that are pushing out and they have a product that is intended to be on servers and at the same time they're leveraging the latest technology that exists to include things like bootsee that allow you to take advantage of updates as soon as that they are available without having to worry that you're going to decimate something underneath so it's kind of like you can have your cake and eat it too for the first time in linux history

[00:18:43] Eric Hendricks: Well and back in the, in the early 2000s mid-2000s when i started my professional career in it i think that first version of stability existed there there were mainframes that i i would pull tapes out of there was there was bsd based servers that would run a website and you would just never touch them it was the kind of thing where your boss would sit you down day one and go, we don't look at the server, we don't mention the server, we don't even think about this server because it could fall over. That doesn't exist any longer. Technology and security is moving way, way, way too fast. So to your point, I think we have to default to the second. So as a new vulnerability is released or as AI drops a new PR into our code repository, as one of our developers comes up with a cool new feature, we need to be able to adopt those, especially from a security perspective. But I think we almost need to stop worrying about, quote, stability and the older sense of the term and think more resiliency. One of the things that I found out back in the days when I'd run Arch, by the way, was, sure, something might break. Because that was part of my process. When I was running Arch on the desktop, every morning when I'd log in, I'd be looking at my calendar, checking my email. And at the same time, my desktop would be running all of its updates. With arch i just accepted that there would be days where something would break i remember trying to edit an audio podcast and audacity was broken some something in the update broke the library dependency or whatever it was and audacity didn't work so it's like, this is arch i could go in and troubleshoot it or i can just check back six hours from now and there's probably a fix and sure enough i came back after lunch and was like hey i've got some time let me see if i can fix that audio file, Pac-Man update. Oh, look, Audacity works again. Ta-da!

[00:20:41] Noah Chelliah: So here's where I think the problem is in the example that you just gave, and here's where I think Fedora fixes it. So in your example, you put your faith In the nerds on the internet and unsurprisingly the nerds on the internet developed or delivered right so you went to lunch and by the time you got back some nerd sat down and went wait wait a second my the code push that i just did it broke audacity that can't be i gotta fix this and then crawled in there and figured out what the thing is and before you knew it bob's your uncle and it's fixed so that's the good part right here's the not here's the not so good part okay back to the xkd cd comics some dude in kansas is holding up half the internet right the downside to that is you're standing in front of your boss or your boss's boss or somebody who has some stupid amount of money on the line your answer cannot be well the nerds on the internet are going to fix it i promise just give them a while well who is it well it's you know and then you give them some screen name that you can't even pronounce like that isn't a great answer right and so to and and the fact that it broke in the first place is going to be unacceptable and so we need some way to be able to run the thing to let the nerds on the internet be able to get their code out into the world into production when it is necessary and at the same time we need some way to to be able to have some predictability to understand when this thing is going to change so to me it would seem that the fedora and and the advent of things like bootc allow for both of those things right

[00:22:13] Eric Hendricks: Yeah, I would agree. My environment reboots every night, typically on a newer version of the image. And that's not a scheduled reboot command. That's triggered by there being a new image available. But that doesn't mean that in the middle of the day, I can't decide that I want to add a new service. Maybe my Minecraft server is also going to run a Valheim instance now, which it does, for those that were wondering. And so it's nothing for me to sit down, open up my ForgeO repository, and paste. In a configuration for running Valheim on top of my gaming server. And then I save that, I commit it to the main line in my Git repository, and my CI pipeline comes out and says, oh, we have a change of code here. Okay, let's rebuild the image. Okay, we built the base, we added all these packages, We've got the SSH keys. We've got the Minecraft container. Oh, look, here's something new. Pull the latest and greatest Valheim image, bake that into the image, push it out to the virtual machine, and reboot it. So you expand that out to a banking situation. We found a vulnerability. Somebody can get into all of their neighbor's banking accounts because of some SQL injection or something. They need to be able to make that fix quickly, but you also don't want to limit, your ability to add new functionality unless you also have the ability to introduce fixes as well.

[00:23:45] Noah Chelliah: Yeah, in software development, the most expensive thing about software development is not putting the thing on the planet, right? It's being able to continue to support it after it has arrived on the planet.

[00:23:57] Eric Hendricks: For sure, especially two years after the thing has been introduced to the world and the development team that originally wrote that feature has turned over three times.

[00:24:08] Noah Chelliah: You know, the other thing that I think has changed that is worthy of discussion, Eric, is that the way that we used computers 20 years ago is just different than the way that we use computers today. And so I'll give you I'll give you two back to back examples. The POS system that I referenced earlier, that has been so far as I know, has existed since the late 2000 ish is been around for that long. And nobody has has been or really that i've seen is real concerned about updating them because in the end they're not really connected to the internet there's a little server that does the thing and it goes out to the POS terminals and they're essentially a little touchscreen thing and you add them the the the items up or whatever and when you hit you know pay all it's doing is generating a total they're taking that number and they're going over to you know shift for whatever payment processor they're using and then they're running that and so the pci compliance and the internet and all the updates and stuff that's happening on totally separate deal so nobody really cares about updates and nobody in that case it really fits option a for stability right it works leave it alone don't touch it To if i were to compare and contrast that i was working with a medical organization and they were using a modern version of called an appliance and also running fedora but in this case the very nature of the appliance the very nature of cloud this that or the other means that it has to be connected to the internet and so by nature then everything becomes a huge every security problem becomes an immediate concern that needs attention right this minute in fact it needed attention five minutes ago and so in that case like people are very concerned and it's a part of all manner of different compliance things to make sure that but it's up to date and has the latest patches and all the rest of it. So it's just changed.

[00:26:01] Eric Hendricks: Yeah, it really has. And everything's always connected all the time now. And everything is talking to everything else. I mean, I had this weird experience the other day where I went to a vendor's website and I was trying to buy a t-shirt or something. But because they use the same payment processing system as another site that, I order from on occasion, they were able to take that and populate all of my name, my shipping address, and then was like, just confirm this is your card. I've never ordered from you before. Yeah. And, to say nothing about social media ads where you talk to a friend about something, not even digitally, like face-to-face, like sitting in the same room, and for the next three days, ads about that thing are all you see.

[00:26:47] Noah Chelliah: Mm-hmm. Not to take us down a side path, but working for a multimillion-dollar media company, advertising company, and marketing company, do you want to know what the secret to that is? Because a lot of people will say, oh, there's a microphone and it's listening to my conversations, and that's not how that works. Do you want to know the secret?

[00:27:05] Eric Hendricks: Ooh, do tell, do tell.

[00:27:07] Noah Chelliah: Yeah, the algorithms, this is going to be hard for you to believe, but I promise you, this is how this works. The algorithms are better at predicting what it is you're going to care about and what it is that you're going to be doing. And so everything about you, everything we know about Eric has a bullion associated with it. That bullion is either true or false. And then we vote on those bullions to see if they're more true or more false. So Eric likes fedora that has a lot of votes for true because Eric hosts the fedora podcast, anti, anti, anti, anti, anti. And so the, it uses that information to predict and it's, it's scary. Good at figuring out what you're going to be thinking about or what you're going to do before you even figure it out. So when, when it's like, Hey, I mentioned cat food and all of a sudden I'm getting ads for cat food. It's not that there's a microphone listening to heard you say cat food. It's there's a whole bunch of other things that it was able to correctly predict that you were going to be thinking about cat food and and to see the curtain pulled back and to see how that works behind the scenes is an absolutely fascinating thing but the moral of the story and the takeaway there and the thing that we should be cognizant of is the computers the robots know us better than we know ourselves and that is worthy of significant pause to say you know Not to get on a soapbox and not to drag us too far off topic. But when I learned that, I was like, oh, interesting.

[00:28:31] Eric Hendricks: We have very much slidden off topic here. We were talking about Fedora as a server. But something to note as we realign ourselves was that what you've described, the algorithms and the processes, you just basically explained to the uninitiated how AI works. If you take that algorithm, that decision-making process, and blow it up, just zoom out. It's just quite a bit. Tiny little statistical mathematical equations. All machine learning. That's all it is, is computers have gotten so good at guessing. It doesn't care what color your wall should be. It just puts together an algorithm based on what it knows, what it's read online, what it knows about you, and has just put together the most likely situation of that wall should be baby blue.

[00:29:23] Noah Chelliah: Okay, so if we were to pull it back completely back onto the rails, if Fedora was going to be a good base, if you were going to base a server or something that you really value and you wanted to set it atop the stability and trust that is Fedora, what would that take for you?

[00:29:42] Eric Hendricks: Well, I'm already there. My home lab's been running Fedora server for years. I think, or not years, for months since July. But to not completely cheat around the question that you asked, as this episode is being released, Fedora Linux 45, desktop, server, bootc images, containers, everything is just imminent. I mean, we are just days away from the newest version of Fedora, and I'm really excited about it. But, you know, ask me how I feel about my entire home lab switching over to Fedora Linux 45 overnight all at once. I'm iffy. Maybe I version pin it to 44 for a week, 10 days. Let the first string of bugs and fixes go out before I put my production home lab onto Fedora 45.

[00:30:36] Noah Chelliah: Okay, so now raise the stakes. Now if I told you I want you to go and suggest it to clients to where money is on the line, what would have to change for it to meet that bar for you?

[00:30:50] Eric Hendricks: If I could give them your phone number for technical support instead of mine.

[00:30:55] Noah Chelliah: You know what's funny? I know that you're, I know that you're trying to, to, to make a little bit of a, I know you're trying to make a little bit of a pass, but like the reality is that in all seriousness, Eric, that was going to be my answer is that for me, there is one thing and really only one thing that is required for me to feel comfortable suggesting it to somebody who's going to put money on the line. And that is a support path. Okay. Yes. So you're doing that in somewhat of a jest by saying like as long as I can give him your phone number. But really what you're saying is when I get over my head, I don't want to get sued. I don't want to be the last guy holding the wall. I don't want to be the last guy holding the leash. And so like who do I hand that off to? Where is my path of escalation? And in the proprietary world coming from it the thing that proprietary world does better than anybody else they have a path of escalation now that oftentimes do not confuse that with a path to success because they are very different but the the path of escalation looks like well that's not our problem it's the hardware vendors problem here let me get and and and it's not uncommon to get two of those that just play volleyball back and forth what's a software vendor it's a hardware vendor it's a hardware vendor it's a hardware vendor and you just go back and forth and then Meanwhile, you're just sitting there going, hey, whose fault it is, and it's not mine. And so that is ultimately unhelpful to the end user, but The having a path of escalation, what I have, my experience has taught me that most end users, most clients don't necessarily expect you to have the answers to all their problems. What they do expect you to say is I'm working on it and here's what I'm doing, even if that's I'm working with the software vendor to find a solution to your problem. And so to the extent that you can get long-term support, to the extent that you can get extended a software security patches so that you don't have to make a change. If Eric isn't ready in seven days when the next version of Fedora comes out and you need to hang back for a little bit, so long as you can continue to get security updates and your stuff isn't going to go unpatched, you are likely going to be comfortable to run that for an extra week, a month, or whatever it is, as long as there's security patches coming in. The second that isn't available the second that it's hey we have an x support cycle and at that date it's like a guillotine it comes down and that's the end of that's no more softer support is coming down the pipe that's a problem to the extent that you can't escalate up the chain to say i know that you have a problem i want to help i don't know how to solve this let me get somebody else who can and there isn't somebody on the other other end of that chain that's to me where there's a demarcation between what is responsible to use inside of an environment and what isn't. IT malpractice is a thing. I just think we have to be aware of that.

[00:33:43] Eric Hendricks: I'll pose a question to you here in a second, but here's what you got me thinking about. Fedora Linux releases a new major version every six months, so twice a year, six months. Their support cycle for each major version is nine months. So, if you think about it, you've got about three-ish months of runway. I'm... I actually have this crazy idea for the podcast. Maybe one Saturday afternoon, I will live stream the upgrade of my home lab from 44 to 45 and we'll watch it live. So there's idea number one. We should make this a day.

[00:34:17] Noah Chelliah: Live demos always go great. That's what I always say.

[00:34:20] Eric Hendricks: As someone who does them for a living, no, they never do. But that's part of the fun, right?

[00:34:27] Noah Chelliah: So the guy watching, yes.

[00:34:29] Eric Hendricks: Yeah. Which I found out more times than not that the audience knows better than you do. I don't know the number of into the terminal live demos that I did as part of the RHEL marketing team where someone in the audience knew how to fix a problem that we ran into live on air quicker than we did. And we were the quote experts. So anyway, the tangent for a different day. But three months is plenty of runway for Eric and his fleet of 12 servers where everything is containerized. You know, maybe I should be more concerned about Podman 6 which which is due out anytime now more than i am fedora, server but that that's two different changes around the same time and with all my workloads being containerized maybe Podman is something i should look into more before i flip that switch. But my point being three months is plenty of time i'm thinking a week but three months is plenty of time for eric's home lab which is running Plex and, Audio Bookshelf and Minecraft to make that major shift. But if I'm a bank or worse, a medical organization where I'm looking at things like FIPS and HIPAA and SOX, if I'm doing some kind of government work or PCI DSS, if I'm taking financial transactions, that's a lot of moving parts because then you've got other things like you're doing booking, you're doing medical imaging, you're doing, You're doing communication. You're doing all these different things. And so all of a sudden, the level of complexity is far, far, far more than just 30-some-odd containers that aren't mission-critical. Outside of my home lab production definition of someone will probably call me and say, hey, your Plex server's down. It's like, I know, I'm doing a live stream, upgrading it right here, right now. So my question to you, Noah, was is Fedora's server really the best answer for that? Technically, I think Fedora's server could run the world. I think it could, but there's no path of escalation beyond jump on a matrix server, jump in a room and ask questions. And you look at turnaround times. It would take a large enterprise three months just to realize there's a new version of the thing out anyway, much less manage to move an entire infrastructure of thousands of servers, tens of thousands of servers. I still know organizations that are running CentOS Linux 7 and that retired like five years ago.

[00:37:03] Noah Chelliah: Mm-hmm. Yeah, it's an interesting question. And here's, and I have, I have a bit of a double-edged answer for you. So the first is my, just my honest to goodness, what I would do if, if, if I were being asked by paying client what to do in that. And my answer there would be, go with some place that has a company that you can buy support from. If only because that's the, that's the path that I know nobody got fired for taking that path. Okay here's the double edge side to my answer though do you have any idea how many very large scale production things are running on debian like major major enterprise things that run atop debian and nobody says to them like well when that breaks where's my path of escalate no you for crying out loud matrix room there you're going to an irc channel and you're going to ask You're going to say, hey, who has the longest, grayest beard? I've got one for you. Right. And whoever that person is speaks up and then they come back and then they get the answer. And I'm not kidding. That is very acceptable in some of the largest enterprises in the world. And so. I, I, I say what I would be comfortable with and where I would draw that line. And then I can immediately point to where I have routinely seen that line violated right, left and center. And so maybe that's not a real fair criteria to judge Fedora against. Also, by the way, it in my, and now again, case study one sample size of one so that it's out there. I've seen very few problems that exist only in Fedora. They don't exist in Ubuntu. It doesn't exist on RHEL. It doesn't exist on CentOS. It doesn't exist on SUSE. It doesn't exist in Debian. It's purely a Fedora problem. I can count on one hand the amount of times I've seen that. 99% of the time, it is something to do with some sort of version mismatch or something where two libraries don't work together, something along those lines. And it affects a whole bunch of different things. And then that's a technological problem and not necessarily a Fedora-specific problem. So again, I struggled to lay the entire weight of that burden on the back of Fedora.

[00:39:18] Eric Hendricks: So, I completely agree with your point about violating that path of escalation. Because what's funny is, I distinctly remember having issues with CentOS 6, which, by the way, the company that I was working for in the late 20-teens was running an entire fleet on CentOS 6 and 7. And I can distinctly remember going to forums, doing Google searches to try and find the answer to a particular problem. And having no real major issue that someone out on the internet hadn't seen and already solved. And I'll take that anecdote a step further. There have been times where I've worked at companies that the one I'm thinking of in particular had PCI DSS requirements. So they did a lot of financial transactions on their servers. So they had to government requirement to meet certain criteria, which meant that we were running Oracle Linux or RHEL or one of them. And I distinctly remember opening a high-severity case one day with the vendor. And this ain't my first road yet. I had all the logs. I had version numbers, kernel versions, uptime. The last time that we went over and brushed the dust off the back of the fans, I mean, I had all the information in the ticket. And then you wait. It's like, no, this is, this is Sev One. This is one of those deals where my boss's boss called my boss who called me as I'm sitting in rush hour traffic, trying to get to the office. And I get in, I open the ticket.

[00:40:54] Noah Chelliah: The weight on the back of my chair right now. I'm sitting vertical, you know?

[00:40:59] Eric Hendricks: And so it's like, okay, support vendor, you are not moving fast enough for me. Let me go out to Stack Exchange and share what information I can. And once you get past the typical round of, well, you didn't format your question correctly for Stack Exchange responsibilities, and, you know, okay, let me edit and then respond. Is that better? Well, now that you put the logs in the right section of your message, now I'll look at your problem. And, oh, look, here's the solution right here. We had a problem identified and a solution implemented into dev stage and prod or dev test and stage and schedule for prod before the vendor even came back and said, well, have you checked this, this, and this? And it's like, what am I paying you for? So I think if we've got a... Unfortunately, as a systems administrator, as a pre-sales engineer, or as a technical marketer, I've given this answer far more often than I really wish I had to. But I think the answer to the question of, is Fedora server enough? I think the question is, or the answer is, it depends.

[00:42:09] Noah Chelliah: Mm-hmm. Yep. I'm inclined to agree with that. I think that I, we know what I would really like to do. I think the best way to settle this, there's been a theme throughout this episode, Eric, and it is that the audience is smarter than we are. So maybe the best way to settle this debate is to ask you, would you run Fedora server? And then, this is the important part, why or why not? So drop comments, find us in the matrix room, send it via carrier pigeon. That is a valid RF. No, that is a valid RFC. You look it up. I am not making this up. There is message. There is, it is, it is, it is via avian carrier. You look it up. That's a ratified RFC. So you can, you can send us a message via carrier pigeon. We will find out to accept that. I'm going to know how that works, but like it's there. It's a ratified standard. It must work. Somebody did the test, right? And just let us know why would you do it? And if so, why or why not bonus points? If you work for a place or have some experience where you've actually done this and what the results are, let us know.

[00:43:14] Eric Hendricks: Yeah, love that.

[00:43:17] Noah Chelliah: Thank you very much for joining us. This has been episode 63 of the Fedora podcast. Make sure to check out the show notes. There's a ton of information in there, guys, links, ways to stay connected to the Fedora community. On behalf of Eric and myself, the entire Fedora project and the community, we'll see you here next time.