text
stringlengths
0
2.67k
**Autumn Nash:** Last one... PairSec?
**Justin Garrison:** Is it ParSec?
**Autumn Nash:** Yeah.
**Justin Garrison:** Kubernetes.
**Autumn Nash:** Dang it. It's really hard to find something in Kubernetes that doesn't have the word Kube in it!
**Justin Garrison:** It's true. It's gonna have a K, or it's gonna sound very Greek...
**Autumn Nash:** Skooner.
**Justin Garrison:** Skooner?
**Autumn Nash:** Printer or Kubernetes?
**Justin Garrison:** Skooner.
**Autumn Nash:** This is so much fun...
**Justin Garrison:** I'm just gonna go with printer on that one.
**Autumn Nash:** That's a Kubernetes. I'm so excited, I had two that you couldn't guess...! Do you know how hard it is to find things of Kubernetes that don't have the word Kube in it? It's so hard.
**Justin Garrison:** Okay, I should have known. Okay. Oh, it's a dashboard. Alright... Good times.
**Autumn Nash:** Look, if you're just pretending to not know to make me feel better, I appreciate it, because it gave me joy. I finally found something that you don't know.
**Justin Garrison:** There are plenty of things.
**Autumn Nash:** Also, now I'm just like "I want to try all these things."
**Justin Garrison:** I'm gonna be quizzing you on whatever our next thing is versus Java.
**Autumn Nash:** Okay, one last thing... Eraser.
**Justin Garrison:** Eraser. That must be the opposite of a 3D printer.
**Autumn Nash:** I know, but it's a cool name.
**Justin Garrison:** So I'm gonna say it's a 3D printer, because I don't know any Kubernetes thing.
**Autumn Nash:** It is Kubernetes.
**Justin Garrison:** Dang. Eraser? There's not even a K in there.
**Autumn Nash:** I know. I was like "This isn't even --" I was like "I don't know if I would think it was either." So I was like-- look, I thought really hard about this.
**Justin Garrison:** Oh, it's an image cleanup tool. Ah, yeah. Okay, that makes sense. There's a couple of those. Good times.
**Autumn Nash:** You almost make me want to do Kubernetes. And then I see things like this, where there's way too much to learn, and I'm like "Oh, no... Maybe I'll stick in Java and data."
**Justin Garrison:** Java and data... That's a small subset of things you have to learn... \[laughs\]
**Autumn Nash:** It makes me feel better than that... I'll just run data things in Java, in a very simple Kubernetes thing. Do you guys do simple? I don't think y'all know what that is.
**Justin Garrison:** Yeah, it's probably not. Thanks, everyone, for joining and listening to the show. One last time we'll shout out that I'm going to be traveling near the end of June, because I guess it's the last episode that's going to come out before SRE Day. So SRE Day in San Francisco, I'm talking there, and we'...
**Autumn Nash:** And hopefully he can sneak me into the Kubernetes thing, because I just love the Kubernetes crowd. I don't know what y'all do special in Kubernetes, but you guys have one of the best open source communities. I love meeting all the Kubernetes people.
**Justin Garrison:** I will 100% give credit to setting up the community in a very deliberate way by people like Sarah Novotny, who was one of the very first community leads for Kubernetes, and she --
**Autumn Nash:** Whatever y'all did community-wise, it worked. People are fire; they're so fun. They've all got cool hair... I was just like "I will learn how to run Kubernetes for y'all. I just want to hang out with you." That and the Postgres community seems very welcoming, too.
**Justin Garrison:** So thanks, Autumn, and thank you all for joining. We'll talk to you again soon.
**Autumn Nash:** Bye, guys.
• Discussion of fatigue from recent events
• CrowdStrike and its impact on hospitals and travel
• Comparison of Justin Garrison's childhood to Autumn Nash's (Disney-related)
• Martin Jackson as guest for upcoming episode
• The rapid evolution of infrastructure and application management in the last 10 years
• Deployment of Puppet at Walmart stores
• Discussion of Netflix's innovation and lack of self-promotion
• Justin Garrison's experience with Walmart's migration to Puppet 10 years ago and its ongoing nature
• The challenges of scaling infrastructure and automation
• Martin Jackson's role at Walmart in implementing configuration management with Puppet
• Lessons learned from deploying infrastructure as code, including the importance of conservative and safe deployments
• Overview of Walmart's internal architecture and server count, and the need for policy assertion and configuration management
• Architecture of deploying servers in individual stores, with half a rack of space used for overhead
• Using Foreman to provide centralized reporting and configuration management at scale
• Kernel version drift across 5000 stores, making it difficult to debug issues
• Legacy application deployment process using RPMs and Puppet
• Historical context of proprietary Unix variants (MP-RAS, HP-UX, AIX) used in store deployments
• Complexity of onboarded applications per store server (ranging from 800 to 5,000)
• Purpose of having servers in individual stores instead of a centralized area due to historical scaling issues and satellite networking constraints
• Walmart's IT infrastructure and the challenges of managing it at scale
• Mainframe-influenced architecture vs modernized systems and configuration management
• Legacy technology and the difficulty of migrating to new systems
• Challenges of deploying software and managing disparate systems across thousands of stores
• The importance of centralizing data and using tools like Puppet for configuration management
• The trade-offs between upgrading old systems and maintaining current infrastructure.
• Scaling Puppet infrastructure at Walmart from 2014-2016
• Aggressive infrastructure management with high risk appetite in 2014
• Deployment of Puppet to over 2000 stores in two months
• Using pagers and enlisting help from other teams for the deployment
• Rolling out Puppet in enforcing mode, minimal catalog, and no downtime
• Preparing other teams to assist with the deployment and training them on configuration management