Forgotten Origins of LXC/LXD

I'm not sure if this is the correct place compared to /Kubernetes or /Docker. Maybe I'm also off base on this... Just wanted to ask about/share some thoughts/rant. I don't even have a problem just talking to the void.

Is LXC/LXD not the origin of both Kubernetes and Docker? Docker originally being released essentially as a wrapper of LXC. And LXD being a container orchestrator supporting standard APIs, networking, cluster support, etc. I'm generalizing a bit, but the concepts behind LXC and LXD were well ahead of later adaptations of similar projects. These are essentially primitives of our current abstractions, yet I never hear/read anything about LXC/LXD anymore? There is no drawing correlations or explaining the alternatives in recent blog posts, guides, etc.

Why is it? Both projects are still alive and supported, seemingly able to solve a lot of the same problems. Is it just not preferred anymore due to more features or usability of newer products (I assume)? Is it performance and I'm stupid to run containers in LXD cluster? There is a certain level of "pureness" that I like.

One last thing - is anybody using Incus or seeing more popularity? Actually reason for this post is due to seeing this project seems to be lead by previous LXD maintainers. Except it seems not tied to Canonical this time. I'm interested but don't see much traction yet.

https://redd.it/1c4kp5o
@r_devops

Question - How do you guys do Patching?

Question on how you all do Patching - both Windows and Linux? is it integrated in any way into your build process or do you use a dedicated patching software? If you use a dedicated software, what do you use?

Any thoughts, regrets, lessons learned that can be shared are helpful.

We still use WSUS with the old Solarwinds add on, but we want to move that into something modern and more effective. Ours is "ok" but needs improvement in all areas - Installation, Compliance, Reporting.

https://redd.it/1c4mayd
@r_devops

How do you define your SLA?

I'm trying to brush up on my basic SRE chops and was reading ye olde Google posts on calculating SLOs based on past performance, and I know that SLA's are supposed to just be an agreement to meet that SLO, but is this really how it works in your organization?

Back in the day the answer often boiled down to 'our biggest enterprise customer forced us to guarantee this SLA,' and since so many other decisions like the cadence of monitoring are based on your SLA, how does your team define the SLA you're trying to deliver?

https://redd.it/1c4j7da
@r_devops

Any ex-TAM(technical account manager) having/had issues to return into the hands on IT market?

The title is self explanatory.

https://redd.it/1c4rwyg
@r_devops

I’m conflicted

Last year I’d taken over the CI/CD for the entire UI team wrangling their massive web monorepo, porting iOS to Az and building everything from the command line, as well as Android. I was handling building and detecting all of the apps on the web, containerizing them, and generating reports for them. Backend was a different story as they do something completely different.

So everything was nice and consolidated into AZDO.

Well, the UI director recently decided that they were taking over CI and I would handle delivery of the artifacts. Now they’re using GitHub Actions.

Now I feel like we’re digging deeper into silos.

I often say that DevOps should be a culture not a title. And part of me is a little relieved not to have to worry about it, but also it’s just going to be further fragmented and harder to maintain.

Do y’all think this is a good change or stepping backwards? Does DevOps handle the entire lifecycle or just the logistics of getting packages from A to B?

https://redd.it/1c4swmo
@r_devops

About Fullstack Web App Deployment.

Hi team,

I have a frontend (react) , backend (spring) and mysql database, how can I deploy. Can I deploy all in any one compute engine like digital ocaen droplets or deploy the frontend in Netlify and backend and db together on Compute Engine. Which is the best approach, if I use netlify how it balances the load, where as in droplet containing all in one VM I use nginx as load balancer.
Kindly le me know which one should I go for. Which optimized and the best way to do deployment. Thanks

https://redd.it/1c4v1eb
@r_devops

Keeping Old Deployments

I don't have too much devOps experience, but I'm wondering what others do when it comes to old deployments on production, staging, and dev servers.


Is there any value in keeping all old deployments on any of these servers?
Is there any harm in deleting them?
What is best practice for "cleaning up" deployments on servers?


I'm starting to run out of hard drive space and deciding whether to cull come deployments or just add more gigs to the servers.
Thanks in advance for any thoughts.

https://redd.it/1c4vo4t
@r_devops

"You Build It, You Run It" is it a lie?

Yo, recently, we had a podcast guest on the show who talked about the evolution of DevOps, and he had a hot take of why the classic "you build it, you run it" approach is a lie and equated it to asking a mechanical engineer who builds a plane to also fly it.

Usually, the "You Build it, You Run it" is a traditional approach of separating development and operations. It has been challenged by the need for developers to take ownership of their code in production. This shift requires devs to not only build the software but also be responsible for its smooth operation and maintenance.
Devs, what are your thoughts? Do we hate it? If you're not a fan of 'you build it, you run it,' what's a better way to go about it? Or maybe that's full of crap, and you truly can both build it + run it.


You can listen to full episode here for context: https://soundcloud.com/ambassador-livinontheedge/s3-ep5-the-evolution-of-devops-ai-why-you-build-it-you-run-it-is-a-lie

https://redd.it/1c4xv6f
@r_devops

Minikube

Hi, so does minikube itself maintain any directory which has all the list of containers that ran on it?

I was just digging around on some of the directories, found out that while executing this command : docker exec -it minikube ls -1 /var/lib/docker/containers

It provided me with a subsequent list of directories which had information about the containers.

Can anyone provide more insights on this?

https://redd.it/1c4tp24
@r_devops

Why not shift to Nomad instead of K8s?

I've been trying to deploy my infrastructure both on K8s and Nomad (+ Consul).

The only benefit I see, that covers up for greater complexity of K8s, is the availability of managed service providers (GKE/EKS). Which makes things easier.

But Nomad seems to fit my use-case of 8-10 services deployed behind a proxy and load balancer, and is so much easier to setup and use.

I've seen a lot of people agreeing that Nomad is more straight-forward, at least for developers who don't want to invest too much time in infrastructure. But there is still hesitance in wider adoption of Nomad.

Is it largely the unavailability of a managed Nomad (or HashiStack) provider who can run the scripts on your behalf, or some other reason?

I want to understand the primary justifications behind why you as current K8 users will not adopt Nomad.

Thanks ❤️

https://redd.it/1c57oot
@r_devops

Theoretical underpinnings of IaC

Tools like Pulumi / Terraform have a very complex state management IMO and diffing algorithms as well as planning to ensure the desired state is met with as little obtrusive changes as possible. Are there any papers that discuss how is this done? Asking for a friend!

https://redd.it/1c5a2rh
@r_devops

Career/Resume Advice Graduated 3 years ago and can't find a job

My resume

---

So I graduated nearly 3 years ago and had to do mandatory military service straight after, which was a total waste of time. Despite listing it as experience on my resume, I was definitely not doing those things everyday and due to no internet and limited infrastructure, I wasn't able to learn new skills and improve myself either.

Anyways, once I finished service, my mom got diagnosed with cancer literally the following week. I was depressed about it and I was staying at home helping her out. She was weak, couldn't use her left arm, and occasionally would just fall due to dizziness from the therapy and whatnot. However, she got better eventually and she's doing fine now. So I feel like I can't use "caring for elderly parent" as an excuse for my huge employment gap because she got better after like 6 months.

Ever since then I've just been applying and trying to learn in-demand skills. However, I never get a call back; rather, it's always the default "we've decided to continue with another candidate" email. Getting a job in my country (Egypt) relies on nepotism for the most part, but no one in my family really knows any people working in tech. In the space of a year, I only got 2 offers, both through nepotism: one was for some Java developer role offering $60/month and an ERP consultant role for $200/month. Given these salaries were essentially unlivable wages and literally had nothing to do with the skills I know or related to the DevOps/Cloud career I'm trying to pursue, I declined those offers. I've been trying to learn new skills, but it seems to not amount to much as employers prioritize on-the-job experience with those skills, and not at-home projects.

So now I'm stuck with essentially a 3 year employment gap, unable to find any job. I'm sorry if this feels like a vent post, I'm just totally lost right now and feel like a burden to my parents living at home while they feed me on their retired pensions. I don't know what to do anymore.

https://redd.it/1c59p3p
@r_devops

Free Azure Devops

I need to teach myself Azure Devops and as part of that I tried https://azure.microsoft.com/en-us/products/devops/ (Start Free). My environment was setup in minutes. Is this totally free? I didnt enter any billing information. Just wanted to make sure because I had a bad experience with AWS.

https://redd.it/1c59uz7
@r_devops

What's the most significant pain you've faced when automating your deployments?

It's driven me nuts how much documentation I've had to read for minimum permission deployments on the three cloud providers within the command line.

A few weeks back, I finished 5 part-time days of constant back and forth between GCP (which I consider more friendly than AWS) to deploy a dockerized application to Cloud Run.

Note that I'm more of a developer who wants to "just run" the application without paying enormous fees to wrapper services like Vercel.

What's been the most considerable pain you've faced over the years?

​

https://redd.it/1c5fzxq
@r_devops

Transitioning to DevOps from Data/AI

Hi guys,

I am a mid-level engineer with about 6 years of experience in building Data and ML applications. In the early years of my career I spent a lot of time on the maths/modelling side (C++/Python) and later on found the systems and engineering (Cloud & devOps) side of building applications more enjoyable.

I haven’t had a formal training in computer science (my bachelors was in Robotics and Master’s in Applied Maths), so I’ve learnt most of what I know on the job. I’ve worked on the devOps side of things - CI/CD, Iac (Terraform), Azure, AKS- for the last couple of years in my roles as a ML/Data engineer.

Late last year I had to resign from this Data related role due to relocation & other reasons and I’m finding it quite challenging (unlike previous times in my career) to land a new role, preferably in DevOps. Having applied to many roles in the last couple of months, I am questioning why I am not even getting interviews, and would like some advice on what I should do to position myself better.

What I enjoyed the most about DevOps work was the satisfaction of helping all others devs work efficiently, and being able to work on system design to some extent.

I understand that DevOps is a somewhat senior role, for those who’ve had experience in other SW roles before, so I feel that I’ve been in the tech industry long enough. Then again, from the rejections I faced Im wondering are companies interested in folks who’ve had 5+ years pure of DevOps experience on a particular tech stack? Or do I apply to really Junior positions?

Thank you!

https://redd.it/1c5hdg7
@r_devops

Recommendation for running a few containers in a small environment

What is the best way to run a few internal (but important) containers in an environment with low/no DevOps experience?

We have a branch office with no DevOps people around. They want to run a few services which are delivered as containers. They have a few IT staff who take care of SaaS apps and on-prem infrastructure.

I'm wondering if a machine running Docker + some startup scripts might be the simplest way to go... Is this such a bad idea?

Kubernetes seems like it would be impossible for them to maintain. Even a managed Kubernetes service like GKE requires "someone" at the helm to operate it.

Is there a happy medium for this kind of situation, or is the general consensus "containers -> welp you better hire a team to run them!"

https://redd.it/1c5fw2q
@r_devops

My self-hosted app for devOps engineers to deal with all the tools and technologies

I created an app for **devOps** engineers called Snipman.io >>> [https://snipman.io](https://snipman.io)


* It is a self hosted code snippet management app (currently free to download on Mac and Windows) that basically lets you store snippets by snippet types.
* I primarily created it because I found myself creating a lot of text files for small code snippets for different **devOps,** **cloud** and **programming** tools and technologies for e.g **AWS**, **GCP**, **Terraform**, **Kubernetes**, **Docker** etc. This not only resulted in a lot of clutter but also a pain when it came to searching.
* My goal was to create something that would allow all the commands, configs and random snippets to be stored in a central repo locally and then have the ability to search them quickly. I think my app helps achieve all of that in through an elegant and simple to use GUI based tool.


I hope all the **devOps** engineers here find it useful and look forward to any feedback!

https://redd.it/1c5huqe
@r_devops

migrating from a cname to an arecord?

Typically when migrating a production lb, I'm able to do some sort of weighted dns routing to ensure everything is running smoothly as I migrate. Anything like this for moving from a CNAMe to an A record? Trying to move from a host backed by vercel to something we serve ourselves.

https://redd.it/1c5kjxj
@r_devops

Have any of you successfully used the Domino's terraform provider?

https://registry.terraform.io/providers/MNThomson/dominos/latest/docs

I'm in Canada and the provider just crashes (but I didn't try that hard to get it working haha). The readme is still good for a laugh though

https://redd.it/1c5me86
@r_devops

What is a good transition plan for moving on from the current job to a new position in the future?

I currently am in a job where I literally have landed in a Swamp with the product I am working on. There is no clear roadmap ahead and because it is a Big Fortune 500 Corp., the amount of bureaucratic work is more than actually Engineering.

I am still relatively young and do not own anything that holds me to the current job / location. I have "quietly quit" the job at the starting this year - I am working towards only one major release due in Start of 2025 and I am planning to move on.

With the buffer of about \~10 months, what would be decent way to plan this transition? I really have invested time into Golang - because I sincerely like the language and I would like to improve my efforts professionally but not in a haphazard manner. Previously, I was jumping on multiple topics and never actually learned something so I am going to go "slow" and keep sticking to Go and the tools that I have strong background in - Ansible, Docker, Docker Compose v2, QEMU, Packer. I do have Open-Source projects that are moderately "popular" (nothing in the 100+ stars but I do get Issues and Patches) and I do contribute to Open-Source Projects.

But any insights from Senior Engineers about a potential way to utilize this buffer time would be highly appreciated.

P.S> I do have about \~8 years experience as a professional. 5 years in R&D and 3 in Engineering (current job)

https://redd.it/1c5n8gu
@r_devops

Kind of freaking out.

I recently got a position as an AppSec/DevSecOps Engineer and I am also focusing on Cloud Security. Yay!! This is so exciting!!

This will be my first role in this capacity as I have never held these titles before. I have been in IT since 2016, but this is my first strictly cybersecurity role.

This is a new department that I have been tasked to help develop.
I am researching and trying to find the best tools and software we can use.

GitHub and GitLab are the top of the list for VC and also for our CI/CD systems.
I am looking for recommendations on SAST, DAST, and SCA tools to use that are low cost as this is a non profit.

We are an AWS shop and right now the developers are using a lot of cloud-native services.

I am kind of freaking out, and keep telling myself I am way over my head. Then I have a proud moment where I think that I am that girl for actually making it!
Then the terror sets in that I actually have to build out a whole department from scratch.

I am taking SEC540 from SANS, and I hope that it sets me up for success!

Any help is greatly, nay, ginormously appreciated!


https://redd.it/1c5vojs
@r_devops