Reddit DevOps
268 subscribers
2 photos
31K links
Reddit DevOps. #devops
Thanks @reddit2telegram and @r_channels
Download Telegram
Can I install a VPN server within WSL to use the ZScaler connection within Windows?

So for the longest time I have worked exclusively on Linux and MacOS. However, the organisation I work for decided to go full Microsoft, and is forcing the entire workforce to work on Windows laptops now. Of course, as a devops engineer, this causes a lot of issues with the tools I am using. For the organisation it allowed them to tighten a lot of security holes, one of which is it now allows them to use ZScaler. They've completely shut off access to the network outside of systems that are connected through ZScaler.

Because we have a ton of different dev(ops) teams within the organisation, it's an almost impossible task to get all of the apps that were being used to Windows, or a Windows alternative. As such the organisation allows us to use WSL, which in turn gives us the option to use certain apps unmanaged by installing them within WSL. An example would be how I installed Chrome through WSL, to be able to use an instance of Chrome where I can use my extensions, such as my password manager and ad blocker.

Now I was wondering: would it be possible to install a VPN server, like OpenVPN, within WSL, connect to it using my old Macbook, and allow me to access the network now only accessible through ZScaler on the Windows laptop that I absolutely despise working on?

https://redd.it/1dto6zc
@r_devops
We did it! Karpor has been Open-Sourced!

🌟 Exciting news! We just open sourced Karpor, an open-source intelligent tool for Kubernetes! 🚀💻 With Karpor, you can bring advanced 🔍 Search, 💡 Insight, and AI to your Kubernetes clusters. Gain crucial visibility across any clouds and empower your development teams to ship faster. Check it out on GitHub: https://github.com/KusionStack/karpor

https://redd.it/1dtpf1q
@r_devops
Best user identity and authentication solutions

My company is currently using freeIPA to manage user credentials for logins to our ec2 boxes and infrastructure.
We are seeking a commercial alternative solution that will continue to be supported but finding a good alternative without knowing what the full implementation and migration road map for each alternative could look like in a landscape or look-alike products is tough.
Does anyone here have experience migrating off freeIPA or even just some wisdom to share on what could be a good solution?

https://redd.it/1dtr6hs
@r_devops
Spacelift worker management

Someone has alleged that if I'm able to get our worker count to 0 from Spacelift's perspective that it can allow us to "bank" time with billing to burst over our contractual worker counts.

We pay for 4 workers with a P95 billing measurement. Spacelift measures per minute and bills per hour. 95% of 60 minutes is 57. So for 57 of 60 minutes I have to be at 4 or less to avoid being over for that hour.

Then in a 30 day month of 720 hours the 95% mark is 684. So for 684 of 720 hours I have to be at 4 or less to avoid being over, and being billed an overage fee. Basically a 36 hour "buffer" where I can burst without a fee.

So if what was alleged is true how would one grow that 36 hour buffer? As far as I can tell if I'm at 0 for 56 minutes in an hour and hit 5 for 4 minute then I'm over and being at zero saved me nothing.

https://redd.it/1dtsx4u
@r_devops
How to handle stress as a Graduate

Hi there! It’s the first time I've posted something here. I’m a bit worried about my future career. As a graduate whose major wasn’t even CS, I got the chance to stay in the devOps team in security. The problem is I always feel like I’m a burden to my team. They are too busy to shadow me so I have to learn things on my own (but the result isn’t ideal since most things I can’t learn from the internet/GPT but from senior dev). As a result, I made a lot of mistakes in my work and I have to finally lean on my colleagues to help me out. Sometimes I have to deal with the pressure from other teams. After work, I’m really tired and I don’t even have the time and energy to learn coding or devOps knowledge. What should I do? I wonder how people become senior devOps from a rookie😭

https://redd.it/1dtu9iu
@r_devops
http domain redirection vs https


We've setup a static website on S3 and hosting it through aws cloudfront

The domain is on godaddy so all the DNS records are on it too.


now the redirection is setup like below,

Domain Forwarding: mydomain.com -> https://www.mydomain.com

CNAME: www -> sf76adsf.cloudfront.net


on cloudfront side, the distribution is configured to have alternate domain names with mydomain.com and www.mydomain.com with ACM

now when i'm accessing the website, i can access it from:

https://mydomain.com

https://www.mydomain.com

https://www.mydomain.com

but when i'm trying to access the https://mydomain.com, it's not opening.


can you guys please help me figure out what could be the potential issue??

Thanks

https://redd.it/1dtttrv
@r_devops
Questions about flux as a platform team

I was wondering if there is anyone here using Flux (probably applies to Argo as well) at scale as a platform team.

For example, my team provides the platform (k8s + several other domain specific components) and aims to make it as easy as possible for 100+ app teams to deploy to the platform.

This means we manage a central, shared repo with a very robust Helm chart. Then, each app has its own repo that contains app code and values files for the chart.

My question is, anyone here using flux in a similar model? If not, can you point me to some useful reading material for running this model at scale? I understand how it all works, I just can't wrap my head around making the model more distributed.

We are not tied to Helm and have explored moving to a Kustomize only approach, but that only seemed to add complexity for the app teams (verbose overlays vs a simple values yaml).

Any info or feedback would be appreciated.

https://redd.it/1dttmgd
@r_devops
Google Cloud Deploy Across GCP Projects?

How (if at all?) are folks using Google Cloud Deploy across GCP projects. My team has separate GCP projects per environment and am looking into Cloud Deploy. I'm envisioning something like this architecture. But I have a few questions before I dive in:

1. Is there anyone out there using Cloud Deploy to release across GCP projects? I can't find any examples or references but it seems possible to have Cloud Deploy in a separate CI/CD project and then have each deploy target point to a different GCP project (dev, staging, prod).
2. If it's common to use Cloud Deploy with multiple GCP projects, how are terraform plan/apply handled as part of the release? It would be ideal if I could conditionally request manual approval for a release if there's a terraform diff. Perhaps hooks could come in handy, but it's unclear to me exactly how.

This feels like a reasonable setup to me, but given the lack of information out there, I'm worried that this is bad practice and that I'm missing the reason why. Any thoughts would be appreciated!

https://redd.it/1dtsbc6
@r_devops
Senior DevOps => DevOps

I've worked at my current placement for 4 years over which time I've been promoted from DevOps Engineer to Senior DevOps Engineer and started to mentor our other guys and perform interviews for new candidates etc.

I am in the market for a new opportunity and I'm wondering if it's "bad" for my resume if I was to accept a role at a company that is hiring a DevOps Engineer vs. a Senior DevOps Engineer.

Like once you're a senior, should you always try to keep this title/role going forward or is it normal to see someone go from Senior to normal just because a company already had Senior engineers (or maybe you just don't want to be a senior)

I would prefer to keep it and I like being a Senior engineer and mentoring folks who are newer to the role but curious what the impact to the optics of my resume would be if a company is only hiring for DevOps vs. Senior DevOps

Any thoughts?

https://redd.it/1du0egd
@r_devops
An experimental, refreshing webserver.

Just drop the stuff here for anyone who desperatly needs a simple web server.

S/O to #nginxunit who built a solid ground for this
light webserver with critical modern features:

- Split configuration in multiple short toml files 😱

- Automatic Ssl certificate (no need to think about it) 😏

https://github.com/pipelight/jucenit

https://redd.it/1du639v
@r_devops
Networking to NetDevOps

Are there any network/infrastructure engineers here that have transitioned from regular Network Engineering to NetDevOps/Network Automation?

How easy was the transition, and what did you have to do to adapt?

https://redd.it/1dubfij
@r_devops
How to use udeploy properties in angular

How can one use udeploy properties in angular? In springboot it's simply ${propertyname}

https://redd.it/1dub646
@r_devops
Openssh vulnerability, can't find a RHEL 9 rpm. Do i need to just make one?

I'm new to dealing with situations like this. I have a bunch of RHEL 9 VMs that need the patched 9.8 version of openssh. I am used to just running dnf check-upgrade to see when new versions of packages are available. But there are no updates showing as available and I can't even find an openssh 9.8 rpm for RHEL 9 just searching out on the internet. How do urgent patches like this normally go? Do I wait for an rpm to be made available via official sources or do I need to figure out how to compile one myself?

https://redd.it/1dueq92
@r_devops
Docker course recommendation- 2024

Hi

I'm a python developer with aws cloud knowledge.

I got an opportunity to work as cloud engineer with more devops and less development.

This is an internal switch and my lead asked to get a grip on docker first and gradually move to k8s ( aws eks ) and others

Need your help in suggesting a udemy course for docker in 2024.

I have kodekloud subscription also. Kindly share a course link that covers everything related to docker.

TIA

https://redd.it/1dufdpq
@r_devops
Looking for advice of coaching staff

I’m a team leader of small group containing no more than 10 people. We had a new member joining team 3 months ago , who was just graduated.

I assigned multiple tasks for him, and none of them worked very well, he worked pretty slowly and seems to be very hard to understand my intention. I’m kind of frustrated coaching him, so I want you guys suggestion. I’ll introduce the tasks I assigned and what I saw.

Task 1: Upgrade grafana from 9.3 to 11.0

So I was imagining just backup data, edit image tag and done.

But actually he went to looking at the database difference between the two versions, trying to upgrade grafana and import the old database, hoping this would work.

Task 2: develop cloud controller manager, sync k8s tag to cloud provider

Cloud controller is an event handler for k8s resources, it listens changes of k8s manifest and do some change on cloud provider, so the simplest solution is on every change, sync the tags.

The problem is that the code he gives has no code reuse, say the resource would be created or updated, both would trigger a tag sync, the basic logic is identical but he did not reuse. He claims that the error handling is different so cannot be reused, but it’s written in Go and certainly we can create a function named sync_tag and return error if anything goes wrong wrong, call it in handlers , check the error and then handle the error gracefully.

Even after I told him the way to do that, it still took him half a day to implement it.


Task 3: replace ntp server and client from ntpd to chronyd

This task is quite straightforward, and he did it ok, writing puppet module, test.

But he cannot troubleshoot network issue, we had a few vm that has limited network access, on those machine time sync would fail, he did not find any way to troubleshoot network issue, at last I told him it might be iptables issue and finally he found it and fixed the issue.

So this is the experience of me coaching other staff, do you think there’s anything I can do to be a better coach?

https://redd.it/1duhlmb
@r_devops
Kubernetes cronjob for periodical Cloudflare backups

I've organized a Kubernetes cronjob to make periodical Cloudflare backups in Terraform format and saving them into your repo for easy and fast business continuity plan.

https://github.com/nullniverse/cloudcron

https://redd.it/1dugi7k
@r_devops
How do you manage Datadog integrations at scale?

We are rolling out Datadog but creating an ansible task to roll out each integration seems like a lot of toil. does anyone have any suggestions on how to manage this?

https://redd.it/1dulacy
@r_devops
Github Actions fails to find UnitTest.csproj

I'm currently trying to set up a GitHub actions such that when a PR is made to the developing branch, a series unit test runs on the PR. This is a dotnet application, and here is the dot.yml:

name: .NET


on:
push:
branches: [ "developing" ]
pull_request:
branches: [ "developing" ]


jobs:
build:


runs-on: windows-latest


steps:
- uses: actions/checkout@v4
- name: Setup .NET
uses: actions/setup-dotnet@v4
with:
dotnet-version: 8.0.x
- name: Restore dependencies
run: dotnet restore
- name: Build
run: dotnet build --no-restore
- name: Test
run: dotnet test --no-build --verbosity normal

Here is how the project is structured: [https://imgur.com/a/EW7YnoG](https://imgur.com/a/EW7YnoG)


When I test out the actions it fails, and I get the following message:

Run dotnet restore

C:\Program Files\dotnet\sdk\8.0.302\NuGet.targets(414,5): error MSB3202: The project file "D:\a\project\UnitTest\UnitTest.csproj" was not found. [D:\a\project\project\project.sln]

Error: Process completed with exit code 1.

I'm not sure how to resolve this, the UnitTest folder is in the project, and it is only in the developing branch, it is not in the Main branch. I have tried adding `run: dotnet restore CBAS/` but that didn't help. Searching online for solution told me that the file name could be too long, but I don't think that's the case with my project. When I run `dotnet restore` locally everything works. Any idea what I should do?

https://redd.it/1duie79
@r_devops
What can you do as a devops to help implement SOC2?

Have you ever been asked to make all the changes necessary to implement SOC2? I read about it and it doesn't seem like something developers or devops can do much about. The only thing I can think of is detecting vulnerabilities, creating unit and integration tests and that's it.

https://redd.it/1duovo9
@r_devops