Devops Prep in one year

The short: I'll be job hunting in a year and would like to transition to devops/SRE. Paths forward for a current systems admin?

Long: Solo systems admin, graduated with a BS in Comp Sci in 2012, went straight to an MSP (i don't know why...), became a solo sysadmin 'jack of all trades' at a 100 employee, two location, medical office in 2015 and have been there ever since. Exclusively Windows besides my ELK stack, PRTG, and an internal wiki. Minor scripting of some repetitive tasks (powershell, cmd) I've done a little, very little, python for a personal project. We have no cloud infrastructure. I feel pretty solid on networking concepts.

The wife and I will be moving to another state next July when she matches to a residency. No idea where. Could be East coast, West, PNW, Utah, PA, we don't know. Relevant maybe? Makes it hard to check most popular technologies in an area.

I'm digging through all the posts, stickies, etc and putting together a pile of resources to start going over. I'm reading The Phoenix Project, I've also got the DevOps and Unicorn books downloaded. Picking out websites, youtube videos, etc. I've got the roadmap, best practices, everything from the weekly thread.


I have a lot of downtime at work that could be devoted to this. (sorry current employer...) I've just got things running smooth enough that I have the downtime. I've got servers that aren't production that I can do whatever on.


What would you do in my shoes? Just read and play with the tech? Jenkins and AWS? Gitlab and Kub? Certs? Classes?


I'm not expecting to walk in to a full blown senior or even midlevel devops position. I expect a pay cut and a "junior" job title, that's what I'm shooting for. Probably a pay cut too (I make ~$70k now).

TLDR: 'Jack of all trades' sysadmin, 6 years as an admin, comp sci degree, wants to move in to devops, has a year to prep. What would your priorities be?

https://redd.it/ok9f5n
@r_devops

Allowing KVMs to reach the internet (Question)

Hi all, I'm having a slight dilemma with a current work situation.

I've got two interfaces on my CentOS 8 machine (internal network facing- eth1), and (internet facing - eth2), and I'm working on setting up a bunch of VMs to use for development purposes.

I considered creating a bridge (br0) and adding eth1 and eth2, but I lose the ability to SSH when I do so. I was researching on other ways, and I came across the use of NAT and macvtap.

I currently have NAT "working" - the VMs can ping the host and eachother, but fail with a "destination port unreachable" when pinging the internet.

Macvtap supposedly is a lightweight way of bridging interfaces, but again I was not able to ping the internet.

I've been writing a .xml template and using `virsh net-define <file>.xml` if the command matters.

Has anyone had any experience with allowing VMs to reach the internet with NAT or macvtap and could give me a bit of assistance?

https://redd.it/okaucj
@r_devops

What to put on Tinder bio?

Completely serious question but I'm trying to figure out what is the most effective

Devops engineer? Most people probably don't know what this means

Cloud engineer? A bit better but similar issues I think

Software engineer? I feel like this might be the one but it's a bit ambiguous

Any advice is appreciated

https://redd.it/okfbxk
@r_devops

How to control access for new users to run certain Ansible Playbooks to setup their work environment only?

&#x200B;

Right now were turning as much low level tasks, like creating users, to locking user accounts with Ansible Playbooks, aka IaC. These playbooks are stored in a git repo, with a BitBucket front end.

We also have two Ansible Playbooks that will automate the creation of a user's .gitconfig file and install software from a RHEL repo, for a new user setup.

My question is that ideally we would like for a new user to sit down at their computer, with Ansible engine installed, have then run whatever playbook that is only needed to get them setup to work.

How can we go about this so that a user can only run certain playbooks and only has privileges to run those playbooks? And once the new user setup is done, that is it.

https://redd.it/okgv9s
@r_devops

moving from a sysadmin/MSP role to DEVOPS

Hi everyone,

I'm posting to see if I can get any insights on how to transition from my current role, into more of a DevOps role at a software company.

Currently, I'm working at a software distributor, within its managed services team. We mainly look after cloud-based environments as a 'software as a service model', where customers utilise the software/platform, and we deploy then eventually manage the servers along with the software.

Because our company doesn't do any development, I feel like I am missing out on the CI/CD-related experience if I were to start applying for DevOps/SRE jobs. For context, my relevant experience after 2 years at my current role include:

- automating ETL processes, data backups, software patches using Python

- created Azure runbooks to schedule Azure environments

- used Terraform and Kubernetes to deploy environments

- used CloudWatch to monitor AWS resources and created Python scripts to parse IIS logs

- managed AD users, networking and security configuration, software licenses, and SSL certificates


I also have all three AWS associate certificates along with the CKA


Essentially, my main worry applying for DevOps jobs is that I have never worked at a company that had developers pumping out development for software. Therefore, I haven't really been involved with the CI/CD process that's a core foundation of DevOps. I have experience developing my own applications and and have deployed them to IIS. Although I don't imagine that is anywhere the same, as deploying heavily-used applications on production environments.


Also, although I have a Comp Sci degree, my role over the past three years hasn't been development-heavy. I hear that you need to be a decent SWE as well.


Does anyone have any insights on what I can do to transition from my current role to DevOps?

https://redd.it/okgamp
@r_devops

Junior Cloud Engineer Interview

I have an interview on Friday for a Junior Cloud Engineer position and I'm currently a an admin. I was wondering if anybody could provide an idea of what would be asked in an interview that I probably wouldn't be asked in a sysadmin or desktoptech interview. At this point, I know what I know so I know I won't learn substantially more from now. The main thing is being nervous about completely embarrassing myself. I'd like to think I interview well but cloud engineer is in a different class of roles that I have interviewed for and I don't want to be blindsided by the unexpected.

I know there are many jobs out there and it's not the end end the world if I don't get it but I live in the "lesser" city of a two city metroplex (in terms of IT job availability) but for once there is a GOOD job that is vertical for me in career at a great company open so my nerves are through the roof. My commute time would go down by an hour+ if I got this job so I'm really gunning for it

The JD qualifications pretty much only asks for some basic to intermediate windows experience and basic virtualization experience and basic networking which I do have. I have done some basic projects in AWS and Azure but the unknown factor of what I could potentially be asked has me worked up

https://redd.it/okijjf
@r_devops

Jenkins X

What are the capabilities of Jenkins X and what is its support for legacy Jenkins shared libraries?

https://redd.it/ojgv3n
@r_devops

AWS NAT Solution for inbound and outbound traffic?

Hi guys! I hope everyone is doing well. I've run into a problem I can't seem to figure out and am looking online for suggestions, help, etc. So any help is well appreciated

**What we need:**

* We have a customer connect to us through a VPN. In our case currently a site to site VPN setup on AWS.
* We need the customer to send traffic/data to one of our resources, but customer has to send this to an IP outside of the VPC CIDR.
* We need a device that NATs this IP into our VPC and routes traffic to a specific resource. We also need the outbound traffic to go through the NAT back to the customer.

**What we've checked:**

* We've looked at the Transit Gateway, NAT Gateway, Client VPN... But we can't find a valid way of doing this.
* The Transit Gateway doesn't seem to do NAT, and we can't figure out a way of using the Transit Gateway together with the NAT Gateway to accomplish what we need.
* It also doesn't seem to be possible to configure the NAT Gateway to NAT specific IPs to specific resources for both in and outbound traffic.
* We've seen the option of using a NAT Instance (which AWS seems to have moved to the NAT Gateway...), and think that maybe this is the least complicated method?

Simple diagram to depict what we're trying to achieve:

[https://forums.aws.amazon.com/servlet/JiveServlet/download/8-343034-989711-34061/aws-nat.jpg](https://forums.aws.amazon.com/servlet/JiveServlet/download/8-343034-989711-34061/aws-nat.jpg)

https://redd.it/oiuhoo
@r_devops

Any team leaders on this sub?

I am curious to know what team leaders (whose teams participate in DevOps) think of a project I'm working on. Please, please, please find holes and critique as if you were aiming to start a flame war.

**Here's a concept summary:**

* It's a continuous feedback sharing and learning tool
* DevOps is the first space I want to address because of its sheer complexity
* You map the Ops activities your team does\*\* then write/link notes to them
* Your engineers spend about 5-minutes per day reviewing notes you and their peers share

\*\* Mapping is done by selecting from a DevOps capability map

Now, you might be thinking, "Why don't we just do this on Slack?". Slack channels better serve ephemeral content, so why not a clean, dedicated space for sharpening your abilities?

**Expected benefits include:**

* Supplements your 1-on-1 coaching and engineer's ongoing certification studies
* Boosts efficacy of work by linking feedback and learning direct to relevant areas
* Help neurodivergent tech workers grasp feedback and learning better due to visual context

So... let me know what you think :)

https://redd.it/oklwda
@r_devops

Been out of the job market for a few years and now everyone requires k8s experience, am I screwed?

Background: I've spent the last several years working in the public sector, which sometimes lags behind the private sector. While we use docker for local development and CI/CD pipelines, we're still deploying to bare EC2/ASGs. We do have plans to deploy to ECS/Fargate but god knows when that will get prioritized by stakeholders.

While I'm not exactly unhappy, I do feel stagnant and the job market is extremely hot right now. I started looking at listings for the first time in 2+ years and it appears everyone and their mom now expects "deep kubernetes experience".

So I ask you fine folk in other industries: is this just HR speak, or have I effectively been locked out from anything but public sector?

https://redd.it/oklqul
@r_devops

Is there a self-paced lab Kubernetes deployment practice like google cloud in AWS?

Hi Devop experts,

I will be going for AWS to try deploying my app which I have done like 2 years ago but have not the time to deploy it for one reason or another.

But, really, I am not sure if Kubernetes is the way I should go for..perhaps Kubernetes is for big corporate and not for my small app?

And after reading on the latest about FASS (Functions as a service) which is like a one time thing...I am not sure what part of the app or normally what people do to make that part of the app to deploy on FASS? Would it be authentication or what ?

Hope I can get some insights here. Thank you guys.

https://redd.it/okm20g
@r_devops

Debugging/Testing CI pipelines

At my company we are using jenkins as our solution for CI. We are using a self-host environment with a large number of pipelines. In many cases our pipelines get large with a lot of content.

A common thing we have are optimizations. For example if a job is rebuilt, e.g some flaky test failed, we want to avoid building the artifacts again. While this lead to a great improvement in the overall performance of the pipelines it added a lot of complexity to the development flow.

We tried to use tools like https://github.com/jenkinsci/JenkinsPipelineUnit. That ended up being even worse. It took a lot of effort to maintain these tests, mocking every single plugin that we use. Add to that the fact that you need to write it all in java. Which isn't the "home environment" for most DevOps at my company.

I saw that other tools, e.g CircleCI, also offer such abilities and potentially complicated plugins.
As can be seen in this guide https://support.circleci.com/hc/en-us/articles/360043638052-Conditional-steps-in-jobs-and-conditional-workflows. While they do offer some support for development, like https://circleci.com/blog/local-pipeline-development/, it still feels like something that will be really hard to use on a real project.

So it got me wondering - how do you handle this issue on your environments?

https://redd.it/oiue44
@r_devops

Permissions are driving me crazy - DevSecOps

I wanna quit work for that and only reason.

I am a DevSecOps engineer. More specifically I secure AWS, deploying native solutions in a multi-account setup. We manage everything through terraform.

I am at this position for 4 months but the work I have managed to complete corresponds to one month of work. The other 3 months have been wasted in waiting for permissions grants, from more senior engineers. The permission grants completely block my tasks everytime.

I cannot connect properly with my team either. Although they are willing to help, they respond to me with delays of 5-6 hours every time. And it may take days to resolve a 10 minute issue. In the end, its me who cant produce work and seems like the unproductive guy.

Have you been in a similar situation? How did you tackle this?

https://redd.it/okoww3
@r_devops

Trying to access redis container remotely, not able to get it to work

Sorry if this isn't the right place to post..

I'm building a redis container with Dockerfile:

FROM redis
COPY redis.conf /usr/local/etc/redis/redis.conf
CMD  "redis-server", "/usr/local/etc/redis/redis.conf" 

The redis.conf says:

bind 0.0.0.0

Under YML file services:

redis:
build: ./services/redis
ports:        
- "6379:6379" restart: always

I can access redis from other containers in the same machine, but I'm not able to access it remotely using redis-cli. It says "Could not connect to Redis at <ip>:6379: Connection timed out".

I have another container with nginx on the same VM, listening to ports 80 and 443. With my limited knowledge I feel like I don't need to make any changes to nginx configs to get redis to work remotely, but I'm not sure.

What should I do?

https://redd.it/okpogw
@r_devops

That's it I am never gonna get a job in this industry

I have a low IQ It takes longer to understand the question I don't do good with ticking time in the interview. I am good programmer I know it. I just can't do these interview coding test, I hate how they amount all the hard work to these couple of these question and decide whether you will get a job or not. I wanna know how can I get better at it

https://redd.it/okqcoy
@r_devops

How to update container when new docker image version is published to docker registry?

I new to CI/CD, I was successfully able to create a pipeline that build a new docker image of my repo and push it to docker registry, but I'm not sure excactly how to update the running container to use the new image? I'm using Drone CI and docker-compose.

https://redd.it/okqefn
@r_devops

Hikaru 0.6b released with support for the 1.16 and 1.17 K8s Python client, newest black formatter

Hikaru is a tool that provides you the ability to easily shift between YAML, Python objects/source, and JSON representations of your Kubernetes config files. It provides assistance in authoring these files in Python, opens up options in how you can assemble and customise the files, and provides some programmatic tools for inspecting large, complex files to enable automation of policy and security compliance.

Additionally, Hikaru allows you to use its K8s model objects to interact with Kubernetes, directing it to create, modify, and delete resources.

https://github.com/haxsaw/hikaru

https://redd.it/oku2qx
@r_devops

Can anyone help with adding --net:container:CONTAINER_NAME option in kubernetes for a container.

Hi all, I have 2 containers, in one I have a java application and in another one I have open-vpn. I need to make the first container use the network of the second one. I managed with --net option locally on my computer and it works. Container 1 shares the network of container 2.

My question is how to do this in kubernetes? I am new to this stuff and also I am not a devops but a java developer so bare with me if explanations are not so good. :)

Thanks!

https://redd.it/okvhck
@r_devops

How does Azure's Deployment groups work?

Recently, I've been working with Azure pipelines to setup releasing our project to our On-Prem Dev environment. I was able to set up the a Deployment group in order to get my release working pretty easily, but I do have questions regarding how that PowerShell script works and how secure it is.


I would like to get a better understanding how this works before I start implementing it in our Prod environment.

This is the article I followed while setting up the Deployment group (https://medium.com/software-development-turkey/deploying-to-on-prem-server-with-using-azure-devops-pipelines-842f4aa226a7)

https://redd.it/okwomj
@r_devops

Service Sizing Calculation

I'm trying to build a sizing calculator for a service which processes work items, so that I can give my customers an idea of what their compute costs will be. For example:

* I have 500 work items arriving per hour. We can assume these arrive at a constant pace.
* Each work item will take \~110 seconds to complete, on average.
* There is a tolerance of no more than 15 minutes of delay between work item submission and result.

Is this enough data to determine the number of workers I need to stay current? I can't quite get my head around it, and can't seem to find the right terms to google.

https://redd.it/okwij8
@r_devops

A live coding exercise

Hey All.

So I will be appearing for the first round of technical interview with a cloud provider company.

As per the process, they are going to do the a live coding exercise.

I have been out of the interview loop for a while and wondering if any one has any suggestions about coding exercise that I can take/do before this interview.

Any other general suggestions will be welcomed as well.

Thanks.

iking

https://redd.it/ol0qqu
@r_devops