How to reassign a secondary NIC to become the primary and vice versa in azure?

is there a way to swap a primary network interface and a secondary network interface, so that the original primary NIC necomes the secondary, and the secondary becomes the primary?

https://redd.it/nl8y8w
@r_devops

Possible solutions for quickly spinning up servers based on new domains

Just a heads up, this question is more for me to know what to learn and study rather than a active problem I am trying to fix. So I have been applying to jobs recently and encountered a question that I didn't really have a automated solution to.

Lets say my client currently has a server running angular, react, etc. for the frontend and it is hosted on this domain: cali.xyz.com. They want to expand to other states but it requires a lot of time since a user has to manually create a brand new server, install all the dependencies, upload the code, hook up the new domain, etc., so they want to automate or simplify the process as much as they can since they will be creating one for each state within the US. This would be hosted on AWS if that helps.

How would you go about this? I was thinking of using Jenkins/CodePipeline to trigger a Docker container to automate the dependency installation part but I am lost as to how to automate or simplify the rest

Unrelated question but how did you guys get experience with devops? Was it through a course or just using the softwares themselves and playing around with them?

https://redd.it/nlpsn1
@r_devops

Making SQL Queries Faster on DoltHub

Just published a blog on improving the performance of SQL web queries on DoltHub. Web queries are a feature that enable users to execute SQL statements against public DoltHub databases. The database format, Dolt, is a versioned SQL database with Git semantics. You can think of DoltHub as the data equivalent to GitHub, and Dolt, basically a hybrid of Git and MySQL.

Anyway, we discuss how we've adjusted the storage layer of Dolt, to boost reads from S3, which helps more of the queries complete before timing out. Enjoy!
https://www.dolthub.com/blog/2021-05-26-improving-web-query-performance-dolthub

https://redd.it/nluf1n
@r_devops

JJ's Star Wars and the Importance of Creative Frameworks

Recently, I was reading this article where JJ Abrams talked about his regret of not having a plan in the storyline of the new Star Wars trilogy. https://collider.com/jj-abrams-star-wars-sequel-trilogy-plan-comments/


And it just recently occurred to me that the new Star Wars trilogy holds a perfect example of the pitfalls of collaborative creation without proper leadership and frameworks.


Yes, what they built was cool. Yes, it made lots of money. Yes, there were lots of great ideas. But because there was no framework, it ultimately lead to a product that is unstable and a bit of a letdown to the customer.


In many organizations, I see similar pitfalls. Can an enterprise create a viable product? Yes. Can it make lots of money? Yes. Will it stand the test of time and lead the industry? Probably for a little while. Until a competitor that does it better comes on to the scene.


Devops is often adopted by enterprises as a "silver bullet" for the problems that ail their hobbling S&T divisions, but it's often implemented without restructuring the organization as well. They believe if they can just put the most brilliant people in the right places, their product will flourish. However, this is often NOT the case. Rather, the product may come together after hours and hours of midnight work and extended deadlines, but the cycle of issues never dies. Instability will plague the product until they understand these concepts. Resources will also be depleted faster. Turnover will be more rapid and execs will see everything as a huge success but with little longevity to show for it.


In conclusion, Disney had every ingredient right there! But they squandered their opportunity because of a lack of a proper framework. Brilliance only gets you so far. Structure makes things last.

https://redd.it/nlosgd
@r_devops

Can you connect to AWS from CircleCI without AWS creds set as secrets in env variables?

Github provides AWS integration which means you don't have to rotate keys and can deploy though codepipelines, at least this is how I understand it.

Github also allows you to deploy self hosted runners into the AWS account for FREE. Another way of by passing AWS keys and their management.

Does the CircleCI AWS integration do the same? I read the docs but don't get it. Also the fact that CircleCI charges for self hosted runners is a joke.

Am I misunderstanding something? I asked support and they just suggested that I write a lambda to rotate keys and then change the env variables. Surely there is a better way!

https://redd.it/nluz51
@r_devops

DevOPS and change management

Just curious what everyone’s opinion is regarding DevOPS and change management.

I’m from an infrastructure background and used to seeing changes submitted, deployment plans/testing and approvals.

Today I had a conversation with a peer after an individual from another Team almost caused a large service outage - we were able to intervene before things went south. The conversation went basically like…

Me - …dude almost broke x-y-z. He didn’t tell anyone he was making these changes, review them or put in a change request.

Other guy - If DevOPS had to start doing change requests then it would defeat the purpose of DevOPS and being agile.

I didn’t really know what to say in the moment. Curious to hear what everyone else thinks about DevOPS and change management?

https://redd.it/nlyeme
@r_devops

How to run docker in GitHub self hosted runners?

So I can get docker running and I can get the image downloaded but then the workflow fails. Something about no HOME and dockercfg...

Is it even possible to set up a workflow with:

Container: some-image:latest

Like you would do for a GitHub hosted runner?

https://redd.it/nlm5yv
@r_devops

How Culture Impacts Technology Choice: A Review of Netflix’s Use of Microservices

Over the last year or so (ever since founding Komodor), I have read many books about culture, trying to make sure we build a great atmosphere in the company from the foundation. One of the books I liked most is no rules rules. Not only is this a GREAT book (and not techy at all), but It also made me think about the rise of microservices, devops, why I (think) it worked for Netflix, and why most companies need to be more careful: https://thenewstack.io/how-culture-impacts-technology-choice-a-review-of-netflixs-use-of-microservices/


The book itself is super recommended for anyone more interested in devops as a movement to empower other people in the organization, and I think it can also help when making tech decision :)

https://redd.it/nlzzqz
@r_devops

Am I too inexperienced to be a DevOps engineer?

I recently got an offer to join as a DevOps Engineer at a AWS Premier Partner firm. But I'm not exactly sure if I have the right foundations to succeed in it though.


I graduated last year with a bachelors in CS, and Covid took away all my good dev offers, and I ended up as a dev at a tiny place with no testing, no CI/CD, no IAC, etc. I started off as a backend dev, did that for a few months. Then volunteered to research and integrate testing frameworks, and wrote all the initial unit tests. Development was very slow we had to wait for an outside consultant and come and fix things when stuff broke.


Out of necessity, I then learnt a bunch of stuff about AWS, and moved the application to it, and implemented CI/CD using CodePipeline, later wrote all of this stuff to Cloudformation. Containerized the application and threw it on ECS, etc. I also got AWS Associate certs using all this experience.


Now a year later, I was looking to change jobs for better pay and got this offer which doubles my salary. I'm happy about the salary, but I'm incredibly scared of getting not being able to live up to it.


If you take me out of AWS, I know nothing. And everything I've done, I've done it only one way with AWS services. I don't know anything about Ansible, Terraform, etc etc etc. The only best practices I know are the very specific ones I researched to implement the specific things at work.


I looked at all the junior DevOps engineers at the firm I'm joining on LinkedIn and the person with the least amount of experience has like 7 years of dev experience. Some of em were sysadmins for years. I know zero stuff about sysadmin. Maybe bare basics about networking, nothing about hardware, switches or anything.


The only thing I'm skilled at is coding skills: I can pick up languages very fast, and I can read and understand codebases very fast. At my current work, this helped me a lot. Since I knew the codebase and was aware of the features being developed, I could usually just guess what and who broke the system with intuition.


I've read that DevOps is a mid career role for highly experienced developers/sysadmins. Am I shooting myself in the foot by leaving development (something I feel confident at) one year out of college? I kinda feel like a jack of all trades, and master of none who just learned a bunch of buzzwords and tricked a company into hiring me.

https://redd.it/nldci3
@r_devops

SFTP versus SSH - one working, one not

Hi,

I'm able to connect to a remote server using SFTP on port 22 with a username and password.

But when I try to SSH connect to the same remote host, I briefly see the welcome message and then it says the connection was closed.

I'm kind of confused by this since they're both over the same port. And the credentials I'm using are the same.

https://redd.it/nlnoo1
@r_devops

How to run bash commands in Jenkins and not go insane with escaping characters?

I have been trying to run some bash commands on multiple servers through JenkinsFile(s). A sequence of pipes involving awk, grep, find etc.

The problem that I have facing is escaping characters in it. While trying to find solutions I came across

https://gist.github.com/Faheetah/e11bd0315c34ed32e681616e41279ef4

This helped a bit. But the problem is when I have single quotes, double quotes, dollars or slashes I spend an inordinate amount of time trying to find the right escape sequence. And the result is hardly readable and thus hard to maintain.

I tried to create a bash file on the server through Jenkins, running it and doing cleanup. But that also had it's own set of escape character issues.

Is there a guide or some articles which delve into this problem? Something with best practices so that I don't waste time fighting escape characters? If I do so many escape characters then it would be hard to edit or understand for anyone unless they also understand the idiosynchrasies of Jenkins. Looking for some best practices which leads to maintainable code.

https://redd.it/nlglbx
@r_devops

Packer + proxmox + ubuntu 20.04.2 autoinstall woes

I have hit some kind of wall with packer builds on a proxmox system. I'm modifying working code that works on VMware locally, vcenter or directly vsphere, but simply fails on my home proxmox node.

The issue is with the autoinstall. I feed it boot commands but it's almost as if it doesn't ever hit enter at the end, and winds up going to the language selection screen.

The relevant code:

 json
      "boot_command": [
        "<enter><enter><f6><esc><wait>",
        "autoinstall ds=\"nocloud-net;seedfrom=https://{{.HTTPIP}}:{{.HTTPPort}}/\"<enter><wait>",
        "<wait><enter>"
      ],
      "boot_wait": "5s",

I've tried probably a dozen different boot command sections from various google answers, but each ends at the same screen.

I call autoinstall in the user-data as well, and like I said, the exact same user-data works find for VMware builds.

Anyone have any ideas or successful examples?

https://redd.it/nlkat8
@r_devops

#002: Weekly Towards AWS Newsletter 🚀

At Towards AWS, we are working so hard to bring quality articles to AWS builders just like you. Last week, we started a newsletter to send the best articles of the week.

Today we released issue #002. Please have a read.

https://towardsaws.com/002-weekly-towards-aws-newsletter-fdbe55d3c7b1

https://redd.it/nm474x
@r_devops

The Mysterious Gotcha of gRPC Stream Performance

At Ably, we use gRPC to streamline our messaging service. But recently, the performance of a gRPC streaming server was worse than expected, so our realtime engineering team rolled up their sleeves and went sleuthing in-house.

This is the full recap, by Paul Cruikshank, one of our Distributed Systems Engineers:

The Mysterious Gotcha of gRPC Stream Performance

https://redd.it/nlf7ip
@r_devops

Hikaru 0.5b released; now with high-level CRUD-style methods for calling Kubernetes

Hikaru's CRUD methods simplify your code and provide the ability to use top-level objects as context managers that can optionally roll back objects to a previous state upon failure. The release includes a number of other requested enhancements and bug fixes.

https://pypi.org/project/hikaru/

https://redd.it/nm5rn9
@r_devops

SOLID Design Principles: The Guide to Becoming Better Developers

Product owners don’t always understand the implications of bad software design, as the burden falls on engineers to consider the best principles of software design.

Let's dive into the 5 SOLID software development principles that will guide you to write code that’s easy to maintain, read and understand, and make it easier to extend the system with new functionality without breaking the existing ones.

Read the full article here: https://adevait.com/software/solid-design-principles-the-guide-to-becoming-better-developers

Note: This post contains a YouTube video from Laracon's EU conference in Amsterdam that goes fully into detail about this important principle. Hope you'll enjoy it and find it useful. 🙏

https://redd.it/nm73ng
@r_devops

Interesting Take: Looking at Netflix as a case study on how their culture enabled their microservices leadership

Just read this great post on the New Stack called How Culture Impacts Technology Choice: A Review of Netflix’s Use of Microservices, and thought it was a really interesting take.

Since I always like reading about Netflix comparisons and how they actually managed to get to where they are from a tech perspective, I thought other folks might feel the same way.

WDYT?

https://redd.it/nm7ozk
@r_devops

Rancher?

Is Rancher still a good option for creating and managing k8s clusters?

https://youtu.be/LK6KbAlQRIg

https://redd.it/nm9jh2
@r_devops

Deply on premise machine with Azure Pipelines, handshake error

Hi, I'm new to Devops, we're trying to deploy using Azure pipelines, but we get this error:

>\##[error\]Failed to connect to remote machine. Verify the SSH endpoint details. Failed to connect to remote machine. Verify the SSH endpoint details. Error: Timed out while waiting for handshake..

We're deploying to an on-premise machine on Azure, we've got a Service Connection in Devops with Contributor permissions.

If we use a hosted agent on other machine in our Azure, there's no problem deploying, but we want to use Azure pipelines, I did not find any prerrequisites, like installing an agent on the machine or open firewall communications, other than using a Service Connection.

We're trying to log into an SFTP server on the machine, that works fine, and we do not see logs of connection attempts, so I suspect that is a conectivity issue, but I don't know how to look for it in Azure.

Thanks in advance guys.

https://redd.it/nm7djw
@r_devops

Why do build stages need to be on different servers

Very new to DevOps so apologies for what is almost definitely a dumb question. I'm running a simple application via jenkins build, test, deploy stages that are all on different openstack instances. Why can't I just build these stages one after the other all on the same server / what are the drawback of this?

https://redd.it/nm78mz
@r_devops

Hashicorp Vault 2-tier CA

Hi guys,

I got a question concerning Hashicorp Vault and building a 2-tier CA. I know, that in a production environment i have to separate the root CA and the intermediate/issuing CA. There are some guides out there how to setup Vault as an intermediate CA, but they all say one should use an "external root CA". Is is advisable to use another Vault instance to be the root CA, turn this instance offline and build up a second Vault instance as the intermediate/issuing CA? Or should you use another piece of software for the root CA?

Thanks in advance!

https://redd.it/nm3upo
@r_devops