Mesos is Dead. Why did Kubernetes Kill it?

title

https://redd.it/enwzrz
@r_devops

How likely was the vent on Death Star added to the design by an out source worker that never cares?

In a galaxy far, far away. There was an out source worker working on a contract project for fixing the thermal simulation failure on the reaction core of the Death Star. He was paid a quarter of the wage of a Jedi trainee and works 12 hours a day, but was assigned to this blocker issue with a 2 week deadline attached.

To his delight, a flaw on the simulation software was discovered. Simply add some straight vents from the surface to the core, the simulation will no longer fail any tests. He discovered that if only one vent can be added for tests to pass, which has the smallest footprint in the design, the vent should be at least 2 meters wide. "This footprint is small enough on a 160-km sphere for anyone to figure out", he had in mind. With the time constrain rolling over him, he submitted a merge request with the change.

The request was reviewed by a full-time imperial engineer, an individual contributor who has no responsibility on contractors. He questioned the design and started an email thread, requiring the contractor to explain WTH he was adding to the design but got no reply. Couple days later, the design owner logged in and checkout the contractor's branch. Run test, no warning, no error! But there is a comment? Screw it for more work, he would just merge it.

2 days to deadline, the Imperial PM responsible for thermal simulation project finally showed up and asked everyone to close on it. He saw some resolved discussion on the updated design, the all-green spread sheet of simulation results and the pressure sent down from the command chain from Darth Vader himself, everything should be good to go.

The design was then built to the mighty ultimate weapon of the Empire, with the vent a contractor added there for fixing some test cases. As the result, the vent was not even blasting out plasma wind from the core and 2 proton torpedoes flew through it with no much resistance.

https://redd.it/env3jz
@r_devops

How much do you need to earn in SF Bay in order to live a decent life, and put away at least 2-4K in savings per month? Also, it is possible to go back to school part-time, to do your Masters, while having a full-time job as a DevOps Engineer? What do you recommend?



https://redd.it/entxe9
@r_devops

DevOps Interview Questions

Title says it all. I am currently a DevOps engineer and I am curious to hear what interviews questions you guys are asked. I would like to build on this to create a guide for myself and, if it's comprehensive enough, share it back with the community. Any guides you guys have would be great.

Now I know interviews vary in terms of difficulty. I come from a developer background and I feel as if those interview have been a bit more exhaustive. They would bombard me with programming questions and coding puzzles. So far my experience with devops roles has been very pleasant. For the most part, they would ask me about my role and the tools I use, and my proficiency with them. Overall, these interviews have be more like a casual conversation as oppose to a hostile programming interrogation.

Would love to hear about your experience and some "Gotcha" questions.
Ill start first, I found that a lot of interviewers have asked me about DevOps best practices, and found that the answers they where looking for came straight from the AWS White-paper "Practicing Continuous Integration and Continuous Delivery on AWS".

https://redd.it/eo5r1q
@r_devops

Starting out with DevOps, what would you do?

We're at the early stages of our DevOps journey, learning as we go. I thought I'd ask you some questions if you could start your journey again:

* What would you do differently, and why?
* What tools would you consider using now that weren't available at the time or you weren't aware of?
* What would your advice be to anyone starting out on this journey, where do you begin, what do you start to build and create first, environments, build tools, automation?
* Finally, do you believe DevOps is essential in today's modern cloud landscape - are there any other real or emerging alternatives?

Thanks!

https://redd.it/eo2sfk
@r_devops

Why Google/AWS/Azure's bandwidth is 10x more expensive than competition?

Cloud provider bandwidth pricing:

​

Google Cloud \[1\]: $0.105/GB (max) - $0.06/GB (min)

AWS \[2\]: $0.09/GB (max) - $0.05/GB (min)

Azure \[3\]: $0.087/GB (max) - $0.05/GB (min)

​

Linode \[4\]: $0.02/GB

DigitalOcean \[5\]: $0.01/GiB

Cloudflare \[6\]: Freemium (but they can suspend you without warning)

​

In comparison, I pay \~60$/Month for 1Gbit/s up/down internet at home, which means \~$0.0002/GB, which is 300x cheaper than Google Cloud's bandwidth.

​

How is this possible? Is it purely because of the top trio's market dominance? Or competitors are just taking the loss to get attract customers?

​

\[1\] [https://cloud.google.com/network-tiers/pricing](https://cloud.google.com/network-tiers/pricing)

\[2\] [https://aws.amazon.com/ec2/pricing/on-demand/](https://aws.amazon.com/ec2/pricing/on-demand/)

\[3\] [https://azure.microsoft.com/en-us/pricing/details/bandwidth/](https://azure.microsoft.com/en-us/pricing/details/bandwidth/)

\[4\] [https://www.linode.com/community/questions/11427/bandwidth-charging](https://www.linode.com/community/questions/11427/bandwidth-charging)

\[5\] [https://www.digitalocean.com/docs/accounts/billing/bandwidth/](https://www.digitalocean.com/docs/accounts/billing/bandwidth/)

\[6\] [https://webmasters.stackexchange.com/questions/88659/how-can-cloudflare-offer-a-free-cdn-with-unlimited-bandwidth](https://webmasters.stackexchange.com/questions/88659/how-can-cloudflare-offer-a-free-cdn-with-unlimited-bandwidth)

https://redd.it/eo43oz
@r_devops

Tools to backup databases in the cloud?

We are running postgres, mongodb, redis, ... databases in our kubernetes cluster on bare-metal servers.

Currently we use jenkins and shell scripts that run mongodump or similar.

Curious what you guys are using and if there is a tool available that helps with that task.

Every tool I found (velero / stash) just backups the persistent volumes. I am worried about data corruption here if there are write operations during backup.

https://redd.it/eo5yzu
@r_devops

Repo architecture

Hello I am about to publish my first open source project and I am a little confused how to organize GitHub repositories for it. Project is composed of 2 (3) elements: Frontend created in angular, backend created in spring boot, and third element which contain init scripts and docker-compose file to run it.

At first I was thinking about publishing 3 repos. It would be easy to create travis-ci scripts to build app and create docker image with separate context of backend and frontend. But more as I think of it it is really hard for anyone else but me understand and navigate through code base.

What is Your experience with such cases? It is better to create multiple repos or one with folders like frontend, backend, scripts? In second example how to handle CI? Building an app in single pipeline when changing only backend will cause to rebuild frontend as well.

https://redd.it/eo1b9q
@r_devops

How do you prefer to get help when you have questions about APIs?

Hi everyone,

I am a user researcher for a cybersecurity company and I am currently working on a project about API support.

I have created a really quick survey (less than two minutes) and would really appreciate it if some of you could answer my questions. I promise that I am not selling anything: I am only doing research. You could think of it like you are helping fellow API users around the world (which you really would be). :)

Here is my survey: [How do you get help with APIs?](https://surveys.hotjar.com/s?siteId=793793&surveyId=149005)

Thank you so much and have a great day! Also please don't hesitate to PM me if you have any questions about the survey.

https://redd.it/eob3nq
@r_devops

Software development management

Hi, I am a third year business student from Rotterdam. For my intership at an IT company I was asking myself... is it any useful to implement software development management processes(like DevOps) on a team of only 2.5 FTE developers. They mostly work on internal software like PSA and incident management systems. Like to know your opinions

https://redd.it/eob0vk
@r_devops

Reducing costs from S3 using Nginx Proxy?

I work at a video streaming company focused on the sports market, currently moving from on-prem storage solutions to AWS. Mostly, it’s because we want to focus on building great things, instead of having to worry about servers and in this specific case scalability. That’s why we chose to move from running our own storage servers to using S3 instead. Right now, we got it all set up and it all works fine, but we are struggling with the higher price tag. Although we embraced ourselves that we would get those monthly fees, I can not wrap my head around the fact that the data traffic is so darn expensive.

In order to reduce this, we setup a server with nginx, acting as a proxy for S3. Since we already had the hardware and bandwidth, so this made sense in order to get the costs down a bit on traffic out from S3.

It sure did get the costs down, but when I look into my logs I see that we have a cache miss ratio of about 30 %, which seems strange since most of the video we deliver is only relevant for about a week until it falls into oblivion (highlights and other sports related content only has a high viewing ratio for like 24 hours until people don’t care anymore).

I’m don’t really know what I am grasping for here, just hope that there is someone else who has stumbled onto this problem and might have a neat solution to it. As I said before, we are currently using nginx with proxy_cache, having the cache size limit to 1300G and a proxy_cache_valid for 200 OK set to 864h. Also worth to note is that we have enabled proxy_cache_lock, in order to only fetch the requested file once.

Anyone having any great ideas?

https://redd.it/eobx5d
@r_devops

Big consulting firm vs small dev group

Just wondering if anybody has done devops for both a large consulting firm and for a small tight knit dev group and can compare them.

I've been on a small dev group for several years now, and am kinda getting burned out and needing a change. In my current situation, I am constantly too busy leaving little time for personal growth and bigger projects, largely stuck on day to day stuff. We talk about big picture things a lot and it is exciting, but we never get to it. Also, the tiny on-call rotation comes around to me way too frequently.

A recruiter from a consulting firm contacted me the other day, and was considering the change. The pros of it seem to be focus on the technologies, and less BS to deal with. The cons being regular travel. What else am I missing and not considering?

https://redd.it/eo8j31
@r_devops

Ruby or Python more prevalent in DevOps? What are some must-know libraries for each that you utilize on a daily basis?



https://redd.it/eoegc0
@r_devops

Asterisk: command not found

Whereis asterisk
/etc/asterisk
I added to env variable with export PATH=$PATH:/etc/asterisk and source env variable file.
Still getting the same error.
I have installed it and wanted to access the asterisk gui, The port is configuration is all good.
Can’t reload asterisk cli. Googled the issue but still no help.

https://redd.it/eoe63h
@r_devops

Feel like Recruiter's make Job Application Process way more complicated?

I have applied to a bunch of job listings for a DevOps Position, and I get a swarm of email's and phone calls, which is totally fine, but because the Recruiter's themselves work for another company, and email you back from that companies email address, it makes it so hard to figure out which job posting they are talking about.

I'm sure it is just as confusing to them as it is for us, so I'm not blaming them or anything.

I want to make this process as smooth as possible, and in between the emails and calendars, it becomes really stressful trying to discern what is what who is who? Am I the only one?

Its 2020, there should be a much better way to go about the job application process already. Like software that keeps track of each application, and automatically creates a calendar for you, by interpreting the emails (AI stuff), and auto-responds based on the initial availability input you provide.

Say, I'm available Mon-Friday 12:00 noon to 5:00 pm, I applied to 100 job postings a month. It should read back the positive feedback, and auto-respond to the interviewer based on first-come-first-server basis with a specific time and date, and auto-populate a calendar for you. That would be so freaking neat!

I'm sure that will eventually happen, but it's taking forever lol.

In the meantime, how do you guys handle the process? I want to make this as less stressful as possible and avoid the complication altogether.

Is there a better way already and I'm not aware of?

https://redd.it/eodgdp
@r_devops

Ideal solution to separate config/data files from source code -Java

Hi everyone!

Hope you guys are doing great :)

I would like your suggestions regarding solving the following problem

I have a Git repository which contains Test Scripts written in Java as well as some config and test data files stored in Yaml and .properties file . The Test Scripts / code doesn’t that often . On the contrary the config and test data changes quite often hence there are way too many commits because of test data /config changes . Hence I am thinking of separating out the test data and config files from the test scripts / code and somehow find a way to access these files during runtime .

Also, if this helps ,I am using Maven .

What would be the best way to do this ? Also the another reason for this is to give a layman ability to access these test data files so that they can modify these files without worrying about the code .

I would prefer an open source solution . If I can host a service on my private network that would be the best as I would have to go through lot of formalities to use public cloud based solution .

Thanks in advance !
Adi

https://redd.it/eo8hve
@r_devops

In Terraform I'm getting a Malformed Terraform Policy Document Error on a seemingly perfect piece of code. What am i doing wrong?

MalformedPolicyDocument: Partition "aws" is not valid for resource "arn:aws:logs:\*:\*:\*

on modules/openshift/05-roles.tf line 26, in resource "aws\_iam\_policy" "openshift-policy-forward-logs":

26: resource "aws\_iam\_policy" "openshift-policy-forward-logs" {

Here is the offending code snippet:

// This policy allows an instance to forward logs to CloudWatch, and
// create the Log Stream or Log Group if it doesn't exist.
resource "aws_iam_policy" "openshift-policy-forward-logs" {
name = "openshift-instance-forward-logs"
path = "/"
description = "Allows an instance to forward logs to CloudWatch"

policy = <<EOF
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"logs:CreateLogGroup",
"logs:CreateLogStream",
"logs:PutLogEvents",
"logs:DescribeLogStreams"
],
"Resource": [
"arn:aws:logs:*:*:*"
]
}
]
}
EOF
}

https://redd.it/eo7pwj
@r_devops

How are you using RANCHER for cluster MGT?

We recently provisioned rancher on our cluster, came with great alert and monitoring tool, any other things you use rancher for?

&#x200B;

Kindly share.

https://redd.it/eo3rku
@r_devops

Canary over K8S - which tool and technique?

We have the need for Canary - get some more confidence before full-blown deployment. What are your recommended tools, techniques and practices? My current Hypothesis is using Istio and a static canary - simply check the next version over 1% of the traffic. Not ideal, but a lean way to go. Thoughts?

https://redd.it/eo29k2
@r_devops

How did/do you automate your Change Management process and evidence or controls for audit and stakeholders?

I'm curious about how your DevOps team went about meeting almost classic Change Management requirements and audit requirements. Im pretty stuck on where to start with this tbh.

https://redd.it/eo1lfg
@r_devops

Where and how are application secrets backed up?

SSH keys, passwords and passphrases, master passwords for passwords managers, API keys etc. Where and how are they all backed up? If my PC cached on fire right now, how would I decrypt data, or SSH to key secured servers?

1. Where do you backup? Only locally on external flash drives, or in the cloud?
2. How do you backup? Backing up to a cloud is meaningless if you do it in a way that exposes you to MITM.
3. In what format do you backup? Plain text, or encrypted (in which case - where do you backup the key you use to encrypt your secrets)? Should you ever bother with obfuscation, or is it ok to backup server private key as "server\_private.pem"?

https://redd.it/eo00y8
@r_devops