Suggestions for a Learning Platform?

My employer is kindly paying for me to join an e-learning platform.

What’s an e-learning platform that’s really excellent? Right now it’s between Linux Academy and ACloudGuru, but I’m open to other suggestions.

https://redd.it/e8vafc
@r_devops

OpenLegacy (or alt) review ?

Hi guys,

Recently, some of the managers I work for went crazy because they heard about OpenLegacy, a solution that can "easily and automatically create open APIs from your main (legacy) systems, such as AS / 400, iSeries, etc. "

I do not really understand how we can trust something like this?

When I check each demo, the use cases are so simple and do not take into account more than 30 years of huge Cobol programs, multiple databases and half-developed Web services.

Has anyone here ever experienced this kind of solution? I am not looking for a sales pitch, which is why I am interested in what you think about it.

https://redd.it/e8xjqk
@r_devops

Amazon Simple Email Service (SES) Are Cron Jobs all you need to manage the rate of newsletter email issuing?

Good day, everyone.

I'm putting together automated notifications about new content on our website and we are looking to automate the process. I have a database with the table of new content and a table of subscribers and I essentially need to send them all an email when new content is uploaded. From what I understand, there are limits to how many emails you can send out using Amazons SES per second. How do you handle email sending? Do you Simpy set up a Cron job to trigger a PHP page every minute and handle sending emails using something like phpmailer?.

https://redd.it/e8tuvi
@r_devops

Can you give advice/feedback on part 2 of our postmortems piece?

Hi guys! Steve McGhee worked with us on creating a second part to our postmortems article. It would be awesome if you took a look and maybe commented on what you liked and found helpful vs what you would like explained more in the future.

[https://www.blameless.com/building-reliability-through-culture-sre-steve-mcghee/](https://www.blameless.com/building-reliability-through-culture-sre-steve-mcghee/)

https://redd.it/e8ts6c
@r_devops

Why run Jenkins master in a Kubernetes cluster?

Hi,

TL;DR: kubernetes novice here, don't see why should I have the master in a cluster.

I'm looking into migrating an existing CI infrastructure (just CI, no CD), which is a fairly classic AWS EC2 based infrastructure, into Google Cloud Kubernetes Engine.

Right now Jenkins master is running on a VM in AWS, slaves are managed by the EC2 Cloud plugin, provisioned with Chef, custom AMIs built with packer, terraform etc. Each node has docker installed and Jenkins pipelines are using the docker plugin to build and execute stuff for different projects inside the containers.

It makes perfect sense for me to have the agents as containers in a Kubernetes cluster. It makes perfect sense to containerise my Jenkins master and run it as docker container, for easier upgrades, etc.

However I cannot understand why it is better to have the Jenkins master inside the Kubernetes cluster as well, it seems an unnecessary complication. I'm finding a lot of documentation on how to run a Kubernetes cluster with Jenkins, but none seems to explain why this is a good idea.

Consider the following picture:

[https://imgur.com/a/0GW8qW1](https://imgur.com/a/0GW8qW1)

vs having the jenkins master container as part of the kubernetes cluster itself.

I'm not planning on having multiple masters or do any fancy load balancing. I still can't find a good reason why everyone seems to run the master as part of kubernetes. I know there are helm apps which make the deploy of it a click install, but to me it seems I'm losing some flexibility here. Having only the agents in the cluster makes sense, since I can destroy and recreate the cluster. Maybe having 2 clusters is a better idea - one for master and one for the agents.

Anyway, sorry for the long post, but I'd really like to understand what are advantages/disadvantages of having the jenkins master container in a kubernetes cluster vs running it as a container in a classic VM.

Thanks!

https://redd.it/e8scng
@r_devops

Adding Serverspec test to Jenkins pipeline

Question from a newbie in the area.

I'm trying to add a Serverspec step in a Jenkins pipeline. The aim is to use the automated test to check if a said server contains the appropriate AMI (AMI being build with Packer) I have no idea where to start with this, and I mean from the very start, i.e: where does it go in Jenkins, folder structure, how the testing could work. Has anyone done this before or have a good reference for it.

https://redd.it/e8pqgi
@r_devops

Code structure to build a single image and apply multiple environments.

Hi, my source code and manifest are stored in the same repository.

Every time the developer commits new code, CI bake and deploys an image to Kubernetes. Target Namespace and Cluster selected based on the source code's branch. ( e.g. branch test/env1-> deploy to namespace env1 @ testbed ).

This strategy is easy and simple but can't archive the immutable infrastructure.. since CI builds an image for development and production separately.

I thinking about having a single image apply to all environments. To archive that I think I need to separate code from the manifest.

That will make the code structure make complicated. So I would like to listen to your advice.

Thank you.

https://redd.it/e8oios
@r_devops

First DevOps project -> Start with using AWS?

TLDR:

Team asked me to set up something that can serve as a dev environment and showcase current status of projects to simplify collaborating with clients. Send halp! Team leaning towards using AWS.

​

Full:

I usually do front-end dev work and am currently working for a consultancy company. Company is cool but some things are very old school and take a lot of meetings to change. One thing my team wants is to have some sort of portfolio-like environment to hold our current projects where we can work locally, push up our changes, and have some sort of pipeline that updates this portfolio so that we can more easily see the most current version of whatever we're working on as well as have an easy way to have our clients see where we are at and get feedback at all stages of development. My team figures it will be quicker to just build our own and, since none of us are experts, just get into AWS because we're gluttons for punishment, it is industry standard, they have a million tools and products, and it is something we could keep building on to try new ideas in the long term.

I have no idea what I'm doing but I'm happy to learn whatever I need to get this done, I just need:

1.) Feedback if AWS is the right choice for us or did we just pick that because we're too clueless to know that there is something better for the long-haul.

2.) What products on AWS I should be focusing on to get this done. So far I think just a simple S3 account to hold static projects and we could have some sort of login that can be created for clients to only see their project. I don't know if we'll have more complicated projects in the future (like ones that require server traffic) and not sure how hard it is to migrate stuff or add that kind of functionality later. I'd rather put in the extra work at the beginning to future proof this some (and because if I'm going to do this I might as well do this right) but also do not want to just waste company time and money on something that we will never use.

Any pointers would be greatly appreciated!

Thanks so much in advance :D

https://redd.it/e9pame
@r_devops

Suggestion request: a kubernetes-like service discovery without the cluster

I am hoping I can get some interesting ideas from people with more experience and knowledge in this domain. What I basically want to achieve is having multiple nodes, on each node having multiple services names a, b, c in any combination (each node can have any combination of a, b, c instances). I want each service to be able to communicate with other services via dns with the service names, so when a makes a request to [https://b/api/resource](https://b/api/resource), it will go to a container with the b service on any node. I am thinking that some kind of mesh proxy sitting locally that will route the traffic as needed would be the best to go, but I can't seem to find anything that can run outside of a kubernetes cluster.

I looked into envoy proxy but it seems like a more cloud native nginx alternative and other solutions seem to target specifically the kubernetes market. Having a kubernetes cluster is not a possibility for me as it has quite some overhead with the control plane and is too much overkill just for what I am planning to do (probably 3 nodes with some replicas).

Another solution I thought about is having a nginx in front on each node that knows about the local services and can route based on host, a "centralized" load balancer where all service name dns will point to and when a services makes a request to [https://b/api/resource](https://b/api/resource), it will be routed by the main load balancer to a nginx that has at least a container with **b** running

What other (open source) solutions are there that be more appropriate for these kinds of requirements?

https://redd.it/e9npsg
@r_devops

When just starting out as a startup, what's the most affordable storage service do you recommend when developing apps?

I am working on a music and movie streaming service for my country but I don't know what would be the most affordable, scalable and efficient storage service for my platform. At first I was thinking either firestore or AWS but companies like Spotify use like Hadoop so I would like to hear your own opinions!

https://redd.it/e9sa9c
@r_devops

How to actually get good at basic DevOps if i'm not already hired

I'm a full stack developer. and i recently was interested in getting into the devops world. learned about docker, clustering, kubernetes but only the very basics, like tutorials level basics. How should i actually get good in Devops if i'm not working for a company that requires me to solve such problems? as a developer, i would just build stuff and scale it since theyre just apps.

https://redd.it/e9nl05
@r_devops

Kemp 360 Central

Hey - we've build Kemp 360 central to allow admins and devops to see traffic across multiple load balancer products - and its 100% free. We think the multicloud might make it more difficult to understand how many users are hitting your applications in one or more locations and across more and more devices.

https://redd.it/e9ppv8
@r_devops

CI testing Linux/Windows integration

I have a application that interfaces with a Microsoft SQL Server database. I can test it on Appveyor under a complete Windows environment. However, I would also like to try testing the code on Linux, while some SQL Server is also out there on a Windows test environment. I know Appveyor does have Linux testing containers. Does anyone know if there is a way to sync two test containers up, one Linux, one Windows to try this? If not, any suggestions on workarounds?

https://redd.it/e9vmrp
@r_devops

Spot instance and custom domain

Hi, apologies if this is not the right sub.

I have a custom domain and currently and EC2 instance server where the application is hosted. Long story short, it's only used a few times a month but I am still paying for the "unused time".

I decided to change it to a spot request instance but I am confused with something: can I point my domain to this instance ( I have an Elastic IP associated with it)? What happens when the instance is not "allocated"?


Appreciate any help. Many thanks.

https://redd.it/e9o6x2
@r_devops

Dockerfile linters wanted!

Looking for some ideas for Dockerfile syntax linters and how they integrate in your CI flow. Do you use one? Are you happy with it or do you see shortcomings in using it?

I've found hadolint ( [https://github.com/hadolint/hadolint](https://github.com/hadolint/hadolint) ) and am trying it out locally at the moment. Curious if you have experience with this one or with other Dockerfile linters.

Where do you integrate it into your CI? Local git-hook before pushing to repository, remote git-hook on PR or build stage in CI? Why?

https://redd.it/e9xwj1
@r_devops

Docker and Sonatype got acquired, what next?

Recently, Docker got acquired by Mirantis and Sonatype Nexus got acquired by a private equity firm. The future of these firms may not be the same and we cannot predict anything since the ownership has changed.

Looks like Docker has gone into the good hands, Mirantis helps companies run Kubernetes and OpenStack with ease and the addition of Docker surely feels like an enhancement in the cloud-native field. But it also concludes that Docker won’t be Docker anymore, like all acquisitions, strategic business objectives will change and this might also affect some microservices developers.

But when it comes to Sonatype Nexus, an equity firm buying it makes no sense to me as a developer.

What can be the future of container registries?

https://redd.it/e9j8vk
@r_devops

DB in containers

What are the pros and cons of running a DB inside a container ? Is it recommended for a production system ?

https://redd.it/e9zcx5
@r_devops

I have one month to prep for a mid/senior level Devops/full stack interview at Capital One

Has anyone gone through this? What was their focus? I am told it will be a full day of 5-7 interviews.

I have about 5 years experience, 2.5 in mobile app development, 2.5 in the systems engineering/devops space. I feel like I have a broad surface level knowledge of almost everything, but I don't know one thing particularly well.

My current goal is to review that Devops learning flowchart that's running around and fill in my gaps on that chart. I will probably practice a hackerrank problem every other day, as I know there will probably be at least one whiteboard session. I will probably review and buff up my AWS and infrastructure knowledge, as I can't currently do a VPC from memory.

What are on-site interviews for Devops like? Will I be asked to bring up a LAMP stack?

Really any general tips for an interview day like this would be appreciated.

​

Note to Mods: this is a repost, my other post was hidden for some reason

https://redd.it/e9zalm
@r_devops

Why Apps like Uber or Tinder need a lot of money to start? is the server cost high?

**Are server costs high? I'm asking to dev-ops people here because you are familiar with server usage and costs in general. I'm finishing to code my own Tinder app (it's for a specific niche, but maybe adoption grows a lot) because I don't have an initial investment like the successful known companies had maybe I'll fail. I'm not sure if I'm going to get money from my users so maybe I won't be able to pay for the server cost with my salary. I'm not looking to make profit with my project, I'm fine if I can pay for the server. I'm going to host it on AWS by the way.**

**I'm not looking for "the perfect answer" just what your dev-ops experience can say about the costs of hosting in general.**

https://redd.it/e9z8yp
@r_devops

Any advice on what's the best Scalable and Affordable database for a startup?

So I am developing an app like Spotify or Tidal, I am thinking about using either AWS/Digital Ocean for Blob storage (For music mp3 files) but would like to know how what would be the best database that's affordable, Scalable and allows data replication (i.e. Move all the data from one database into another). I welcome all opinions and suggestions.

https://redd.it/e9viwj
@r_devops