Відео

Hi Mark. I just have some questions outside of this video .I am planning to buy your book. Just want to know if the Azure DevOps is included in the book. Like automating pipelines and that kind of stuff and deploying to Azure using Azure DevOps. Also , I just found your channel and want to know in which order should I watch your channel. Should I directly start with Automating the automation playlist? For context : I have good knowledge in storage accounts and VMs and I have setup Terraform workflow using GitHub actions (not using Azure DevOps yet). Sorry for such a long question😅

why the fuck do you have so less views.😢 you deserve to be popular like hell bro😂.

Great video! Re: 7:24 in your video: I disagree about reusable modules using ">=" instead of pessimistic constraint "~>". The best practice here would be to have the reusable module specify "~> 1.51", which means this module can work with >1.51.x but less than 2.0.0. The code here makes the mistake of using ~> too granularly by specifying the patch also ("~>1.51.1"), which is too constraining. As a reusable module author, I can't guarantee compatibility with the next major provider version. Major versions introduce breaking changes, so it's risky to promise support beyond what I've tested. It's responsible to set upper bounds to prevent unexpected issues for users in the future.

Please report this to Cloud Strike and Microsoft if they don't or haven't already have figured it out perhaps that will help someone do so. My computer currently isn't utilizing my 365 apps correctly. They do not communicate effectively.

Good deal. I already pre-ordered

Putting the plan in the artifacts might not be the best idea depending on who has access to the pipelines. The plans tend to have secrets and values in plaintext and anyone could download it and leak these. A better approach we use is to add a task that uploads it to the storage account (the same one we use for backend) and download it before apply. You can use the --overwrite flag if you only care about the last plan and keep on overwriting stuff but a safer way is to put everything in a folder like PipelineArtifacts/ in the storage account and put the plan name to be Build.BuildId or something to be unique per run and for apply to find the corresponding plan from the previous stage. Also you can configure a lifecycle policy on the storage account with a path filter to delete old plans in the PipelineArtifacts/ path periodically.

If someone is watching this in the future, you do not need a separate pipeline. You could use the pre-defined. Build.Reason variable in the condition and disable the Apply stage if the value is "pr"

Man. Symantec Ghost mention. Childhood memory unlocked.

I'm not opposed to switching to YAML, but there are certain things I am trying to figure out, as it relates to a release with YAML - primarily, when I tag a branch in main, I want to update associated work items that have been added to main since our last prod deployment, and update their status to "Deployed Prod", or something of the like - that way we can easily have visibility of deployed features, and automate these tasks. Any idea how best I can move this direction with YAML?

Great👍👍

Congrats, this book will be available in India

Can't wait!

I will buy it

Awesome video!! immutable infrastructure is great, but I think it only works well with Linux apps. Do have examples with Windows Apps that are domain joined?

Congratulations again!

Hi Mark, Congratulations with your first book! Are you going to describe the best practices how to build a framework on top of TF modules? When user can use combination of TF modules and build the entire infrastructure for their solution providing the configuration files?

Thanks for this. I must be missing something, though. In order to access the settings at that link, you need to be signed in to the account. But, if 2FA is enabled (and unless you have an alternate verification method), you can't sign in without the working Authenticator account--it's a catch-22.

Thanks a lot for your efforts

Thanks!

yeah; i write terraform and GitHub actions workflows most of my day. I consider myself an Infrastructure Person, DevSecOps Engineer and a Software Engineer

Great presentation. A shame I wasn't there to say hi in person, but hope you enjoyed Sydney all the same

That's pretty cool. You could also do the something similar with VM SKUs e.g. validating the SKU is available in the region you want to use, validating the SKU meets CPU and Memory requirements, validating there is sufficient capacity within the region/availability zone to deploy the SKU.

Very cool module!

great video :)

Great module! Especially the availability zones flag

Awesome Video mate!

Great video. Learned lots. Thank you

And what if you didn't have backup already setup? Then what? ty

password batao

one can only imagine hacking with hitler.

How about using tf to setup an extension to join the domain

Thanks! Big help!

can I do the same using azure bicep?

can you share some tips for HashiCorp Ambassador for 2025 selection process because I'm targetting it.

Yup didn't do that

I love this, thanks man 🎉

Helped a friend with this process. They went through the same "freak out" experience you described when trying to get set up on a new device. QR codes? Where? How? All the instructions were worthless. Thank you for this video and the link for the sign in. Mission accomplished!

Hey! What an amazing video! It has been extremely helpful. I was wondering how can we accomplish the same but using https instead of ssh? I'm working on a project and keep on getting the TF401019 error 🥲 and I need to accomplish this by using a service principal

Hi Mark. Love your videos, in particular the code review ones. I have a video request, please can you do a video on the best ways to manage multiple environments with terraform (ie Dev/Test/Prod or Dev + NorthEurope/Dev + WestEurope etc). Theres a lot of differing opinions on it across the internet and it's terribly confusing. Some people seem to copy their whole codebase for each environment which makes the software engineer in me scream out in pain. Some people seem to use workspaces but then others say dont use workspaces unless your on Terraform Cloud. Others use a module that wraps all their other modules and then point an env root module to it. And then there's terragrunt which seems to introduce a whole other layer of complexity (Also do people actually split there projects into like 50 state files?!) From the videos of yours I've watched, it seems that having a folder per environment that points to a module that wraps the others is what most people are doing but I'm not 100% sure and would love a demo of how it works if this is your preffered method. As a current Pulumi user who has never used Terraform but wants to learn to make fair call on which to continue using in my current org the one thing holding me back from Terraform is this. In pulumi I can just add a new yaml file, add my subscriptionId to it and a few variables and be done. If you're not able to do a video on it I'd still love to hear your thoughts on a blog/comment or whatever is best for you and maybe some of the gotchas and how to resolve them (chicken and egg with shared infra etc) Thanks!

Thanks! This helped me a lot..

Useful Mark, but anu idea how I can parameterize script paths? In the shell provisioner I'm using: provisioner "shell" { execute_command = local.execute_command scripts = var.shell_scripts } ... and the var file looks like: shell_scripts = [ "./scripts/configure-repos.sh", "./scripts/install-httpd.sh" ] This works great when when I run packer from "." because the path is relative. If I try to run it from somewhere else higher up the tree I get "file not found" errors. With the file provisioner I can use: provisioner "file" { source = "${abspath(path.root)}/files/repo-config/" destination = "/tmp/install" } This doesn't seem to work with the scripts block though. Looking for inspiration!

As a junior devops engineer I shared your channel with my seniors they are shocked that you only have 4K subscribers

Great video! Very informative and to the point. Keep up the good work!

Would love a updated video that does this with the latest Unifi UniFi Network Application 8.1.127. I can get both Azure and Unfi to show the tunnel connected but cannot get the VM to ping..

I've been playing with AOI for a bit now and it is pretty interesting. Forcing myself to learn NodeJS and grow my Python skills.

Amazing video!!! This helped me to understand a lot about the azure networking more than any documentation

With a couple of our projects we started to provision small vnets. Only fitting one or two subnets. What do you think. Now I think of it it would make sense that vnet becomes "dumb" and just becomes a group of subnets for routing purposes.

This is a great great explanation about the core of resilience in Microsoft Azure. Thank you!