GitLab Speed Runs, 18 Sep 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GitLab Runner Speed Run - Auto provision VM with Runner on GCP

Description

In this video we review the new capability that allows users to use Docker and a simple command to auto provision a virtual machine on Google Compute Engine with GitLab Runner.

https://gitlab.com/gitlab-org/ci-cd/distribution/runner/self-hosted/-/tree/master

A

Hey everyone darren smith here, product manager for gitlab runner, for today's speed run, we'll cover the use of our recently published docker image that automates the provisioning of a virtual machine with the runner and that virtual machine will have the github runner and that one will be automatically pre-registered against your time gitlab project. This is picking up on the speed run. From last time, we were previewing a bit of that capability that we had been building out of spot about in 1304 release.

A

The target cloud platform in this case is google compute engine for folks that are interested in looking at the details of this particular work. um Here is the link to the project repo. So you can take a look at the details there again. This is part of our broader strategy of simplifying or making cic super easy for customers by simplifying the or automating the installation of git, lab runners to target environments and then long term simplifying the management capabilities, simplifying the management of run as a scale.

A

So today, we'll just be covering um one of the first features of that evolution in our roadmap. Is this idea of having a docker image where everything is pre-packaged and then by simply using a simple docker command?

A

You will auto private provision, a virtual machine that will have gitlab runner and that one will be pre-registered for you against a target um gift lab project so in the particular project reaper here you'll see the readme file, and so today we're just simply we're covering this very simple double command and what it's going to do is going to pull down if it hasn't done so already.

A

If this is the first time that you're running this command, it'll pull on the image from the gitlab registry and then I'll run the commands using these parameters, and only it takes three parameters: your registration token, which is a registration to confirm your time to get that project, your google compute project id and your google compute zone, so I'm just gonna bring this over into a command prompt and on the command.

A

I already have everything set up on my end and again, the three parameters are your registration token, a project id and you click on compute zone hit, enter.

A

Give it a couple seconds and if again, this is the first time you're running this, it will have um you would it would need to download the image from a registry, in my case, it's already done so and so now it's actually in the script, and so now we will prompt you to copy this url into your browser.

A

I'm going to choose this account hit, allow and now I've gotten the code that I would need back in the script, the verification code I paste it in here- hit enter and now what's happening in the background. Is that we're using terraform the ottoman greatest virtual machines? They will have to give that runner they'll be registered against your project. I've already done this before so I wouldn't go forward and do it here.

A

But in your case, once you complete the um if you hit enter you'll, be able to go into google compute, your your in your google compute um compute account you'll, see the virtual machine and you'll get that project. You'll notice that a runner has been registered against your project. um So I'm looking for looking forward to continuous feedback from your users in our community around this type of this type of capabilities, as we look to again simplify um the automation, the installation of runners automate that process for you and so on.

A

So in future speed runs we'll continue. um Looking at some of the additional capabilities that we're building out in this space, so look forward to sharing with you, then cheers.