Red Hat OpenShift Red Hat OpenShift 4, 20 Apr 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Getting Started with ArgoCD on OpenShift 4

Description

In this video, Christian Hernandez shows you how to get started with ArgoCD on OpenShift 4 as he deploys an application across two clusters.

To learn more, visit OpenShift.com

A

Hi, my name is squisher Hernandez from the cloud platforms business unit over at Red Hat in this video I'm, going to show you how to get started with agro CD on open ship for argo. Cd is a get-ups tool that can be used to sync your application deployments to one or many clusters. Argo CD can be used standalone or as part of your CI CD pipeline. Let's get started.

A

In order to get started, I need to log into my open ship cluster.

A

Here, I'm logging in as cube admin as I need to be an administrator in order to install an operator note that I logged into a cluster called go dot, o CP that Red Hat workshops that IO.

A

In order to install the operator, we need a namespace in order to house all the components here, I'm going to create a project called agro, CD.

A

Once the project is created, I can go over the operator hub and search for agro, CD and click on install note that I'm installing it in the previously created argo, CD, namespace.

A

The operator will go from pending to installing to succeeded.

A

Once the operator is installed, we can use the operator to deploy an instance of our NGO CD.

A

My configuration has some customizations like deploying in the same namespace as the operator creating a route hooking it up to the OpenShift oo, auth and mapping my local admin to the Argo CD admin.

A

Once I could create it'll go into a pending state and I'll start the install going over to my git repo I have a few manifests that I want Argos CD to sink into my cluster.

A

First I have a deployment that will deploy an application that will draw a blue box. For me.

A

I also have a namespace configuration which will create a namespace for me.

A

Next, here out specification that has the URL for this specific cluster.

A

And then it's backing service, the service connects to my back-end application on port 8080.

A

Also note that I'm running on a branch called video.

A

Going over to work loads and pods, you can see now that all my pods are up and running and once I verify that all my pods are running. I can go to networking routes to display my route once I click. My route I'll accept that the self signed certificate.

A

Note that I have a login with OpenShift since I specified in that. In my configuration.

A

Once logged into our go, I am ready to use the CLI in order to add a cluster to use. The CLI I will have to log in I login by using Argos edy login with a SSO option using the SSO I need to accept the self signed certificate and copy the URL.

A

Once I copy, the URL I can open up a browser and paste the URL in once I paste it in and see the green banner I am authenticated.

A

Switching back over to the CLI I can now run Argos CD cluster ad to see my clusters I, don't see any clusters because I need to log in with openshift from the CLI. So let's do that.

A

Now, when everyone Argo CD cluster ad I, can see my NGO cluster.

A

Now I can run agro CD cluster ad in the name of the cluster.

A

Once added I can see it in Argo CD cluster list.

A

So she back over the UI I can now deploy an application.

A

My application creates a namespace called bgd, which you note is not created yet in my cluster, so this deployed when I click, new app I will give it a name of bgd and put it in the default project. This is in relation to our NGO CD, not open. Shipped I will set a manual sync policy for the repository URL I need to give it the get URL.

A

And then specify the video branch.

A

I'm running all my manifests inside a directory called bgd and I want to deploy it in the cluster. We just added the namespace bgd does not exist, but our go CD will create it. For me, once I click create my application is created. An Argo CD will start to look for differences, note that it says I'm out of sync and in fact everything is missing.

A

In order to gets my application to my cluster I'll use, Argo CD, sync.

A

Argo CD will then sync my upstream manifests to my cluster in this case, creating everything from scratch. It created the namespace, my deployment, which is running a pod and the route, and indeed my application is running with the blue box.

A

It's introduced to change and see how our go CD handles it.

A

I would tell my application that I want a green box, now note that I'm making an infrastructure change, not a change to the application itself.

A

Once this is patched our go. Cd right away recognizes that something is out of sync.

A

My application is healthy, but is out of sync going over. Indeed, I am running a green box and the change hasn't been made to remedy this I will click sync and sync all.

A

What's the succinct I will wait for my roll out of my application deployment and once my application deployment is rolled out, I can see that everything is back to normal.

A

And I do have a blue box.

A

So how does Argo CD add other configurations that it's not that it doesn't know about notice here that I have a prune option here for my sink. So let's see this in action. If I do an OC get config map in my namespace you'll see that I have no config map set I have a configuration here that sets a sample. Config map and I will go ahead and create it in my namespace.

A

It is not created in my bgd namespace argo CD C, said I'm out of sync.

A

Now, when I click sync and choose a prune option, Argo CD will not only sync to my upstream git repository, but I'll also delete any superfluous configurations found in the namespace.

A

Argo CD can also be used to recover from catastrophic failure. For instance, if I accidentally delete my name space of a production work of book workload.

A

I'm now in a terminating state on my namespace and if I go to Argo, CD you'll see that it's out of sync.

A

In fact, my manifests are missing for my namespace.

A

Once I click sync and using proving for good measure, you'll see that I can quickly recover from a catastrophic failure.

A

My name space is back along with my workload and my route is there as well.

A

The only way to delete an application with article CD is deleting it within our go.

A

It goes into a deleting space and to make sure everything is removed for my cluster.

A

Now, if I have multiple clusters, how do I deploy it from a single repo for this we'll use customized in this directory, we have a base directory with all my base configurations. So, for instance, I have a deployment that will go to both clusters. I have a namespace I'll go with both clusters.

A

The service as well will go to both clusters and even the route, but the route will change, depending on which cluster I'm in so I'll leave the host field empty.

A

Going back to my overlays you'll, see that in my first cluster I want to override whatever my deployment says and set it to blue and whatever my route says and set it to the proper route. In this case, it'll go from empty to the route, and my second cluster I want to run green over writing that deployment and then in my route, having a route specific for that cluster.

A

And note that all this is within the video branch.

A

So going over to my second cluster note that it's running cluster to dot o c p dot, Red Hat workshops that IO.

A

Switching over to my CLI I need to add this cluster and again just like before I. Don't see it in my list because I'm not logged in so let's login with the CLI.

A

Once I logged in with the CLI, my cluster, two I can now see it in the list. So I do Argo CD cluster ad and indeed I am now seeing it in my configuration so now running, Argo CD cluster ad and specifying the cluster to configuration.

A

It'll add a cargo CD once I do Argo CD cluster list you'll see that now I'm controlling to clusters cluster two and go.

A

So creating a new application, I'll call this VG dk1 in the project name, space defaults again within our NGO CD.

A

Using the same URL get URL and the branch this time I will do the path be GDK overlays cluster, one.

A

And I'll deploy it to the cluster, go in the name. Space be Gd.

A

Not creating another application.

A

I'll specify vgk to.

A

And I will use the exact same repo and the exact same branch, but this time I'll choose the directory cluster and in the cluster cluster in the namespace bgd.

A

Now I can sync the applications.

A

Sinking PG, k, 1 and then sinking b, GD, k, 2.

A

My first cluster you'll see that bheegi bheegi D is deployed with the route and the blue box once my second cluster is synced. I can go and see that it created the project, created the pointman and created the custom route.

A

I have a greenbox, I hope you enjoyed this video and invite you to try our go CD in your get ops workflow on open ship, for. Thank you.