VMware Carvel Community Meetings, 14 Nov 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Getting Started Tutorials: ytt Part 2

Description

Getting Started Tutorials: ytt Part 2
Are you new to ytt and would like to know where to start? Varsha Munishwar (@vmunishwar), a ytt maintainer, continues to walk you through the process of getting started with ytt. Make sure to check out Part 1 here: https://youtu.be/DvApsPy0IrI

▬▬▬▬▬▬ T I M E S T A M P S ⏰ ▬▬▬▬▬▬
0:01 – Summary of Part 1
0:50 – Sample deployment configuration
1:16 – Labels and selectors
1:52 – Deployment on Kubernetes cluster
2:22 – Problem demo
3:50 – Solution using a ytt feature
4:26 – Implementation details
5:46 – Better world with more edits :)

A

Hello, everyone welcome to the part of the getting started with YTD tutorial Series. In the first part, we saw introduction to YTD what kind of problems ytt helps to solve. How uh templating, with variety, helps reduce boilerplate, and we saw ITT in action with few examples on playground and how to run ytt on CLI. We saw how to extract variables, use functions, to extract common code using conditionals and for Loops. Now in this tutorial, let's look into a slightly more involved scenario, where ytt can help us avoid kind of a problem.

A

Have you ever made the same edits in multiple places in your kubernetes configuration manually? As you might be aware, this can be error prone and can lead to misconfigurations here. Ytt can help us avoid this kind of problem. Let's say from an example, so here is a manifest file ready to deploy to our kubernetes cluster. It has two objects, deployment and Service. As you might know, the deployment contains um a section that is a template for all the parts that are going to be created here.

A

We see there are three parts and each of these parts are going to have this set of labels.

A

uh The deployment also has a selector and it is used for relocating any parts that it should manage once the parts are up and running uh the corresponding service. It's used to direct traffic to those parts. It does so with a set of selectors that matches the labels uh or the subset of labels on those parts.

A

So, in short, the labels in the selector of the service must match sums upset of labels on the parts that were generated ultimately from the deployment. So this is important to keep these labels in sync, uh let's see how this deployment works. So for that, let me deploy this to a kubernetes cluster.

A

So my deployment is subsided here. Let me open the browser and let's see we get a response from server now. Let me go ahead and make some changes in my deployment configuration.

A

So let's say: if I want to make some change, in instance, name here in service I make made the change to XYZ, but now I forgot to make same changes in the deployment. For example, let's see what happens when I deploy this.

A

So after the deployment I can see, the deployment is subsided, I can see the status and when I get endpoints, I can see that there are no end points generated and let's look at the browser window refresh and the site cannot be reached. So obviously the deployment is successful, but the service doesn't know which parts it should route the traffic to there would be no IP addresses or output. So um that's why, when we look at the application we get nothing. Now, let's make the change in in our deployment configuration to match all the labels.

A

Once these are matched and I'll redeploy.

A

So now, after matching the labels, I redeployed and now, let's look at the browser and I can get the response back. So, let's get back to the deployment configuration notice that I have to remember to make the changes in all three places, and so the problem is that there must be some way that guarantees that all these changes stay in sync- and we can do this in YTD by extracting the yaml fragment into a function.

A

So ytt provides a way to extract the entire chunk of yaml and return it from a function. This function is called yaml fragment function and this will eliminate the risk of misleveling certain keys and values at multiple places. So how do we write the ml fragment function?

A

So we start with hash at Define labels and n, and we just copy these labels.

A

My students here and then call them from all the places where we at these labels.

A

So this is how we write the yaml fragment functions. These are yaml structures that are wrapped in starlock function, definition, and these are different than starlock functions. That we saw in first part notice that there is no return statement over here and everything inside the function will be the return value these are used when you need to return a nestodium structure or key value pairs.

A

Now, let's see the output from ytt and then 3D Plus, the tables are getting inserted at all the places where we call the labels from. Let me read, apply the application, so I redeployed, the application and, let's see on the browser we get the response back.

A

So let's make one more change in the deployment EML configuration now, if I want to add some more labels. To this, let's say: I will add version to this notice that I just have to update label at one place, and we can see that the labels are updated at all places. So this is, our configurations can be in sync with yaml fragment functions, so try using these functions and let us know if you have any questions please reach out to us through our kubernetes slack Channel. Thank you and see you in the next one.