WEBVTT 00:00.110 --> 00:05.150 Okay, so now that we've pushed our images up to Gcloud, we're going to go ahead and start by running 00:05.150 --> 00:09.590 our own local Kubernetes cluster and have our images running locally. 00:09.620 --> 00:14.270 Now I'm using Docker desktop as my Kubernetes provider here. 00:14.270 --> 00:19.970 So if you use Docker desktop, you should be able to go into the settings and see a Kubernetes field 00:19.970 --> 00:26.240 here where you can enable a Kubernetes cluster on Docker desktop to make it very easy to get started 00:26.240 --> 00:28.640 with Kubernetes back in the command line. 00:28.670 --> 00:35.840 You can see if I run Kubectl get namespaces, I have the default namespaces for Kubernetes cluster and 00:35.840 --> 00:41.000 if I run Kubectl get pods, I have nothing running in my default namespace right now. 00:41.000 --> 00:44.840 So let's go ahead and change that by starting to create our first deployment. 00:44.870 --> 00:51.110 We're going to be using Helm as our dependency manager for Kubernetes so that we can easily deploy our 00:51.110 --> 00:56.030 application anywhere locally or on Gcloud very easily using helm. 00:56.030 --> 01:03.960 If you don't have helm already installed, simply go to helm sh slash docs intro install here where 01:03.960 --> 01:10.080 you can get instructions for how to install helm easily on any system, including from a script or through 01:10.080 --> 01:12.930 package managers like Homebrew or Chocolaty. 01:12.960 --> 01:16.770 It's very easy to install helm so go ahead and make sure you have this running. 01:16.770 --> 01:22.440 So to start, let's go ahead and create a new folder in our project called K80s, which is where all 01:22.440 --> 01:25.560 of our Kubernetes deployments and manifests will live. 01:25.680 --> 01:26.790 Now in our terminal. 01:26.820 --> 01:36.270 I'll CD into that folder and then I'll run helm Create Sleeper here to create a new helm chart called 01:36.270 --> 01:37.080 Sleeper. 01:37.080 --> 01:42.540 And what this is going to do is it's going to create some starter files for Sleeper here and give us 01:42.540 --> 01:48.900 some starting templates as well as an all important chart YAML file which defines our application and 01:48.900 --> 01:51.390 all the dependencies for this chart. 01:51.390 --> 01:53.400 So we're going to go ahead and clean things up here. 01:53.400 --> 01:58.560 I'll go ahead and remove everything in the templates folder that was generated. 01:58.560 --> 01:59.700 We don't need any of this. 01:59.730 --> 02:04.890 We can keep the helm ignore here, which will make sure we don't commit any of these unneeded files. 02:04.890 --> 02:08.640 And then in the values file, I'll go ahead and delete everything as well. 02:08.640 --> 02:13.080 So now we just have a completely empty chart called Sleeper here. 02:13.110 --> 02:13.440 Okay. 02:13.440 --> 02:18.630 So the next thing we want to do is actually create deployments for each of our services. 02:18.630 --> 02:24.420 Deployments are what's going to define the manifest for each of our microservices and make sure we always 02:24.420 --> 02:27.390 have a pod running for each one of these apps. 02:27.390 --> 02:29.340 So let's go ahead and start creating one. 02:29.340 --> 02:36.000 We're going to go ahead and use the Kubectl command here and run create deployment and we'll start off 02:36.000 --> 02:40.890 with reservations here and then I'll provide the image name here. 02:40.890 --> 02:46.290 And of course, we know we need to get the image name from our Gcloud repository. 02:46.290 --> 02:52.290 So we'll go ahead and copy the full path to the image here and paste in the image and of course make 02:52.290 --> 02:55.160 sure we add slash production to the end here. 02:55.170 --> 03:01.230 Next, we're going to set this flag called dry run equal to client here to make sure that we don't actually 03:01.230 --> 03:06.630 execute this, but we want to actually output this to YAML so that we can source control and commit 03:06.630 --> 03:12.030 it as part of our helm chart and make sure it's always going to be run when we do helm install. 03:12.030 --> 03:17.790 So let's go ahead and add dry run equal client and then we'll add output equal to YAML here and pipe 03:17.790 --> 03:21.240 this to a file called deployment dot yaml. 03:21.390 --> 03:27.660 So now if we go back to our folder, we have this generated yaml here which is the manifest for the 03:27.660 --> 03:33.240 reservations deployment and you can see we have everything we need out of the box to create this deployment. 03:33.240 --> 03:34.470 Go ahead and clean this up a bit. 03:34.470 --> 03:40.140 I'll get rid of the creation timestamp up here, then I'll get rid of the strategy section. 03:40.140 --> 03:46.350 We can get rid of the creation timestamp resources status and then I'll change the name of the container 03:46.350 --> 03:48.240 here to be reservations. 03:48.270 --> 03:54.330 Next, let's go ahead and in the sleeper directory we need to move this deployment manifest into our 03:54.330 --> 03:58.650 templates folder, which is where everything is going to be actually deployed from. 03:58.650 --> 04:04.830 So let's create a new directory in the templates folder here called Reservations, and then we'll move 04:04.830 --> 04:08.310 the Deployment.yaml into that reservations directory. 04:08.310 --> 04:09.630 So let's go ahead and give this a try. 04:09.630 --> 04:16.230 Now if we CD into Sleeper and then we can run helm, install Sleeper and then provide the path to the 04:16.230 --> 04:16.830 helm chart. 04:16.830 --> 04:22.080 In this case, it's in the current directory, we can see our deployment was successful and now if we 04:22.080 --> 04:25.440 do kubectl get pods, we can see our reservations. 04:25.440 --> 04:27.120 Pod is running, which is awesome. 04:27.120 --> 04:32.340 However, we have an image pull backoff issue here, which means we're not able to correctly pull the 04:32.340 --> 04:33.030 image. 04:33.030 --> 04:39.990 So if we do describe pod and enter it, you can see we're getting denied access on our image. 04:39.990 --> 04:42.150 So let's go ahead and take care of this next. 04:42.150 --> 04:47.520 So remember that out of the box that our artifact registry is completely private, which means that 04:47.520 --> 04:50.910 people can't pull our images without correct access. 04:50.910 --> 04:57.900 And that means even on our local Kubernetes cluster, we need a way to configure access to these images, 04:57.900 --> 04:59.730 just like we did for local Docker. 04:59.950 --> 05:01.240 We were pushing and pulling. 05:01.240 --> 05:02.920 So let's go ahead and do just that. 05:02.920 --> 05:10.450 We'll go into the G Cloud console and go to APIs and services, then go ahead and click on credentials, 05:10.600 --> 05:16.300 and then we're going to click Create Credentials and click on service account. 05:16.420 --> 05:22.670 I'll go ahead and call this artifact image pull, then click create and continue. 05:22.690 --> 05:29.500 Then for roles here, click on roles and scroll down until we find the artifact registry. 05:29.500 --> 05:35.860 And then we're going to click on Artifact Registry Reader so that we can read items from our artifact 05:35.860 --> 05:36.550 registry. 05:36.550 --> 05:37.970 Then click on Continue. 05:37.990 --> 05:40.510 Then we'll go ahead and finish by clicking Done. 05:40.510 --> 05:45.940 So now that we've created the service account, go ahead and click on it and then we're going to click 05:45.940 --> 05:50.410 on Keys here and add a new key and click Create new Key. 05:50.440 --> 05:51.700 This will be JSON. 05:51.700 --> 05:56.530 So go ahead and click Create and then we'll click Allow to download this. 05:56.740 --> 06:01.380 So make sure you download this JSON and save it somewhere on your machine where you can find it. 06:01.400 --> 06:07.160 Okay, so now we need to tell Kubernetes to use our JSON key file when pulling from GCR by creating 06:07.160 --> 06:08.300 a new secret. 06:08.300 --> 06:17.540 So we'll call kubectl, create secret of type Docker registry and we'll call this GCR JSON key and provide 06:17.540 --> 06:20.180 a flag here called Docker Server. 06:20.180 --> 06:26.960 And this is going to be equal to the domain of where our repository is living, which we can also get 06:26.960 --> 06:36.870 by clicking on a repository, setup instructions and copying this here the US East four in my case.package.dev. 06:36.890 --> 06:38.600 Copy this and paste this in. 06:38.630 --> 06:46.400 Then we'll provide a Docker username and we'll set this equal to underscore JSON key. 06:46.970 --> 06:50.660 And then importantly the Docker password. 06:50.690 --> 06:57.440 This is going to be equal to open up quotes and inside of here we'll do dollar sign and then open up 06:57.440 --> 07:04.700 parentheses and in the parentheses we're going to run a command cat and then enter the path to the JSON 07:04.700 --> 07:05.630 key file. 07:05.630 --> 07:11.090 In my case, I've put it in this exact directory here, so I will paste the path here. 07:11.090 --> 07:16.510 Finally, we'll provide a Docker email here and I'll set this equal to our email that we set up Sleeper 07:16.520 --> 07:19.190 Nest app at gmail.com. 07:19.400 --> 07:24.050 And then you can see our secret JCR JSON key was created. 07:24.140 --> 07:28.880 Finally, we need to add the secret to our default service account so that it's actually going to be 07:28.880 --> 07:30.320 used to do this. 07:30.350 --> 07:41.060 We can run Kubectl patch service account default and add dash P, open up single quotes here and inside 07:41.060 --> 07:49.760 of here we're going to open up a new object and we're going to use open up quotes and put in image pull 07:49.790 --> 07:51.140 secrets here. 07:52.200 --> 07:55.110 And set this equal to an array. 07:55.380 --> 08:01.230 And inside of that array we'll have another object with a name here. 08:01.620 --> 08:03.930 And the name is going to be equal to the key. 08:03.930 --> 08:07.510 We just created GCR, JSON Key. 08:07.530 --> 08:09.680 So go ahead and patch this. 08:09.690 --> 08:14.370 Now we can go ahead and run kubectl rollout, restart deployment. 08:16.250 --> 08:24.290 Reservations so that a new reservations pod will be created can now run kubectl get pods to see our 08:24.290 --> 08:25.100 reservations. 08:25.100 --> 08:30.980 Pod is running and if we take a look we can see it's currently crashed loopback offing. 08:30.980 --> 08:37.400 If we run kubectl logs and paste the pod name in, we can see that the startup failed because we're 08:37.400 --> 08:42.740 not supplying the pod with the correct environment variables, which is what we expect to see because 08:42.740 --> 08:45.020 our validation is working correctly. 08:45.020 --> 08:48.670 Let's go ahead and continue fleshing out our deployment. 08:48.680 --> 08:53.000 Okay, so we're going to go ahead and create the deployments for the rest of our services before we 08:53.000 --> 08:54.680 define our environment variables. 08:54.680 --> 09:00.770 So let's go ahead and in our folder, in the templates directory, let's go ahead and create a new folder 09:00.770 --> 09:02.390 for each of our services. 09:02.390 --> 09:08.030 We of course already have the deployment here for the reservation service, so we'll go ahead and add 09:08.030 --> 09:13.130 an additional templates folder here for the auth service. 09:13.670 --> 09:20.910 Now in the terminal, we can CD into the templates folder and let's start off with the auth deployment. 09:20.910 --> 09:26.580 So CD into the auth folder and we'll run the command we already ran earlier for the reservations deployment. 09:26.580 --> 09:33.510 We'll run Kubectl, create deployment and we'll provide the image name now so we can go back to Google 09:33.510 --> 09:35.790 Cloud into artifact registry. 09:35.790 --> 09:42.420 So we'll click on Auth and click the copy button here to get the path to the auth repository and then 09:42.420 --> 09:44.250 we'll paste in the image name here. 09:44.250 --> 09:48.750 Make sure we also add the name of the deployment here in this case auth, and then we'll go ahead and 09:48.750 --> 09:56.640 specify dry run equal to client here and specify the output should be in YAML format and then we will 09:56.640 --> 10:00.600 pipe this to a deployment dot yaml file. 10:00.600 --> 10:05.850 So we should now see in the auth folder we have the deployment now filled out, we'll get rid of the 10:05.850 --> 10:07.200 time stamps here. 10:07.200 --> 10:14.070 So go ahead and delete the two creation time stamps, the resources status and strategy object and go 10:14.070 --> 10:20.160 ahead and add the slash production to the end of our image here to make sure we're using the production 10:20.160 --> 10:20.610 image. 10:20.610 --> 10:22.290 We'll go ahead and back in the terminal. 10:22.290 --> 10:26.010 We'll repeat the same process next for the payments deployment. 10:26.010 --> 10:32.820 So go ahead and regenerate this create deployment command and we will go ahead and just swap out the 10:32.820 --> 10:40.260 repository URL here to make sure we're targeting the payments image and go ahead and change the name 10:40.260 --> 10:42.480 of the deployment to payments as well. 10:42.480 --> 10:48.600 Then we'll go ahead and open up this new deployment and repeat the same process, delete the strategy, 10:48.630 --> 10:55.200 time stamps, resources and status, and go ahead and add a slash production to the end of the image 10:55.200 --> 10:55.950 name here. 10:55.950 --> 11:00.900 So finally, we need to do the same thing for the notifications service. 11:00.900 --> 11:08.580 So let's go ahead and create a new deployment here, change payments out and add notifications and change 11:08.580 --> 11:09.810 the name of the deployment. 11:09.840 --> 11:12.210 Now to notifications as well. 11:12.450 --> 11:19.800 Let's go ahead and open up the notifications deployment, delete the time stamps, strategy, resources 11:19.800 --> 11:24.720 and status, and add the slash production to notifications image. 11:24.720 --> 11:26.700 So now we have all of our deployments. 11:26.700 --> 11:28.410 We're going to need to deploy our app. 11:28.410 --> 11:34.030 Let's go ahead and upgrade our helm installation so that we can get these new deployments. 11:34.050 --> 11:41.670 To do this, we can make sure we're at the root of the chart here and we'll go ahead and run helm upgrade 11:41.670 --> 11:49.230 Sleeper and provide the path to the chart YAML In this case, in the current directory you can see that 11:49.230 --> 11:56.790 our deployment here was upgraded and we can run kubectl get pods to see all of our different deployments 11:56.790 --> 12:02.880 running and you can see they're all in an errored state and that's because we have no environment variables 12:02.880 --> 12:05.280 as we've already seen with the Reservations pod. 12:05.310 --> 12:09.570 We can run kubectl logs on the auth pod here and see the same issue. 12:09.600 --> 12:12.240 We're not passing the environment variable validation. 12:12.240 --> 12:15.900 So let's go ahead and continue to set up these environment variables properly.