1 00:00:00,001 --> 00:00:02,901 [No Audio] 2 00:00:02,901 --> 00:00:06,101 And we are back to GCP dashboard. We are still 3 00:00:06,101 --> 00:00:09,834 running on our credits, our bill is still 0, so 4 00:00:09,834 --> 00:00:11,867 we are quite safe on that front, let's go to the 5 00:00:11,867 --> 00:00:15,134 navigation menu, or the hamburger icon or the 6 00:00:15,134 --> 00:00:19,167 3 horizontal lines on the top left corner of 7 00:00:19,167 --> 00:00:22,334 our dashboard. Scroll down to the Kubernetes 8 00:00:22,334 --> 00:00:26,434 Engine and click on the Clusters. We don't have 9 00:00:26,434 --> 00:00:28,934 any clusters created yet. So we are having this 10 00:00:28,934 --> 00:00:31,867 screen, it is giving us a few options. 11 00:00:31,867 --> 00:00:36,201 Take a quick start, deploy a container directly, or create 12 00:00:36,201 --> 00:00:38,967 cluster. The most friendly looking one seems to be 13 00:00:38,967 --> 00:00:41,234 Create cluster, so let's go for it. 14 00:00:42,034 --> 00:00:46,901 Just like VM instances, GCP is prompting us to define some of 15 00:00:46,901 --> 00:00:48,967 the specifications of our cluster. 16 00:00:49,167 --> 00:00:52,367 But before that, let's see what do we have under cluster 17 00:00:52,367 --> 00:00:56,401 templates. We have Standard cluster, CPU intensive 18 00:00:56,401 --> 00:00:59,267 application cluster, Memory intensive cluster, 19 00:00:59,267 --> 00:01:02,567 GPU accelerated cluster, and Highly available 20 00:01:02,567 --> 00:01:05,301 cluster. All of these are useful for different 21 00:01:05,301 --> 00:01:08,434 applications. But we will stick to Standard cluster. 22 00:01:09,001 --> 00:01:11,734 The default given name is standard-cluster-1, 23 00:01:11,734 --> 00:01:14,867 but that's too cliche. Let's name it something else. 24 00:01:15,301 --> 00:01:17,467 Call it k8s which stands for 25 00:01:17,467 --> 00:01:22,234 Kubernetes cluster. Next choice is Location type, 26 00:01:22,234 --> 00:01:25,501 we can have either a Zonal or a Regional cluster, 27 00:01:25,667 --> 00:01:28,367 which means, do you want to spread out your cluster 28 00:01:28,401 --> 00:01:31,734 across different zones of a region, or do you want 29 00:01:31,734 --> 00:01:33,867 to spread out your cluster across different 30 00:01:33,867 --> 00:01:37,401 regions themselves? The choice of location type is 31 00:01:37,401 --> 00:01:40,134 permanent. For higher reliability, Regional might 32 00:01:40,134 --> 00:01:42,567 be the better choice. But we are not opting for 33 00:01:42,567 --> 00:01:44,867 such performance intensive applications at the 34 00:01:44,867 --> 00:01:47,134 moment, so we'll just run with zonal. 35 00:01:48,201 --> 00:01:52,267 Let's choose our zone, you can choose any zone you prefer. 36 00:01:52,401 --> 00:01:57,701 We'll pick europe-north1-a, next is Master version. 37 00:01:58,234 --> 00:02:00,667 This means the Kubernetes version, which 38 00:02:00,667 --> 00:02:03,267 will be running on the master instance. 39 00:02:03,867 --> 00:02:09,734 The default is 1.9.7 or gke.6 and we will 40 00:02:09,734 --> 00:02:13,434 keep it at that. Now let's define our node pool, 41 00:02:13,434 --> 00:02:16,601 which means the number of nodes and their machine types. 42 00:02:17,134 --> 00:02:20,167 We are creating a three node cluster and 43 00:02:20,167 --> 00:02:23,567 all of these nodes will help Container Optimized OS 44 00:02:23,567 --> 00:02:26,434 or cos install on them. 45 00:02:27,034 --> 00:02:31,467 We are using machines with one virtual CPU and 3.75 46 00:02:31,467 --> 00:02:36,367 gigs of memory and our boot disk size is 10 GB per node. 47 00:02:36,367 --> 00:02:39,367 We have enabled auto repairing to avoid 48 00:02:39,367 --> 00:02:42,501 any potential failures and our service account is 49 00:02:42,501 --> 00:02:45,201 Compute Engine default service account. 50 00:02:45,201 --> 00:02:48,367 Let's allow full access to all of the Coud APIs. 51 00:02:48,367 --> 00:02:51,934 We can also provide metadata to our cluster, like 52 00:02:51,934 --> 00:02:55,167 label and taint, but that feature is under beta 53 00:02:55,167 --> 00:02:58,967 right now. So let's skip it and click on Save. 54 00:02:59,334 --> 00:03:01,501 Finally, let's click Create. 55 00:03:02,767 --> 00:03:07,167 And here we are, our cluster is created. We have the name of the 56 00:03:07,167 --> 00:03:11,234 cluster which is k8s cluster, our zonal location, 57 00:03:11,267 --> 00:03:14,867 which is europe-north1-a, amount of nodes or 58 00:03:14,867 --> 00:03:17,967 cluster size, which is 3, total number of 59 00:03:17,967 --> 00:03:21,201 cores which is 3 vCPUs, because each of the 60 00:03:21,201 --> 00:03:26,367 node has 1 vCPU, and total memory, which is 11.25 gigs, 61 00:03:26,367 --> 00:03:28,534 because each of the nodes has three 62 00:03:28,534 --> 00:03:31,301 point 75 gigs of memory, we can look at the 63 00:03:31,301 --> 00:03:34,601 description of this cluster by clicking on the name of it. 64 00:03:35,134 --> 00:03:37,367 Starting from the top, we have Master 65 00:03:37,367 --> 00:03:40,734 version, which is the default one, endpoint for 66 00:03:40,734 --> 00:03:43,201 this particular cluster and some other 67 00:03:43,201 --> 00:03:47,134 information such as Pod address, range, and 68 00:03:47,134 --> 00:03:50,567 and the affirmation of Stackdriver Logging and Monitoring. 69 00:03:51,067 --> 00:03:55,567 Below that we also have Node Pool specs, which are 70 00:03:55,567 --> 00:03:57,601 pretty much what we had provided. 71 00:03:57,601 --> 00:04:01,467 Let's go to storage section. Well, this is not about the 72 00:04:01,467 --> 00:04:04,834 storage disk size of nodes. It is about the 73 00:04:04,834 --> 00:04:08,267 storage objects within the cluster. Since we 74 00:04:08,267 --> 00:04:11,101 haven't done anything to the cluster. Both of the 75 00:04:11,101 --> 00:04:14,301 fields of persistent volumes and storage classes 76 00:04:14,334 --> 00:04:17,301 don't have any special entries. And if you notice 77 00:04:17,301 --> 00:04:21,367 closely, there is a storage class called standard. 78 00:04:21,367 --> 00:04:24,434 Which provisions Google Compute Engine Standard 79 00:04:24,434 --> 00:04:27,201 persistent dis will use the storage class in 80 00:04:27,201 --> 00:04:30,201 future to provision some persistent volumes. 81 00:04:30,734 --> 00:04:32,834 Finally, let's go to Nodes section. 82 00:04:33,534 --> 00:04:37,434 These are the details of all three nodes of our cluster. 83 00:04:37,834 --> 00:04:41,801 These long and complicated names are the names of VMs 84 00:04:41,801 --> 00:04:43,767 which are used to create the cluster. 85 00:04:43,867 --> 00:04:45,934 All of their status is Ready. 86 00:04:46,767 --> 00:04:51,567 Next up, we have requested and allocatable CPU size, while the 87 00:04:51,567 --> 00:04:55,367 requested CPU sizes are different, allocatable size 88 00:04:55,367 --> 00:05:00,301 is common which is 940 mCPU, which means more 89 00:05:00,301 --> 00:05:04,034 or less 1 vCPU. Next up, we have requested 90 00:05:04,034 --> 00:05:08,101 memory and allocatable memory. 2.7 gigs are 91 00:05:08,101 --> 00:05:12,234 allocatable whereas we had provided 3.7 gigs. 92 00:05:12,234 --> 00:05:15,267 So what about the rest of the memory? Well, you can 93 00:05:15,267 --> 00:05:19,001 comprehend allocatable memory as user space memory 94 00:05:19,001 --> 00:05:22,167 of all of these virtual machines, which means 95 00:05:22,167 --> 00:05:25,134 remaining 1 gigabyte will be used by the kernel 96 00:05:25,134 --> 00:05:28,301 space of these machines. Till now we haven't 97 00:05:28,301 --> 00:05:30,967 requested or allocated any storage yet. 98 00:05:31,401 --> 00:05:34,534 It seems like we know our cluster better than before. 99 00:05:34,534 --> 00:05:37,634 In next lecture, we will navigate to the cluster 100 00:05:37,634 --> 00:05:39,434 and play around with it a bit.