1 00:00:01,623 --> 00:00:04,240 So having a deployment is nice, 2 00:00:04,240 --> 00:00:05,780 but not enough. 3 00:00:05,780 --> 00:00:07,830 I will add a second file here, 4 00:00:07,830 --> 00:00:10,160 and that's the service.yaml file, 5 00:00:10,160 --> 00:00:12,450 where I wanna define the service 6 00:00:12,450 --> 00:00:15,990 for my deployment for this application. 7 00:00:15,990 --> 00:00:18,300 And the service.yaml file of course, 8 00:00:18,300 --> 00:00:21,240 is very similar to the deployment.yaml file, 9 00:00:21,240 --> 00:00:23,460 at least to some extent. 10 00:00:23,460 --> 00:00:26,320 Since it's still a Kubernetes Config File, 11 00:00:26,320 --> 00:00:29,400 we still need to define the API version. 12 00:00:29,400 --> 00:00:33,080 Here, the difference compared to deployment is, 13 00:00:33,080 --> 00:00:35,130 that it's just v1. 14 00:00:35,130 --> 00:00:38,380 And we can see this in the official reference documentation, 15 00:00:38,380 --> 00:00:43,380 there, if you have a look at the service resource, 16 00:00:43,560 --> 00:00:46,610 you see the group is the core group, 17 00:00:46,610 --> 00:00:49,140 and that essentially means that you can omit it. 18 00:00:49,140 --> 00:00:53,440 So it's not core/v1, it's just v1. 19 00:00:53,440 --> 00:00:55,770 Now, I as mentioned before that alone, 20 00:00:55,770 --> 00:00:58,200 also in conjunction with the file name, 21 00:00:58,200 --> 00:01:01,050 won't tell Cooper, Nita is what you wanna create. 22 00:01:01,050 --> 00:01:03,150 Because it ignores the file name, 23 00:01:03,150 --> 00:01:06,020 since that name is totally up to you. 24 00:01:06,020 --> 00:01:09,160 Instead, you need to specify a kind in this case, 25 00:01:09,160 --> 00:01:10,170 the service kind. 26 00:01:10,170 --> 00:01:12,710 Since we wanna create a service resource 27 00:01:12,710 --> 00:01:16,513 and you again see that here in the reference docs as well. 28 00:01:17,760 --> 00:01:20,230 There, you'll also see what else can be defined, 29 00:01:20,230 --> 00:01:21,760 and that's of course again, 30 00:01:21,760 --> 00:01:26,350 the metadata and the specification. 31 00:01:26,350 --> 00:01:28,280 So for the metadata, I again, 32 00:01:28,280 --> 00:01:32,960 wanna give this service a name and I'll name it backend, 33 00:01:32,960 --> 00:01:35,080 the name is of course, up to you. 34 00:01:35,080 --> 00:01:36,400 And then for the spec, 35 00:01:36,400 --> 00:01:39,993 we will again define and configure this service. 36 00:01:42,260 --> 00:01:45,830 And here, if we have a look at the reference doc, 37 00:01:45,830 --> 00:01:47,420 we see that there's again, 38 00:01:47,420 --> 00:01:49,960 a lot we can configure because of course 39 00:01:49,960 --> 00:01:53,720 kubernetes is very powerful and gives us a lot of options. 40 00:01:53,720 --> 00:01:58,070 But again, we typically don't need all these options. 41 00:01:58,070 --> 00:02:01,040 What we do need here though, is a selector. 42 00:02:01,040 --> 00:02:04,120 And I did talk about selectors before. 43 00:02:04,120 --> 00:02:07,790 A selector identifies which other resources 44 00:02:07,790 --> 00:02:12,790 should be controlled or connected to this resource. 45 00:02:13,220 --> 00:02:16,550 In this case here, we wanna define which pods 46 00:02:16,550 --> 00:02:19,180 should be part of this service. 47 00:02:19,180 --> 00:02:20,950 Because as we learned before, 48 00:02:20,950 --> 00:02:23,210 a service object exposes pods 49 00:02:23,210 --> 00:02:26,460 to the cluster or to the outside world. 50 00:02:26,460 --> 00:02:28,930 So you don't control deployments with a service, 51 00:02:28,930 --> 00:02:30,900 but instead pods, but of course, 52 00:02:30,900 --> 00:02:33,850 pods are typically generated by a deployment 53 00:02:33,850 --> 00:02:36,050 nonetheless with that selector here, 54 00:02:36,050 --> 00:02:39,100 we wanna select the individual pods. 55 00:02:39,100 --> 00:02:41,170 Now this elector of the service object 56 00:02:41,170 --> 00:02:43,070 works a bit differently than the one 57 00:02:43,070 --> 00:02:44,860 from the deployment object. 58 00:02:44,860 --> 00:02:47,190 There, we had matchlabels 59 00:02:47,190 --> 00:02:50,050 and then our label key value pairs, 60 00:02:50,050 --> 00:02:51,733 which we wanted to connect. 61 00:02:52,700 --> 00:02:54,940 Now, we had this syntax here 62 00:02:54,940 --> 00:02:59,320 because the deployment resource type is a bit newer. 63 00:02:59,320 --> 00:03:02,350 It has been added more recently, you could say, 64 00:03:02,350 --> 00:03:04,560 even though it's also over here around 65 00:03:04,560 --> 00:03:06,010 for quite some time now. 66 00:03:06,010 --> 00:03:08,450 But the service resource is simply older 67 00:03:08,450 --> 00:03:11,530 and there, the selector is simpler. 68 00:03:11,530 --> 00:03:14,110 You don't have matchlabels below it 69 00:03:14,110 --> 00:03:17,100 because you can only match labels here. 70 00:03:17,100 --> 00:03:18,890 So therefore you directly have 71 00:03:18,890 --> 00:03:20,963 to label key value pairs here. 72 00:03:21,810 --> 00:03:25,120 And I wanna control my ports, 73 00:03:25,120 --> 00:03:26,680 which are created like this, 74 00:03:26,680 --> 00:03:29,540 so which have these labels. 75 00:03:29,540 --> 00:03:32,880 Hence we can copy the labels here and add them here 76 00:03:32,880 --> 00:03:36,743 to selector, just make sure the indentation is correct. 77 00:03:37,730 --> 00:03:39,810 Now, you don't always have to select 78 00:03:39,810 --> 00:03:41,860 by all the labels though. 79 00:03:41,860 --> 00:03:45,740 Maybe that is what you want, but maybe in your project, 80 00:03:45,740 --> 00:03:47,700 you also have another deployment, 81 00:03:47,700 --> 00:03:50,820 which creates ports with an app of second app, 82 00:03:50,820 --> 00:03:53,790 but which does not have a tier of backend. 83 00:03:53,790 --> 00:03:55,230 And whilst these ports 84 00:03:55,230 --> 00:03:57,270 would be part of different deployments, 85 00:03:57,270 --> 00:04:00,160 you maybe wanna group them into the same service. 86 00:04:00,160 --> 00:04:03,930 So you could choose to just select by the app label 87 00:04:03,930 --> 00:04:05,900 and ignore the tier label. 88 00:04:05,900 --> 00:04:10,470 So that all ports which have that app label with that value, 89 00:04:10,470 --> 00:04:12,830 would be controlled by that service, 90 00:04:12,830 --> 00:04:15,520 even if they then have different tiers. 91 00:04:15,520 --> 00:04:17,300 Not something we have here, 92 00:04:17,300 --> 00:04:19,670 but something you can do with selectors, 93 00:04:19,670 --> 00:04:22,280 and definitely the idea behind selectors. 94 00:04:22,280 --> 00:04:25,920 The idea is that you have great flexibility when it comes 95 00:04:25,920 --> 00:04:30,920 to expressing which other resource should be connected to 96 00:04:31,110 --> 00:04:33,853 or controlled by some other resource. 97 00:04:34,780 --> 00:04:37,870 So in this case I'm saying this service selects 98 00:04:37,870 --> 00:04:42,870 and therefor exposes all ports with an app label 99 00:04:43,150 --> 00:04:45,370 with a value of second app. 100 00:04:45,370 --> 00:04:48,040 And that of course means that the ports generated 101 00:04:48,040 --> 00:04:51,893 by this deployment will be exposed by this service. 102 00:04:53,230 --> 00:04:56,520 So that's great, now we know which pods should be exposed, 103 00:04:56,520 --> 00:04:59,920 but we don't know yet how they should be exposed 104 00:04:59,920 --> 00:05:02,123 and which ports should be exposed. 105 00:05:03,140 --> 00:05:07,200 And keep in mind that when we used the imperative approach, 106 00:05:07,200 --> 00:05:09,440 we used the expose command 107 00:05:09,440 --> 00:05:12,700 to expose our first app deployment. 108 00:05:12,700 --> 00:05:15,460 And there we added dash dash port 109 00:05:15,460 --> 00:05:18,280 to decide which port should be exposed. 110 00:05:18,280 --> 00:05:22,170 And we add a dash dash type loadBalancer 111 00:05:22,170 --> 00:05:25,970 to decide how it should be exposed basically. 112 00:05:25,970 --> 00:05:27,670 And we can do the same things here. 113 00:05:27,670 --> 00:05:30,680 Let's start with the ports and you can expose more 114 00:05:30,680 --> 00:05:33,920 than one port, which is why this is named ports. 115 00:05:33,920 --> 00:05:36,770 And therefore you then have a list of ports 116 00:05:36,770 --> 00:05:39,220 that can be exposed. 117 00:05:39,220 --> 00:05:40,540 And there, you typically 118 00:05:40,540 --> 00:05:43,680 always define the protocol being used, 119 00:05:43,680 --> 00:05:45,793 which by default here is TCP. 120 00:05:46,820 --> 00:05:50,650 And then on the same list item the port, 121 00:05:50,650 --> 00:05:52,210 you wanna expose it on, 122 00:05:52,210 --> 00:05:56,090 let's say port 80 and the target port 123 00:05:56,090 --> 00:05:59,160 inside of the container in this case 80 80, 124 00:05:59,160 --> 00:06:00,730 because that is the port, 125 00:06:00,730 --> 00:06:03,293 my application is listening on. 126 00:06:05,520 --> 00:06:07,960 So therefore that's the target port, in the container 127 00:06:07,960 --> 00:06:10,500 and that's then the outside world port 128 00:06:10,500 --> 00:06:12,866 on which this should be exposed. 129 00:06:12,866 --> 00:06:15,180 And if you had multiple ports, 130 00:06:15,180 --> 00:06:17,970 you could also of course do that, right? 131 00:06:17,970 --> 00:06:22,470 So if you had another port, if you had another port, 132 00:06:22,470 --> 00:06:27,470 let's say port 443 you could of course also exposed that. 133 00:06:29,020 --> 00:06:31,663 But we don't have it here, so I'll comment this out. 134 00:06:32,920 --> 00:06:37,920 So that's the ports option now for the type option, 135 00:06:39,200 --> 00:06:42,300 you simply add that next to ports. 136 00:06:42,300 --> 00:06:46,550 So not nested below it, but on the same level as ports, 137 00:06:46,550 --> 00:06:48,360 you add your type. 138 00:06:48,360 --> 00:06:50,970 And there you got the types available, 139 00:06:50,970 --> 00:06:54,360 which I described, clusterIP, which is the default, 140 00:06:54,360 --> 00:06:56,900 which is an internally exposed IP. 141 00:06:56,900 --> 00:07:00,458 So only accessible from inside the cluster, 142 00:07:00,458 --> 00:07:03,750 NodePort to basically expose it on the IP 143 00:07:03,750 --> 00:07:05,300 and port of the worker node 144 00:07:05,300 --> 00:07:10,100 on which it runs and most commonly used loadBalancer. 145 00:07:10,100 --> 00:07:12,883 At least if you want outside world access. 146 00:07:13,760 --> 00:07:16,620 This then utilizes the loadBalancer 147 00:07:16,620 --> 00:07:19,870 off your cluster infrastructure, in this case, 148 00:07:19,870 --> 00:07:24,870 minikubes loadBalancer, and you will get an address, 149 00:07:24,900 --> 00:07:27,470 an outside world reachable address, 150 00:07:27,470 --> 00:07:29,830 which then allows you to reach this service 151 00:07:29,830 --> 00:07:32,900 and therefor the ports exposed by that service. 152 00:07:32,900 --> 00:07:37,343 And in addition, incoming traffic is distributed equally. 153 00:07:38,620 --> 00:07:42,410 And with that, we got our service object defined. 154 00:07:42,410 --> 00:07:44,470 That's our service resource 155 00:07:44,470 --> 00:07:46,733 and that's the service we now wanna create. 156 00:07:47,580 --> 00:07:51,730 And we created now just as we created our deployment, 157 00:07:51,730 --> 00:07:56,730 with kube control, apply dash file service.yaml. 158 00:07:56,850 --> 00:07:58,710 It's always the same command 159 00:07:58,710 --> 00:08:00,670 because it's simply looks into the file 160 00:08:00,670 --> 00:08:03,320 to then find out what it needs to create. 161 00:08:03,320 --> 00:08:06,760 So for now had enter this services created. 162 00:08:06,760 --> 00:08:10,450 So if we now get our services, we will see it here. 163 00:08:10,450 --> 00:08:15,340 And of course we can now expose it with minikube service 164 00:08:15,340 --> 00:08:18,603 and then the name you chose here, so in this case backend. 165 00:08:20,200 --> 00:08:22,860 and this will open up a new browser tab 166 00:08:22,860 --> 00:08:25,940 and you can view your application here. 167 00:08:25,940 --> 00:08:27,780 If it doesn't open a new tab, you can, 168 00:08:27,780 --> 00:08:31,340 of course manually use this URL here. 169 00:08:31,340 --> 00:08:34,159 So this is now working. 170 00:08:34,159 --> 00:08:36,950 This is now how we can create our deployment 171 00:08:36,950 --> 00:08:39,340 and our service in a declarative way. 172 00:08:39,340 --> 00:08:43,610 And this is the way and the approach you typically use 173 00:08:43,610 --> 00:08:45,490 because it's easier to maintain. 174 00:08:45,490 --> 00:08:48,370 You don't always have to re-type commands 175 00:08:48,370 --> 00:08:50,340 and it's less error prone. 176 00:08:50,340 --> 00:08:53,130 If you type a command like to Kube control, 177 00:08:53,130 --> 00:08:56,590 expose command, it's easy to mistype something, 178 00:08:56,590 --> 00:08:58,270 or to forget something. 179 00:08:58,270 --> 00:09:00,410 If you have it in your conflict file, 180 00:09:00,410 --> 00:09:03,630 making changes is easy, you don't have to repeat everything. 181 00:09:03,630 --> 00:09:05,440 You can just change one thing, 182 00:09:05,440 --> 00:09:08,300 and it's also very easy to share. 183 00:09:08,300 --> 00:09:10,400 In addition, if a colleague wrote this, 184 00:09:10,400 --> 00:09:13,560 you can quickly see how something needs to be configured, 185 00:09:13,560 --> 00:09:17,010 and you don't have to ask your colleague on how you should 186 00:09:17,010 --> 00:09:19,950 run this command the next time you're running it. 187 00:09:19,950 --> 00:09:22,030 So that's why the declarative approach 188 00:09:22,030 --> 00:09:25,040 is the recommended approach and why it is the approach 189 00:09:25,040 --> 00:09:27,063 We will use throughout this course.