1
00:00:00,620 --> 00:00:10,480
‫All right, now that we've learned all about Swarm scaling with overlayed networking and routing mesh,

2
00:00:11,350 --> 00:00:18,910
‫I want you to actually go through the process of creating the assets, or the objects, that you need to

3
00:00:18,910 --> 00:00:20,620
‫deploy in a Swarm scenario.

4
00:00:20,620 --> 00:00:26,110
‫So we're no longer using the docker run command, but the networking and the volumes are kind of the

5
00:00:26,110 --> 00:00:28,640
‫same but slightly different.

6
00:00:28,660 --> 00:00:36,030
‫I want to give you a real-world scenario of a multi-tier app, and we're going to use Docker’s Distributed

7
00:00:36,030 --> 00:00:36,820
‫Voting App.

8
00:00:36,840 --> 00:00:38,830
‫It's actually called the Example Voting App.

9
00:00:38,910 --> 00:00:43,980
‫You might have seen it in Docker keynotes, or in other labs or tutorials.

10
00:00:43,980 --> 00:00:49,860
‫It's a pretty common, and I think, a great app, because it's got at least five roles to it, or different

11
00:00:49,860 --> 00:00:51,540
‫containers that you need to deploy.

12
00:00:51,720 --> 00:00:56,760
‫And they all serve a special function and they're actually doing stuff that actually works, actually does

13
00:00:56,760 --> 00:00:57,740
‫what it's supposed to do.

14
00:00:57,750 --> 00:01:03,000
‫And so you can play with it and it has very typical, distributed architecture components like the web

15
00:01:03,010 --> 00:01:04,870
‫frontend and the worker backend,

16
00:01:04,870 --> 00:01:12,190
‫and the key value in a database. I think it's really helpful to do. In this case, you're going to use

17
00:01:12,370 --> 00:01:18,960
‫the directory swarm-app-1. It's actually in our course repo. In there is nothing but a readme and

18
00:01:18,960 --> 00:01:26,590
‫an image. From that readme, you're going need to create the volume and a couple of networks and five

19
00:01:26,590 --> 00:01:30,280
‫services that are needed for this app to fully function.

20
00:01:30,370 --> 00:01:36,580
‫It may help you to try some of these commands locally, or actually write them in your text editor

21
00:01:36,790 --> 00:01:40,290
‫and sort of compose them, because they are going to be a little complex, right. When you do that docker

22
00:01:40,300 --> 00:01:45,650
‫service create, you’re going to have to connect it to a specific network, and open up ports, and call it

23
00:01:45,650 --> 00:01:51,490
‫a specific name, and use a specific image. So, it might be easier to edit it locally and then just cut

24
00:01:51,490 --> 00:01:54,400
‫and paste into your shell, wherever your Swarm is at.

25
00:01:54,520 --> 00:01:56,240
‫But that's just me.

26
00:01:56,260 --> 00:02:01,390
‫Everything is going to use Docker Hub images because in the production Swarm you don't want to be building.

27
00:02:01,450 --> 00:02:05,260
‫That's not a process you want on your production Swarm, typically.

28
00:02:05,290 --> 00:02:10,390
‫Usually, you want to do your building either through some automated online service like Docker Hub or

29
00:02:10,390 --> 00:02:17,280
‫Docker Cloud, or somewhere in your CI/CD pipeline that's not eating resources on your production Swarm.

30
00:02:17,290 --> 00:02:22,630
‫In this scenario, we can imagine that our images are being built elsewhere and stored on Docker Hub and

31
00:02:22,630 --> 00:02:28,720
‫we just need to pull those down, whatever the latest versions are, into our Swarm.

32
00:02:28,840 --> 00:02:34,720
‫And I'll just say this, like many things in Computer Science, this is one half science, one half art form.

33
00:02:34,930 --> 00:02:40,650
‫Your commands could be in totally different orders, or have totally different option orders than mine.

34
00:02:40,660 --> 00:02:43,200
‫You might call your networks different things.

35
00:02:43,270 --> 00:02:43,940
‫That's fine.

36
00:02:43,950 --> 00:02:48,010
‫At the end result is that the application works, that they're able to talk to each other in the proper

37
00:02:48,010 --> 00:02:52,180
‫way, that you've got all the ports published and that you’ve got the data protected in a volume.

38
00:02:52,180 --> 00:02:55,020
‫Let's take a quick look at what you're up against.

39
00:02:55,030 --> 00:02:59,170
‫Imagine you're having to design the commands and the architecture for this.

40
00:02:59,170 --> 00:03:01,970
‫And this is what the developer would give to you that wrote this app.

41
00:03:02,140 --> 00:03:05,280
‫Well in this case, it's probably lots of developers that wrote this app, right?

42
00:03:05,290 --> 00:03:10,120
‫Because we've actually got a web frontend that's in Python we've got a Web backend that's running

43
00:03:10,120 --> 00:03:11,140
‫in Node.

44
00:03:11,140 --> 00:03:17,170
‫We've got a worker process that's running in .NET and then we've got a Redis key value store and

45
00:03:17,180 --> 00:03:19,600
‫a PostgreSQL database.

46
00:03:19,630 --> 00:03:24,610
‫This is actually something I see more and more often, where different teams are writing different parts

47
00:03:24,610 --> 00:03:29,510
‫of a solution, and they're able to just pick the language that they think is best for that scenario.

48
00:03:29,660 --> 00:03:34,660
‫And in the world of containers, this kind of application really shines. Because our applications are

49
00:03:34,660 --> 00:03:35,750
‫all segmented,

50
00:03:35,770 --> 00:03:40,630
‫they can still technically run on the same machine but they're all protected from each other so that

51
00:03:40,630 --> 00:03:45,170
‫you don't end up with Node conflicting with dependencies on Python or whatever.

52
00:03:45,190 --> 00:03:49,810
‫You’ll also notice here the traffic flow, and the voting-app doesn't have it,

53
00:03:49,810 --> 00:03:53,680
‫but it should have a little arrow coming in that this is where the users would be coming in on the voting-app

54
00:03:53,710 --> 00:03:54,320
‫part.

55
00:03:54,460 --> 00:03:59,530
‫And you can see that the voting-app actually has to push your votes for this application which, by the way, you’re

56
00:03:59,530 --> 00:04:01,300
‫going to be voting for cats and dogs.

57
00:04:01,400 --> 00:04:07,780
‫It'll be pushing that into Redis and then the worker will actually be checking the Redis system for anything

58
00:04:07,780 --> 00:04:15,060
‫in the queue and then be pushing the results into Postgres which will then be shown on a WebSockets

59
00:04:15,100 --> 00:04:19,960
‫backend that's actually live-updating as you're voting.

60
00:04:19,990 --> 00:04:25,540
‫If we jump into the Directory real quick I've put some notes in here to help you along the way because

61
00:04:25,540 --> 00:04:28,400
‫there's a few things that are a little different and new.

62
00:04:28,510 --> 00:04:34,240
‫But, the services area would be kind of written like it was from someone who knew their application but

63
00:04:34,240 --> 00:04:38,390
‫didn't know actually how to implement this in Docker. And that's a pretty common scenario.

64
00:04:38,530 --> 00:04:43,110
‫You can see that in the services I've listed which image you need to deploy,

65
00:04:43,270 --> 00:04:46,310
‫the actual name that you need to give it, because those are important.

66
00:04:46,360 --> 00:04:51,130
‫These are images that are, unfortunately, hard coded for what names they are going to look for on the network.

67
00:04:51,130 --> 00:04:55,350
‫Ideally, you'd want applications that you could configure on the fly for what they need to do for names,

68
00:04:55,360 --> 00:05:01,360
‫but, in this case, vote is going to specifically look for a DNS record called redis. You can see that

69
00:05:01,360 --> 00:05:06,460
‫there's actually two networks here. We have a frontend and a backend network. That's pretty typical

70
00:05:06,460 --> 00:05:11,500
‫also of a larger application where you maybe want to protect your database on a backend network so

71
00:05:11,500 --> 00:05:14,580
‫that the frontend user app doesn't have direct access to it.

72
00:05:14,680 --> 00:05:19,790
‫And that's what we are showing here and you’ll notice that the worker actually is connected to two networks.

73
00:05:19,930 --> 00:05:25,830
‫So when you use your create command, remember you're using docker service create not docker run.

74
00:05:25,840 --> 00:05:29,950
‫So when you using your docker service create command you're going to need to specify the --network

75
00:05:29,980 --> 00:05:33,460
‫twice, once for the frontend, and once for the backend.

76
00:05:33,490 --> 00:05:36,100
‫Obviously, you know to create your networks first.

77
00:05:36,100 --> 00:05:41,290
‫So some of that’s specified up here. And if you need the image again the architecture diagram is also in

78
00:05:41,290 --> 00:05:45,350
‫here, just in case that helps conceptually with how you're going to deploy this.

79
00:05:45,700 --> 00:05:50,680
‫And, of course, you're going to have the Docker documentation. You might need to use the --help on

80
00:05:50,680 --> 00:05:54,790
‫the service create command that will give you all the various options that you might need.

81
00:05:54,790 --> 00:06:00,580
‫You only are going to need three or four of those. If you get stuck with the commands themselves, you

82
00:06:00,580 --> 00:06:06,430
‫might want to review the actual documentation, which I will put into the resources of this section.

83
00:06:06,520 --> 00:06:13,250
‫But this particular document under Swarm actually goes into creating services and configuring services.

84
00:06:13,330 --> 00:06:18,610
‫So, it's a little wordy, but it's got a lot of information in it that might help you if you get stuck

85
00:06:18,610 --> 00:06:19,160
‫in here.

86
00:06:19,180 --> 00:06:22,700
‫The next video I will actually run through this whole thing myself as if I was you.

87
00:06:22,750 --> 00:06:23,980
‫Good luck and have some fun.