0
1
00:00:00,390 --> 00:00:06,060
Let's see is there any well known wait and containerized a legacy VM based monolithic app and make it
1

2
00:00:06,060 --> 00:00:06,720
cloud native.
2

3
00:00:06,720 --> 00:00:12,330
And how do you redirect the incoming consumer request from the VM to those containers.
3

4
00:00:12,330 --> 00:00:14,910
So this is a very loaded question and
4

5
00:00:20,170 --> 00:00:21,480
going to make this the last question.
5

6
00:00:21,730 --> 00:00:30,040
So because everyone else is taking my questions and I'm running out of time so this is a long topic
6

7
00:00:30,130 --> 00:00:31,600
and it's a great topic.
7

8
00:00:31,600 --> 00:00:38,860
Basically what you're asking here is we read the question again Is there any well known way to containerized
8

9
00:00:38,920 --> 00:00:43,720
a legacy VM based on monolithic applications and make it clone native.
9

10
00:00:44,200 --> 00:00:48,640
And how do you redirect the incoming consumer requests from VM to those containers.
10

11
00:00:48,640 --> 00:00:51,160
So that is two separate questions.
11

12
00:00:51,370 --> 00:00:58,560
I would do the I'm going to answer the first one first to containerized a legacy VM Docker provides.
12

13
00:00:58,600 --> 00:01:03,850
If you're an enterprise customer they provide a certain level of tools depending on what the app is
13

14
00:01:04,150 --> 00:01:09,670
and it's very framework specific because you know obviously a python app has certain requirements certain
14

15
00:01:09,670 --> 00:01:17,080
dependencies you cannot containerized a full VM that's not a thing because a VM includes the kernel
15

16
00:01:17,320 --> 00:01:22,390
and the kernel drivers and all sorts of things that don't go in containers so you wouldn't wouldn't
16

17
00:01:22,390 --> 00:01:27,850
want to containerized the rest of the OS anyway because that's probably at least a gig of unnecessary
17

18
00:01:27,850 --> 00:01:30,930
stuff inside that VM right.
18

19
00:01:31,000 --> 00:01:35,860
If you don't know if it's a Windows machine it's probably 10 gig of unnecessary stuff so you wouldn't
19

20
00:01:35,860 --> 00:01:41,890
want to do that either but it's not that hard to containerized a legacy app because surely somewhere
20

21
00:01:42,220 --> 00:01:47,740
you have some written documentation or even just screenshots of how you installed that.
21

22
00:01:47,860 --> 00:01:53,620
And as long as you have that basic information you can recreate that in a docker file and really a docker
22

23
00:01:53,620 --> 00:01:58,540
file is just like an automated way of installing something on Linux or Windows.
23

24
00:01:58,660 --> 00:02:01,820
And so if you have those instructions you can do that.
24

25
00:02:01,990 --> 00:02:07,140
And there are tools at Docker they kind of sell them with Docker enterprise.
25

26
00:02:07,240 --> 00:02:12,540
You can basically go back to Docker con twenty eighteen in San Francisco from a year ago.
26

27
00:02:12,550 --> 00:02:17,530
You can look that up on YouTube and they demo some of these tools in their keynote speech where they
27

28
00:02:17,530 --> 00:02:23,980
talk about taking legacy web apps mostly especially on Windows side because those can be like Dot Net
28

29
00:02:24,130 --> 00:02:25,910
framework apps and stuff like that.
29

30
00:02:26,020 --> 00:02:30,910
And then finding all of the parts and pieces to them based on their own package files that come with
30

31
00:02:30,910 --> 00:02:31,700
those apps right.
31

32
00:02:31,730 --> 00:02:36,850
I use these apps have their own package installation tooling and they look at those files and then find
32

33
00:02:36,850 --> 00:02:38,270
all the right pieces and put it in.
33

34
00:02:38,470 --> 00:02:43,870
But I don't know of a tool that I would recommend that is going to just look at the entire operating
34

35
00:02:43,870 --> 00:02:46,440
system and then suck it all in into one container.
35

36
00:02:46,450 --> 00:02:53,290
So I would say look at the information on how that app is installed or even just basically if you don't
36

37
00:02:53,290 --> 00:02:54,240
even have that.
37

38
00:02:54,310 --> 00:02:59,440
Just start looking at you know just start trying with a docker file and you can work on and iterate
38

39
00:02:59,470 --> 00:03:02,380
to get to the point that you actually have it working in the container.
39

40
00:03:02,380 --> 00:03:08,080
Now if you're dealing with the data moving the data around is a separate conversation.
40

41
00:03:08,080 --> 00:03:10,360
You've got to get that data into a volume first.
41

42
00:03:10,690 --> 00:03:12,850
And so that's that's going to be a separate process.
42

43
00:03:12,850 --> 00:03:17,800
I would first first focus on getting the app working then I would worry as a sort of a sub project on
43

44
00:03:17,800 --> 00:03:23,450
how to move the data over if that's indeed a problem where you have unique data inside that old VM.
44

45
00:03:23,530 --> 00:03:28,690
And then thirdly how do you redirect it incoming consumer requests from the VM to those containers that
45

46
00:03:28,690 --> 00:03:31,260
is going to be completely dependent upon your infrastructure.
46

47
00:03:31,360 --> 00:03:33,820
And that is really not a whole lot to do with containers.
47

48
00:03:33,940 --> 00:03:36,420
That has to do with is this a web request.
48

49
00:03:36,430 --> 00:03:41,860
Is there a front end proxy in front of it like an elastic load balancer or a application load balancer
49

50
00:03:41,890 --> 00:03:44,880
or your own third party F5 proxy.
50

51
00:03:44,980 --> 00:03:46,780
Is it is it not a web app.
51

52
00:03:46,780 --> 00:03:51,040
If it's not a web app then you've got to look at how do those connections come in.
52

53
00:03:51,040 --> 00:03:54,160
Because Docker is not going to do that for you Docker doesn't do connection.
53

54
00:03:54,280 --> 00:03:59,710
You know redirect from an old VM that's not a thing that containers or container orchestration solves
54

55
00:03:59,770 --> 00:04:01,450
that that I've ever seen at least.
55

56
00:04:01,450 --> 00:04:05,260
So that's really three parts of a major project.
56

57
00:04:05,260 --> 00:04:10,990
And so when you hire a consultant to help you migrate this is kind of what I do migrate old applications
57

58
00:04:11,020 --> 00:04:15,590
into containers Docker has an entire part of their company dedicated to this.
58

59
00:04:15,670 --> 00:04:21,520
They call it the MTA program migrating traditional apps and they basically send out Tiger Teams to companies
59

60
00:04:21,520 --> 00:04:24,100
for a fee and they'll spend a week on site.
60

61
00:04:24,100 --> 00:04:28,770
I think it is and they'll take your app migrate it and you'll learn everything about it.
61

62
00:04:28,950 --> 00:04:33,970
How help you essentially migrated into a container or copy it in a container get it running get the
62

63
00:04:33,970 --> 00:04:37,630
data moved over and then swap the connections in your case if you don't have any sort of front end load
63

64
00:04:37,630 --> 00:04:43,780
balancing you've got to have basically you've got to have that app running in the container and then
64

65
00:04:43,780 --> 00:04:49,780
figure out a downtime you know it could be just minutes where you're going to redirect the DNS traffic
65

66
00:04:49,780 --> 00:04:53,950
or whatever you need to get that flipped over and that has nothing to do with containers and everything
66

67
00:04:53,950 --> 00:04:59,230
to do with just routing and DNS and networking that basically the same process you would do for moving
67

68
00:04:59,230 --> 00:05:06,670
from an old VM to a new VM when you have to replace VMS over time but that's the Cliff Notes on essentially
68

69
00:05:06,820 --> 00:05:11,890
what it's like to be a docker consultant because you're always migrating old apps and the new stuff
69

70
00:05:11,890 --> 00:05:13,390
for other people and helping them do that.