1
00:00:00,001 --> 00:00:01,901
[No Audio]

2
00:00:01,901 --> 00:00:05,301
Now that we have deployed our Nginx service across

3
00:00:05,301 --> 00:00:07,967
the Swarm cluster successfully, let's think of

4
00:00:07,967 --> 00:00:11,767
some more innovative use cases. For example, what

5
00:00:11,767 --> 00:00:13,767
if I want to take down one of my nodes for

6
00:00:13,767 --> 00:00:17,967
maintenance, or what if one of my nodes actually

7
00:00:17,967 --> 00:00:21,767
goes down. Here, we will test it out.

8
00:00:22,867 --> 00:00:28,401
The safe way to make a node leave the cluster is to drain it,

9
00:00:28,634 --> 00:00:32,366
we can do it with docker node update availability

10
00:00:32,366 --> 00:00:36,034
command, followed by the action and the name of

11
00:00:36,034 --> 00:00:39,434
the node. Here, the command will work like

12
00:00:39,434 --> 00:00:45,601
docker node update --availability flag, drain worker-2,

13
00:00:46,101 --> 00:00:49,467
and what we get is the name of the node as the

14
00:00:49,467 --> 00:00:53,067
confirmation that it has drained out. Still, we

15
00:00:53,067 --> 00:00:57,367
can verify it by typing docker node ls, and we can

16
00:00:57,367 --> 00:01:00,967
see that the status of worker-2 node is still

17
00:01:00,967 --> 00:01:05,134
Ready, but the availability is Drain. Which means

18
00:01:05,233 --> 00:01:08,367
that the node is up, but its availability is

19
00:01:08,367 --> 00:01:10,867
drained, which means that no containers can be

20
00:01:10,867 --> 00:01:13,901
scheduled on it. When we drained the node, the

21
00:01:13,901 --> 00:01:16,767
container or the task scheduled on the node gets

22
00:01:16,767 --> 00:01:19,667
transferred or rescheduled to one of the other

23
00:01:19,667 --> 00:01:25,201
nodes. Let's verify it using docker service ps web-server.

24
00:01:25,201 --> 00:01:28,901
And as you can see, the web-server.3

25
00:01:28,901 --> 00:01:32,301
container has been shifted from worker-2

26
00:01:32,301 --> 00:01:36,234
to manager and it has been running since 42 seconds,

27
00:01:36,234 --> 00:01:40,434
which is about the time when worker-2 was drained.

28
00:01:40,701 --> 00:01:44,467
On the other hand, if we use docker ps

29
00:01:44,467 --> 00:01:47,401
on worker-2 now, which has been drained,

30
00:01:47,467 --> 00:01:51,667
we will see that the container has exited the node

31
00:01:51,767 --> 00:01:55,801
and is now in quite dead state. Now let's remove

32
00:01:55,801 --> 00:01:58,034
this node from cluster whatsoever.

33
00:01:58,734 --> 00:02:03,067
And when we try to do that, we get this error from Docker daemon.

34
00:02:03,334 --> 00:02:08,067
The reason behind that is, the node might be in the

35
00:02:08,067 --> 00:02:11,401
drain condition, but it is still up. Docker is

36
00:02:11,401 --> 00:02:15,201
still serving its API. So we need to make sure

37
00:02:15,201 --> 00:02:18,467
that it leaves the Swarm cluster first, then it

38
00:02:18,467 --> 00:02:21,867
gets removed from the master's list. Let's use

39
00:02:21,867 --> 00:02:25,201
docker swarm leave command from worker-2 node.

40
00:02:25,201 --> 00:02:29,334
Once we do so, we get a pretty clear nod that Node

41
00:02:29,334 --> 00:02:32,267
has left the swarm. If we try to run the same

42
00:02:32,267 --> 00:02:35,601
command again on manager node, we will see that

43
00:02:35,601 --> 00:02:38,434
the worker-2 node has been removed successfully.

44
00:02:38,501 --> 00:02:41,667
We can verify it by listing the nodes again and

45
00:02:41,667 --> 00:02:44,667
what we will find is our cluster made of just two

46
00:02:44,667 --> 00:02:47,167
nodes, manager and worker-1.