1
00:00:00,401 --> 00:00:03,867
Kubernetes provides container lifecycle hooks to

2
00:00:03,867 --> 00:00:07,267
trigger commands on Container lifecycle events.

3
00:00:07,267 --> 00:00:11,301
If we recall, container lifecycle had 5 stages,

4
00:00:11,301 --> 00:00:17,034
Created, Running, Paused, Stopped and Deleted.

5
00:00:17,034 --> 00:00:20,767
Out of these 5, kubectl provides lifecycle hooks

6
00:00:20,767 --> 00:00:25,234
for two of the states, which are Created and Stopped.

7
00:00:26,301 --> 00:00:28,234
Let's explore both of these using

8
00:00:28,234 --> 00:00:32,433
lifecycle-pod.yaml file. This is a standard

9
00:00:32,433 --> 00:00:36,067
nginx Pod named lifecyc-pod. And under the

10
00:00:36,067 --> 00:00:39,567
containers spec, we have two lifecycle hooks called

11
00:00:39,567 --> 00:00:43,201
postStart and preStop. These hooks functionality

12
00:00:43,201 --> 00:00:46,467
is pretty much as their name suggests. Both of

13
00:00:46,467 --> 00:00:48,867
them have handlers attached to them, which are

14
00:00:48,867 --> 00:00:52,334
executable commands, postStart hooks handler will

15
00:00:52,334 --> 00:00:57,167
'echo Welcome' to a file called poststart-msg and it

16
00:00:57,167 --> 00:00:59,734
will trigger after the container will enter the

17
00:00:59,734 --> 00:01:02,834
Created state. This is the state where resources

18
00:01:02,834 --> 00:01:05,667
for read-write layer are reserved, but the

19
00:01:05,667 --> 00:01:09,267
container is not running yet. In other words, the

20
00:01:09,267 --> 00:01:13,301
latest cmd or entrypoint instruction is yet to be

21
00:01:13,301 --> 00:01:16,001
executed. The hook works synchronously with the

22
00:01:16,001 --> 00:01:19,701
Pod's container creation process. Which means if

23
00:01:19,701 --> 00:01:22,934
for some reason the handler of the hook hangs or

24
00:01:22,934 --> 00:01:26,301
fails to execute, the pod will remain in container

25
00:01:26,301 --> 00:01:29,467
Created state and won't go into the Running state.

26
00:01:29,867 --> 00:01:32,567
To brief things up, first of all the container will

27
00:01:32,567 --> 00:01:35,001
be created, then the postStart hook will be

28
00:01:35,001 --> 00:01:38,501
handled, and the message will be printed. And then

29
00:01:38,501 --> 00:01:41,401
the container will start running by executing cmd

30
00:01:41,401 --> 00:01:44,667
or entrypoint command. A general use of this hook

31
00:01:44,667 --> 00:01:47,567
is for better debugging, just like try and catch

32
00:01:47,567 --> 00:01:51,101
clause in the programming. But it also brings the

33
00:01:51,101 --> 00:01:53,967
burden of stalling the container if the hook

34
00:01:53,967 --> 00:01:57,634
doesn't get handled properly. So if Pod events and

35
00:01:57,634 --> 00:02:00,534
logs are sufficient for your debugging, you might

36
00:02:00,534 --> 00:02:04,801
want to skip using this hook. Lastly, we have preStop hook,

37
00:02:04,801 --> 00:02:07,234
which triggers before termination of the

38
00:02:07,234 --> 00:02:10,001
container. We are simply quitting the nginx

39
00:02:10,001 --> 00:02:12,866
process before terminating the container. But if

40
00:02:12,866 --> 00:02:15,301
you want to strongly verify this hook, you can

41
00:02:15,301 --> 00:02:17,734
apply SIGKILL to one of the container's crucial

42
00:02:17,734 --> 00:02:20,667
processes and you would find the container exited

43
00:02:20,667 --> 00:02:23,867
with a respective error code. Let's exit the file

44
00:02:23,867 --> 00:02:25,234
and create the Pod.

45
00:02:25,234 --> 00:02:29,034
[No Audio]

46
00:02:29,034 --> 00:02:31,601
30 seconds down and the Pod is ready.

47
00:02:31,934 --> 00:02:33,801
I know we have sold the benefits of

48
00:02:33,801 --> 00:02:37,167
containers a lot, but it is always amusing to see

49
00:02:37,167 --> 00:02:39,934
such a level of managed isolation being created

50
00:02:39,934 --> 00:02:43,167
with such less efforts and within such a short time.

51
00:02:43,267 --> 00:02:46,134
Now let's execute the Pod with kubectl exec

52
00:02:46,134 --> 00:02:51,401
command and run bash on it. cat the file poststart-msg

53
00:02:51,401 --> 00:02:55,234
and bingo the hook was executed successfully.

54
00:02:55,434 --> 00:02:59,367
The message is loud and clear. Well not that loud,

55
00:02:59,467 --> 00:03:01,901
but it is quite clear. In the next lecture,

56
00:03:01,901 --> 00:03:04,767
we will see how to replace container cmd command