1

00:00:04,180 --> 00:00:07,150

Hello and welcome to this lecture in this lecture.



2

00:00:07,180 --> 00:00:10,000

we will talk about Kube Scheduler.



3

00:00:10,000 --> 00:00:15,540

Earlier we discussed that the kubernetes scheduler is responsible for scheduling pods on nodes.



4

00:00:16,300 --> 00:00:18,640

Now don't let the graphic mislead you.



5

00:00:18,670 --> 00:00:26,710

remember the scheduler is only responsible for deciding which pod goes on which node. It doesn’t actually



6

00:00:26,710 --> 00:00:29,010

place the pod on the nodes.



7

00:00:29,050 --> 00:00:34,690

That’s the job of the kubelet. The kubelet or the captain on the ship is who creates the pod on the



8

00:00:34,690 --> 00:00:35,610

ships.



9

00:00:35,610 --> 00:00:39,860

The scheduler only decides which pod goes where.



10

00:00:40,000 --> 00:00:43,270

Let's look at how the scheduler does that in a bit more detail.



11

00:00:43,420 --> 00:00:48,770

First of all why do you need a scheduler? When there are many ships and many containers,



12

00:00:48,940 --> 00:00:53,450

You want to make sure that the right container ends up on the right ship.



13

00:00:53,590 --> 00:00:57,670

For example there could be different sizes of ships and containers.



14

00:00:57,670 --> 00:01:04,130

You want to make sure the ship has sufficient capacity to accommodate those containers different ships



15

00:01:04,150 --> 00:01:06,280

maybe going to different destinations.



16

00:01:06,280 --> 00:01:11,740

You want to make sure your containers are placed on the right ships so they end up in the right destination



17

00:01:12,540 --> 00:01:17,170

In kubernetes, the scheduler decides which nodes the pods are placed on



18

00:01:17,170 --> 00:01:22,660

depending on certain criteria.  You may have PODs with different resource requirements,



19

00:01:22,720 --> 00:01:27,180

You can have nodes in the cluster dedicated to certain applications.



20

00:01:27,280 --> 00:01:33,700

So how does the scheduler assign these PODs? The scheduler looks at each POD and tries to find the



21

00:01:33,700 --> 00:01:35,280

best node for it.



22

00:01:35,290 --> 00:01:37,560

For example, let’s take one of these PODs.



23

00:01:37,810 --> 00:01:41,970

The big blue one. It has a set of CPU and Memory requirements.



24

00:01:42,040 --> 00:01:47,980

The scheduler goes through two phases to identify the best node for the pod in the first phase.



25

00:01:48,010 --> 00:01:52,940

The scheduler tries to filter out the nodes that do not fit the profile for this pod.



26

00:01:53,050 --> 00:01:59,830

For example, the nodes that do not have sufficient CPU and memory resources requested by the pod.



27

00:01:59,860 --> 00:02:06,100

So the first two small nodes are filtered out.  So we are now left with two nodes on which the POD



28

00:02:06,130 --> 00:02:07,870

can be placed.



29

00:02:07,900 --> 00:02:14,660

Now how does the scheduler pick one from the two the scheduler ranks the node to identify the best fit



30

00:02:14,660 --> 00:02:15,580

for the pod.



31

00:02:15,620 --> 00:02:21,240

It uses a priority function to assign a score to the nodes on a scale of 0 to 10.



32

00:02:21,260 --> 00:02:28,130

For example the scheduler calculates the amount of resources that would be free on the nodes after placing



33

00:02:28,130 --> 00:02:29,570

the pod on them.



34

00:02:29,570 --> 00:02:36,230

In this case, the one on the right would have 6 CPUs free  if the pod was placed on it which is 4 more



35

00:02:36,560 --> 00:02:39,430

than the other one so it gets a better rank.



36

00:02:39,530 --> 00:02:40,820

And so it wins.



37

00:02:40,910 --> 00:02:43,220

So that's how a scheduler works at a high level.



38

00:02:43,280 --> 00:02:48,380

And of course these can be customized and you can write your own scheduler as well.



39

00:02:48,380 --> 00:02:53,660

There are many more topics to look at such as resource requirements, limits, taints and tolerations,



40

00:02:53,690 --> 00:03:00,290

node selectors, affinity rules etc. Which is why we have an entire section dedicated to scheduling coming



41

00:03:00,290 --> 00:03:04,940

up in this course where we will discuss each of these in much more detail.



42

00:03:04,940 --> 00:03:10,040

For now we will continue to focus on the scheduler as a process at a high level.



43

00:03:10,760 --> 00:03:13,160

So how do you install the kube-scheduler?



44

00:03:13,160 --> 00:03:18,740

Download the kube-scheduler from the kubernetes release page. Extract it and run it as a service.



45

00:03:19,220 --> 00:03:20,600

when you run it as a service.



46

00:03:20,600 --> 00:03:23,090

You specify the scheduler configuration file.



47

00:03:23,570 --> 00:03:26,240

So how do you view the kube-scheduler server options?



48

00:03:26,240 --> 00:03:31,820

Again, if you set it up with kubeadm tool, kubeadm deploys the kube-scheduler as a pod



49

00:03:31,820 --> 00:03:33,450

in the kube-system namespace



50

00:03:33,470 --> 00:03:39,030

on the master node. You can see the options within the pod definition file located at etc



51

00:03:39,090 --> 00:03:41,270

kubernetes manifests folder.



52

00:03:41,420 --> 00:03:46,280

You can also see the running process and the effective options by listing the process on the master



53

00:03:46,280 --> 00:03:49,250

node and searching for kube-scheduler.



54

00:03:49,250 --> 00:03:51,020

Well that's it for this lecture.



55

00:03:51,200 --> 00:03:52,550

I will see you in the next.