WEBVTT

00:00.900 --> 00:01.733
-: In this section,

00:01.733 --> 00:04.050
we're going to very quickly put the configuration together

00:04.050 --> 00:06.390
for our multi worker deployment.

00:06.390 --> 00:08.340
After that, we're going to make sure

00:08.340 --> 00:10.260
that we take all of our different config files

00:10.260 --> 00:11.093
that we've put together,

00:11.093 --> 00:13.560
and throw them all back into Cube CTL,

00:13.560 --> 00:14.760
through that Apply command,

00:14.760 --> 00:16.380
and just get kind of a general idea

00:16.380 --> 00:18.300
of where all of our work is at.

00:18.300 --> 00:19.860
So, let's put the configuration together

00:19.860 --> 00:21.690
for the multi worker pod.

00:21.690 --> 00:24.690
As you might guess, this file is gonna look very similar

00:24.690 --> 00:26.130
to the config that we put together

00:26.130 --> 00:27.977
for the other two deployments.

00:27.977 --> 00:30.210
So inside my K8's directory,

00:30.210 --> 00:31.590
I'm going to make a new file

00:31.590 --> 00:33.427
called WorkerDeployment.eml.

00:38.864 --> 00:41.130
Then inside of here, we'll put together config very similar

00:41.130 --> 00:42.480
to the other deployments.

00:42.480 --> 00:46.500
We'll say API version is going to be apps view one.

00:46.500 --> 00:50.130
The kind of object that we are creating is a deployment.

00:50.130 --> 00:51.660
We'll give a metadata section

00:51.660 --> 00:55.533
to this thing with a name of worker-deployment,

00:56.910 --> 00:59.700
and then we'll provide a spec to this thing, as well.

00:59.700 --> 01:00.870
So for replicas,

01:00.870 --> 01:03.120
we're gonna start off with just as single replica,

01:03.120 --> 01:06.690
or a single pod running the multi worker image.

01:06.690 --> 01:08.550
Remember that if there's anything inside

01:08.550 --> 01:10.710
of our application that really needs to scale out

01:10.710 --> 01:11.985
and have multiple instances,

01:11.985 --> 01:13.769
it probably is the multi worker

01:13.769 --> 01:16.650
because that was the image or the container inside

01:16.650 --> 01:19.740
of our application that was doing that fibonacci calculation

01:19.740 --> 01:22.110
and that was the absolutely slowest part

01:22.110 --> 01:23.790
of our entire application.

01:23.790 --> 01:25.950
And so at one point in time, or at some point in time,

01:25.950 --> 01:28.440
we might want to come back and put a little bit more work

01:28.440 --> 01:31.050
into that thing to make sure that it is scaled properly.

01:31.050 --> 01:32.610
But for right now, we're gonna start off with just

01:32.610 --> 01:35.670
a single replica, so that we can eventually see that,

01:35.670 --> 01:37.980
yeah, this thing needs to be scaled and we can figure

01:37.980 --> 01:40.030
out how to do that at that point in time.

01:41.430 --> 01:44.080
All right, so after that, we'll provide our selector.

01:45.180 --> 01:47.680
We'll say that we want to attempt to match a label

01:49.455 --> 01:50.363
on some other object inside of our cluster.

01:53.183 --> 01:54.283
And for this one, we'll give a component of worker

01:54.283 --> 01:56.152
because that's really its job.

01:56.152 --> 01:58.470
It is a worker container inside of our application.

01:58.470 --> 02:00.390
After that, we'll put together our template

02:00.390 --> 02:01.680
for the pods that are created.

02:01.680 --> 02:03.360
Remember that the template property goes

02:03.360 --> 02:05.823
on the same indentation layer as selector.

02:07.560 --> 02:09.336
We'll provide a metadata to the template,

02:09.336 --> 02:11.640
where we'll provide the different labels

02:11.640 --> 02:14.850
that need to be applied to every pod that gets created.

02:14.850 --> 02:16.650
And of course, we're going to provide a label

02:16.650 --> 02:18.750
that matches the selector

02:18.750 --> 02:20.610
that we had specified just a second ago.

02:20.610 --> 02:24.093
So under labels, I'll say component worker.

02:27.060 --> 02:30.420
After that, I'll un-indent back out to the metadata section

02:30.420 --> 02:33.600
or metadata kinda indentation layer, I suppose,

02:33.600 --> 02:36.360
and will provide a spec that's going to configure exactly

02:36.360 --> 02:38.460
what this pod is running.

02:38.460 --> 02:39.510
So we're going to say

02:40.637 --> 02:41.970
that we want to run a single container,

02:41.970 --> 02:45.180
but remember it is plural containers, like so.

02:45.180 --> 02:47.905
And then the single container that we want to run

02:47.905 --> 02:50.220
is going to have a name of worker.

02:50.220 --> 02:55.187
It's going to use the image of your Docker ID, multi worker,

02:55.187 --> 02:57.240
and then just as before,

02:57.240 --> 02:59.760
we are going to eventually need to stuff in

02:59.760 --> 03:01.800
some environment variables into this thing

03:01.800 --> 03:03.780
to tell it how to connect to Redis.

03:03.780 --> 03:04.620
But for right now,

03:04.620 --> 03:06.930
we're just going to kind of ignore that for a second.

03:06.930 --> 03:08.220
We're going to first set up Redis

03:08.220 --> 03:09.720
and make sure that's working correctly.

03:09.720 --> 03:10.890
And then we're going to come back

03:10.890 --> 03:14.280
to both our worker and express API deployments

03:14.280 --> 03:17.220
and add in the environment variables at that point in time.

03:17.220 --> 03:19.200
Now, the other thing I wanna mention here is,

03:19.200 --> 03:21.242
if you look at this diagram,

03:21.242 --> 03:23.850
remember there is nothing inside of our worker image

03:23.850 --> 03:26.100
or the worker container that's running here

03:26.100 --> 03:27.330
that needs to be accessible

03:27.330 --> 03:30.270
from anything else inside of our cluster.

03:30.270 --> 03:33.510
In other words, there's no other object whatsoever,

03:33.510 --> 03:36.750
no other service, no other image, no container, no nothing,

03:36.750 --> 03:39.480
that needs to directly connect to the multi worker

03:39.480 --> 03:41.940
and try to get some information out of it.

03:41.940 --> 03:44.477
The multi worker is going to connect

03:44.477 --> 03:45.810
to something else inside of our application,

03:45.810 --> 03:48.420
but nothing is going to make a unprompted request

03:48.420 --> 03:50.400
into the multi worker.

03:50.400 --> 03:52.800
And as such, the multi worker does not need

03:52.800 --> 03:54.660
to have any port assigned to it,

03:54.660 --> 03:55.800
and it does not need to have

03:55.800 --> 03:58.140
any service assigned to it either.

03:58.140 --> 03:59.850
We only make use of services

03:59.850 --> 04:01.890
when we want to have requests going

04:01.890 --> 04:05.940
into a set of pods or into a single pod for that matter.

04:05.940 --> 04:08.670
And because that's not the case here, no need for a service,

04:08.670 --> 04:10.533
no need to worry about any ports.

04:11.940 --> 04:13.092
Okay.

04:13.092 --> 04:14.238
So with that in mind,

04:14.238 --> 04:15.647
that's pretty much it for this file.

04:15.647 --> 04:16.480
No ports required.

04:16.480 --> 04:18.030
So we're just going to finish it off like so.

04:18.030 --> 04:22.380
And we do not need to create a associated service file.

04:22.380 --> 04:24.060
So let's save this as is,

04:24.060 --> 04:25.350
we'll take a quick pause right here,

04:25.350 --> 04:26.790
and when we come back to the next section,

04:26.790 --> 04:29.190
we're going to take the three config files

04:29.190 --> 04:30.900
that we just put together in the last couple videos,

04:30.900 --> 04:35.116
and throw them into Cube CTL through the Apply command.

04:35.116 --> 04:36.623
So quick break and I'll see you in just a minute.