solo.io Gloo Platform, 8 Aug 2023

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: Simplify Istio Lifecycle on Multiple Clusters with Gloo Platform

Description

No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).

A

Foreign [music] we're going to deploy istio in the different workload clusters using glue's.

A

Lifecycle manager features if istio is already installed on the cluster. Glue can discover it, but.

A

Glue has the capability of automating the life cycle of istio. So let's check out that feature.

A

We can let glue mesh, deploy everything you need, but in this case in this lab we're going.

A

To follow a production-ready pattern of keeping the ingress gateway service objects separate. We.

A

Do this, because, most of the time when users deploy istio and clusters on the cloud, they use a service.

A

Of type load balancer for gateways, which creates a hostname or ip address for you automatically.

A

And you want to be able to reuse that hostname or ip, even if you jump between different versions of.

A

Istio, so if I start with istio 116, I want to go to 117, I don't want to create a new.

A

Load balancer, when I upgrade so that's why it's a best practice to create these service objects.

A

Up front and put all the annotations that you need based on the service provider, that you're.

A

Using and then we'll see later that we, when we deploy istio using glue lifecycle management, we.

A

Can tell glue mesh that we want to use these custom services that we've created manually.

A

So we'll start by creating this namespace called listio gateways where we're going to deploy the.

A

The istio ingress gateway for north south traffic, as well as for cross-cluster communication.

A

Again, you could create everything in one namespace like it's your system, but we'll try to follow a.

A

Best production ready best practice so we'll create one service for istio, ingress gateway.

A

For the north south traffic, that's going to be sent, that's going to be listening on 443 and 80.

A

And then we do that also for another gateway, um another service for the gateway for the.

A

Cross-Cluster communication that will use these different boards so we'll do that for.

A

Cluster one- and we do that for cluster two, so basically create two services in each cluster.

A

And in this example, we're using metall b. But if you're using kubernetes from a.

A

Cloud provider, they typically generate the creation of the cloud load balancers. So now.

A

If I go look at this my services here, I can see that my ingress gateway service has an.

A

External ip and then my east-west gateway service has this other ip.

A

Now we're ready to use the lifecycle manager object.

A

To tell glue mesh that you know what version of istio we wanted to install and in which cluster.

A

You could have one life cycle manager where we define multiple clusters, which is good if.

A

You want to upgrade all of your clusters at the same time or you can have one object per cluster.

A

It's really up to you. um We won't go through all the details of. What's here, you can see, there's a.

A

Lot of settings that we've added just because in this lab, we want to show all the various.

A

Features we'll do that in other videos. You know things like integrating virtual machines.

A

Etc, so that's why we have a little bit of extra option set of options here.

A

Note that this we're using this hub and tag which says that we want to we want to use, we don't want.

A

To use the default istio docker images, but instead we want to use the ones that are built by solo. So.

A

The solo images they're not a fork of istio they're, just images for each version of istio.

A

Where solo ads extra envoy filters, especially for things for providing extra capabilities.

A

Around gateway, there's also like custom builds of istio like fips version arms, build version. There's.

A

Fips distro-less different combinations, etc. So here we're just saying that we.

A

Want to deploy in cluster one, and we want to deploy istio 1164 using solo images.

A

And when you apply this, we're we're telling glue management plane to automate the deployment of.

A

The istio operator, so it's going to deploy an istio operator on the on the first cluster.

A

So you can see here this gm iop 116. So this is the istio operator. That's going to create the.

A

Istio operator objects and these objects would correspond to what we've just described in the.

A

In the istio lifecycle manager, so this should be very similar, so this is same as kind of what.

A

You said in the steel lifecycle manager object and then we can see that the operator itself.

A

Is going to deploy istio-d with the options that we've configured.

A

And we'll show in a different video, how we can do a proper upgrade like a canary.

A

Upgrade without any any disruption, but right now we're just deploying istio on cluster one.

A

And we will deploy the two gateways so for that we'll use another object, called gateway lifecycle.

A

Manager object where we can say we want to use this. The istio 116 revision that we just deployed.

A

And you can see, we want to use that custom service that we created in the very first step.

A

And this is just tells them that we want to reuse that service, that we created.

A

And again, we're using a different gateway lifecycle management for ingress of cluster.

A

One and then we'll and then the east west of cluster one as well, but you could really do it.

A

um In whichever way you prefer, you can have one for both of them. If you wanted.

A

So next we're going to create these objects and the operator is already there so we're just.

A

Going to apply all of these, these new istio operator objects and the operator is going.

A

To deploy the gateways on our behalf, so we can see we already have the ingress gateway deployed.

A

And the east west gateway is deployed as well.

A

So we just did cluster one, so we'll do exactly the same for cluster two, so we deploy istio d.

A

And ingress gateway and the east, west gateway and everything's declarative. We don't need to.

A

Worry about the order of applying things you can just create the gateway lifecycle manager before.

A

Or whatever everything's going to be reconciled at the end, so just apply now, which you can just.

A

Apply everything at once and very quickly. We should see the operator come up, istio-d come.

A

Up and all the gateways etc, and if we switch to the ui, you can also see that things are being.

A

Discovered so you can see it discovered, istio 1164 that we just deployed. It's got solo images.

A

And now the same for the second cluster, so that means everything should be ready to go.

A

We also defined some extra environment variables that we're going to use in.

A

In the later labs and videos where, when we want to access.

A

The ingress gateway when we want to do different tests, access the services, etc.

A

So I mean these: these environment variables correspond to the services.

A

That we just created so it's a way for us to get the ip address of the of the the ingress.

A

Gateway of cluster, one, for example, we're going to assign that to a to a variable.

A

And we'll also set the port for 4443.