Load-balancer for virtual machines on bare metal Kubernetes clusters #837

RamLavi · 2022-04-03T14:23:22Z

What this PR does / why we need it:

Special notes for your reviewer:

kubevirt-bot · 2022-04-03T14:23:24Z

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

RamLavi · 2022-04-04T10:00:16Z

/hold

still needs a few adjustments, but structure wise this is sound

phoracek · 2022-04-04T10:46:17Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+## Introduction
+
+Over the last year, kubevirt has integrated with MetalLB in order to support fault-tolerant access to an application 


Suggested change

Over the last year, kubevirt has integrated with MetalLB in order to support fault-tolerant access to an application

Over the last year, KubeVirt has integrated with MetalLB in order to support fault-tolerant access to an application

I believe that when we refer to the KubeVirt project, we use the CapitalSnakeCase

phoracek · 2022-04-04T10:48:06Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## Introduction
+
+Over the last year, kubevirt has integrated with MetalLB in order to support fault-tolerant access to an application 
+through an external IP address. As a Cluster administrator using a bare-metal cluster, now it's possible to add MetalLB 


Is not the use-case for MetalLB on-premise clusters? Baremetal servers can be deployed in public cloud too.

@fedepaol please keep me honest here

Both "on premise" and "baremetal" may be non 100% correct (because there can be distributions deploy-able on premise with their own network loadbalancer), but on premise is probably better. If we want to be accurate, we can say "on premise without a load balancer", or "on premise clusters normally do not come with a network loadbalancer" or something along that line.

perhaps we need to be more accurate, but AFAIR according to metalLB page:

MetalLB is a load-balancer implementation for bare metal [Kubernetes](https://kubernetes.io/) clusters, using standard routing protocols.

So I think we're not miss-advertising here. no?

I actually did not see @fedepaol 's comment when I posted mine on top of it. I'll change it accordingly of course :)

phoracek · 2022-04-04T10:48:30Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## Introducing MetalLB
+
+[MetalLB](https://metallb.universe.tf/) allows you to create Kubernetes services of type `LoadBalancer`, and provides network load-balancer 
+implementation in clusters that don’t run on a cloud provider, such as bare-metals.


bare-metals can run on a cloud provider

Let's wait to see what @fedepaol says on your previous comment

phoracek · 2022-04-04T10:49:07Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+- Layer 2 mode (ARP/NDP)
+Layer 2 mode does not implement Load-balancing behavior provides a failover mechanism where one node owns the `LoadBalancer` service, until the node fails and 
+another node is chosen. In a strict sense, this is not a real load-balancing behavior, but instead it makes the IPs reachable to 
+the local network. This method announces the IPs in ARP (for IPV4) and NDP (for IPV6) protocols over the network. From a 


Suggested change

the local network. This method announces the IPs in ARP (for IPV4) and NDP (for IPV6) protocols over the network. From a

the local network. This method announces the IPs in ARP (for IPv4) and NDP (for IPv6) protocols over the network. From a

phoracek · 2022-04-04T10:49:43Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+the local network. This method announces the IPs in ARP (for IPV4) and NDP (for IPV6) protocols over the network. From a 
+network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.
+
+- BGP mode


This renders as a single paragraph "BGP mode BGP mode provides..."

yep I'll change it.. thanks!

phoracek · 2022-04-04T10:50:33Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## Demo: VirtualMachine with external IP and MetalLB load balancer
+
+With the following recipe we will end up with a nginx server running on a virtualMachine, accessible outside the cluster 
+using MetalLB loadBalancer.


Suggested change

using MetalLB loadBalancer.

using MetalLB load balancer.

phoracek · 2022-04-04T10:50:59Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+### Demo environment setup
+
+We are going to use [kind](https://kind.sigs.k8s.io) provider as an ephemeral Kubernetes cluster.
+To start it up follow this [guide](https://kind.sigs.k8s.io/docs/user/quick-start)


Suggested change

To start it up follow this [guide](https://kind.sigs.k8s.io/docs/user/quick-start)

To start it up follow this [guide](https://kind.sigs.k8s.io/docs/user/quick-start).

phoracek · 2022-04-04T10:52:14Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+### Demo environment setup
+
+We are going to use [kind](https://kind.sigs.k8s.io) provider as an ephemeral Kubernetes cluster.


That's a very long guide, do I need to follow the Advanced part? Or build my own images?

Please mention which parts of the quick start should I follow

phoracek · 2022-04-04T10:53:43Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+### Spin up a Virtual Machine running Nginx
+
+Now it's time to start-up a virtualMachine running nginx, using this yaml:


The colon is confusing here

phoracek · 2022-04-04T10:54:02Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+
+### Expose the virtualMachine with a typed LoaBalancer service 


Suggested change

### Expose the virtualMachine with a typed LoaBalancer service

### Expose the VirtualMachine with a typed LoaBalancer service

already changed to virtual machine

RamLavi · 2022-04-05T11:54:49Z

/hold cancel

maiqueb

Overall looks very good, and is easy to understand.

The section about the ARP configuration mode is a bit convoluted IMO, and I could not follow it.

maiqueb · 2022-04-05T13:53:47Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## Introducing MetalLB
+
+[MetalLB](https://metallb.universe.tf/) allows you to create Kubernetes services of type `LoadBalancer`, and provides network load-balancer implementation in clusters that don’t run on a cloud provider, such as bare-metals.
+MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 working modes: Layer 2 and BGP modes:


Suggested change

MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 working modes: Layer 2 and BGP modes:

MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 modes: Layer 2 and BGP:

maiqueb · 2022-04-05T14:09:44Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+This mode does not implement Load-balancing behavior provides a failover mechanism where one node owns the `LoadBalancer` service, until the node fails and another node is chosen. In a strict sense, this is not a real load-balancing behavior, but instead it makes the IPs reachable to the local network. This method announces the IPs in ARP (for Ipv4) and NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.
+
+- BGP mode
+This mode provides a real load-balancing behavior, By establishing BGP peering sessions with the network routers. They in turn advertise the external IPs of the `LoadBalancer` service, distributing the load over the nodes.


Suggested change

This mode provides a real load-balancing behavior, By establishing BGP peering sessions with the network routers. They in turn advertise the external IPs of the `LoadBalancer` service, distributing the load over the nodes.

This mode provides real load-balancing behavior, by establishing BGP peering sessions with the network routers - which advertise the external IPs of the `LoadBalancer` service, distributing the load over the nodes.

maiqueb · 2022-04-05T14:11:31Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+#### Installing MetalLB on the cluster
+
+There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator). Note
+that you will need `cluster-admin` privileges.


Would it make sense to have a section before with prerequirements ? Having this privilege would be one of them.

maiqueb · 2022-04-05T14:15:44Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator). Note
+that you will need `cluster-admin` privileges.
+You can confirm the operator is installed by entering the following command:


I think it would be preferable to block until it is in the correct phase, something like:

kubectl wait -nmy-metallb-operator csv metallb-operator.v0.12.0 \ --for=jsonpath='{.status.phase}'=Succeeded --timeout=2m

I actually think both options are not so great. The reason is that after installing the MetalLB operator, it takes a few moments for the csv to appear, so you'll get

$ kubectl wait -nmy-metallb-operator csv metallb-operator.v0.12.0 --for=jsonpath='{.status.phase}'=Succeeded --timeout=2m Error from server (NotFound): clusterserviceversions.operators.coreos.com "metallb-operator.v0.12.0" not found

Moreover, your suggestion will break as soon as they do a new operator release.

I think we should keep it, as the confirmation command kubectl get csv -n my-metallb-operator I mentioned actually comes from the MetalLB installation guide, so I figure they know best..

Moreover, your suggestion will break as soon as they do a new operator release.

Using a variable would address that - and still, having the version of something as part of its name is .... not a smart move.

I think we should keep it, as the confirmation command kubectl get csv -n my-metallb-operator I mentioned actually comes from the MetalLB installation guide, so I figure they know best..

I'm just trying to make the reader's life easier: use a blocking command that unblocks when state == desiredState is more comfortable than repeating a command until the reader sees what's expected, but feel free to discard this opinion; I won't insist.

Moreover, your suggestion will break as soon as they do a new operator release.

Using a variable would address that - and still, having the version of something as part of its name is .... not a smart move.

I think we should keep it, as the confirmation command kubectl get csv -n my-metallb-operator I mentioned actually comes from the MetalLB installation guide, so I figure they know best..

I'm just trying to make the reader's life easier: use a blocking command that unblocks when state == desiredState is more comfortable than repeating a command until the reader sees what's expected, but feel free to discard this opinion; I won't insist.

using the version is indeed... a choice... but it's not in our hands..
I do agree that wait is cooler, but still prefer the "get csv" option in this particular case. I will add a line that it may take a few seconds for the CSV to appear. Hope you find it ok by you.

maiqueb · 2022-04-05T14:17:15Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+> Notes:
+> - If you're running a bare-metal cluster in a colocation factory, you need to first reserve this IP Address pool from your hosting provider. Alternatively, if you're running on a purely private cluster, you can use one of the private IP Address spaces (a.k.a RFC1918 addresses). Such addresses are free, and work fine as long as you’re only providing cluster services to your LAN.


What is a colocation factory ? I would expect an explanation along the term whenever it is introduced.

I would also drop the word purely. It adds nothing imo.

it's basically the opposite of a private cluster you spin up on a machine and you have controll over the entire network segment.
colocation factory means you share the network with others (clusters, hosts, etc..), and so you need to approach your (factory) network admin to give you a lease of IPs for metalLB to use.

An example of this is our cnv.eng cluster - you can't just grab a range of IPs and call it yours, you need to ask for a lease otherwise the DHCP can give it to one of the machines, and you'll be in trouble =) .

I rephrased and moved to prerequirements, tell me if it's more understandable now

maiqueb · 2022-04-05T14:20:56Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+              #!/bin/bash
+              echo "fedora" |passwd fedora --stdin
+              sudo yum install -y nginx
+              sudo systemctl enable nginx
+              sudo systemctl start nginx


I would prefer to see a cloud config instead of a bash script snippet:

Suggested change

#!/bin/bash

echo "fedora" |passwd fedora --stdin

sudo yum install -y nginx

sudo systemctl enable nginx

sudo systemctl start nginx

#cloud-config

password: fedora

chpasswd: { expire: False }

packages:

- nginx

runcmd:

- [ "systemctl", "enable", "--now", "nginx" ]

maiqueb · 2022-04-05T14:23:13Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+  allocateLoadBalancerNodePorts: true
+  type: LoadBalancer
+  ipFamilyPolicy: SingleStack
+  sessionAffinity: None


nit: This is the default; it can be omitted.

only sessionAffinity?

allocateLoadBalancerNodePorts also defaults to true: https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.22/#service-v1-core

If I were you, I'd check all of them are not the defaults.

ack, thanks!

maiqueb · 2022-04-05T14:48:58Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 working modes: Layer 2 and BGP modes:
+
+- Layer 2 mode (ARP/NDP)
+This mode does not implement Load-balancing behavior provides a failover mechanism where one node owns the `LoadBalancer` service, until the node fails and another node is chosen. In a strict sense, this is not a real load-balancing behavior, but instead it makes the IPs reachable to the local network. This method announces the IPs in ARP (for Ipv4) and NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.


Suggested change

This mode does not implement Load-balancing behavior provides a failover mechanism where one node owns the `LoadBalancer` service, until the node fails and another node is chosen. In a strict sense, this is not a real load-balancing behavior, but instead it makes the IPs reachable to the local network. This method announces the IPs in ARP (for Ipv4) and NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.

This mode - which does not implement Load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as service owner. This configuration mode makes the IPs reachable from the local network.

This method announces the IPs using ARP (for Ipv4) or NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.

I am a bit confused about the second paragraph though ... I unfortunately cannot provide any assistance in re-writing it.

rephrased, tell me what you think.

maiqueb · 2022-04-06T16:20:42Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## Introduction
+
+Over the last year, Kubevirt has integrated with MetalLB in order to support fault-tolerant access to an application through an external IP address. 
+As a Cluster administrator using an on-prem cluster without a network load balancer, now it's possible to add MetalLB operator and gain a load-balancer capabilities (with Services of type LoadBalancer) to virtual machines.


Suggested change

As a Cluster administrator using an on-prem cluster without a network load balancer, now it's possible to add MetalLB operator and gain a load-balancer capabilities (with Services of type LoadBalancer) to virtual machines.

As a Cluster administrator using an on-prem cluster without a network load balancer, now it's possible to use MetalLB operator to provide load-balancer capabilities (with Services of type LoadBalancer) to virtual machines.

maiqueb · 2022-04-06T16:24:19Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+- You should have `cluster-admin` privileges on the cluster.
+- IP Address allocation: 
+  How you get IP address pools for MetalLB depends on your environment:


I find this sentence strangely positioned. Are you sure you want this to read:

IP Address allocation: How you get IP address pools for MetalLB depends on your environment: ?

rephrased. what do you think?

maiqueb · 2022-04-06T16:38:11Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator). Note
+that you will need `cluster-admin` privileges.
+You can confirm the operator is installed by entering the following command:


Moreover, your suggestion will break as soon as they do a new operator release.

Using a variable would address that - and still, having the version of something as part of its name is .... not a smart move.

I think we should keep it, as the confirmation command kubectl get csv -n my-metallb-operator I mentioned actually comes from the MetalLB installation guide, so I figure they know best..

I'm just trying to make the reader's life easier: use a blocking command that unblocks when state == desiredState is more comfortable than repeating a command until the reader sees what's expected, but feel free to discard this opinion; I won't insist.

maiqueb

Short. concise, and useful.

Thank you.

phoracek · 2022-04-11T12:02:45Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+- You should have `cluster-admin` privileges on the cluster.
+- Getting IP Address pools allocation for MetalLB depends on your environment:
+  - If you're running a bare-metal cluster in a shared host environment, you need to first reserve this IP Address pool from your hosting provider.
+  - Alternatively, if you're running on a private cluster, you can use one of the private IP Address spaces (a.k.a RFC1918 addresses). Such addresses are free, and work fine as long as you’re only providing cluster services to your LAN.


I tried following the guide and this is where I got stuck.

I followed the Installation section, so now I have kind. But now the blog post speaks about cluster-admin and allocated pool. Are these requirements for kind? How do I apply them? Why is it a subsection of "Demo environment setup".

People should by able to follow through the blog-post simply by copy-pasting examples, either from here or from some other documentation.

Please stick to what is required to get the kind demo cluster running here (including the note from below "ttps://kind.sigs.k8s.io/docs/user/loadbalancer/"). If these pre-requirements are for a real (non-kind) cluster, it should be clearly stated and outside "Demo environment setup"

You're right, these pre-requirements certainly do not fit here. Removed
Regarding the note - I will add only the necessary things needed for this demo.
Done.

phoracek · 2022-04-11T12:05:39Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+    - 172.18.1.1-172.18.1.16
+EOF
+```
+> Note: Since this demo is using a kind cluster, we want this range to be on the docker kind network. For more information, see [link](https://kind.sigs.k8s.io/docs/user/loadbalancer/)


This guide describes how to install MetalLB. It is confusing to list this under a section that waits for MetalLB to get installed

I agree, this deserves a proper section. Done

phoracek · 2022-05-02T09:00:01Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+### Demo environment setup
+
+We are going to use [kind](https://kind.sigs.k8s.io) provider as an ephemeral Kubernetes cluster.
+To start it up follow this [installation guide](https://kind.sigs.k8s.io/docs/user/quick-start/#installation).


This should be in the prerequirements section.

Also, could we change this to "First install kind on your machine following its installation guide". The installation is platform specific and may change, so it makes sense to refer an external source.

However, I would ask to list everything happening after installation here, starting with the commands starting a kind cluster.

If we split it like that, it would be clear "install kind however you want, then come back and we will prepare the environment"

I agree. Done.

phoracek · 2022-05-02T09:02:43Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+This demo runs Docker on Linux, which allows sending traffic directly to the load-balancer's external IP if the IP space is within the docker IP space.
+On macOS and Windows however, docker does not expose the docker network to the host, rendering the external IP unreachable from other kind nodes. In order to workaround this, one could expose pods and services using extra port mappings as shown in the extra port mappings section of kind's [Configuration Guide](https://kind.sigs.k8s.io/docs/user/configuration#extra-port-mappings).


Once the kind installation is moved to this section, could we move this paragraph to a another subsection "External IPs on macOS and Windows"?

yes, but I think it this subsection should remain under Prerequirement section.
Done.

phoracek · 2022-05-02T09:03:25Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+This demo runs Docker on Linux, which allows sending traffic directly to the load-balancer's external IP if the IP space is within the docker IP space.
+On macOS and Windows however, docker does not expose the docker network to the host, rendering the external IP unreachable from other kind nodes. In order to workaround this, one could expose pods and services using extra port mappings as shown in the extra port mappings section of kind's [Configuration Guide](https://kind.sigs.k8s.io/docs/user/configuration#extra-port-mappings).
+


Suggested change

### Deploying cluster

... How to start kind

phoracek · 2022-05-02T09:04:37Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+docker network inspect -f '{{.IPAM.Config}}' kind
+```
+
+You should get the subclass you can set the ip range from. The output should contain a cidr such as 172.18.1.0/16


Could you explain how this pool relates to my docker installation, and is this different on macOS, windows?

The pool uses the IP space provided by kind on docker, this is only relevant for Linux.
macOS, windows will need to set their own network, as mentioned on External IPs on macOS and Windows section.
I will rephrase to make that clearer, tell me what you think.

phoracek · 2022-05-02T09:05:02Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+#### Setting Address Pool to be used by the LoadBalancer
+
+In order to complete the Layer 2 mode configuration, we need to set a range of IP Addresses for the LoadBalancer to use. We want this range to be on the docker kind network, so by using this command:


Suggested change

In order to complete the Layer 2 mode configuration, we need to set a range of IP Addresses for the LoadBalancer to use. We want this range to be on the docker kind network, so by using this command:

In order to complete the Layer 2 mode configuration, we need to set a range of IP addresses for the LoadBalancer to use. We want this range to be on the docker kind network, so by using this command:

phoracek · 2022-05-02T09:05:12Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+docker network inspect -f '{{.IPAM.Config}}' kind
+```
+
+You should get the subclass you can set the ip range from. The output should contain a cidr such as 172.18.1.0/16


Suggested change

You should get the subclass you can set the ip range from. The output should contain a cidr such as 172.18.1.0/16

You should get the subclass you can set the IP range from. The output should contain a cidr such as 172.18.1.0/16

phoracek

Worked for me, some minor comments. I'm yet to re-review the text

phoracek · 2022-05-03T11:14:21Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+kind create cluster
+```
+
+Then to start using the cluster:


How is the following command using the cluster?

mmm.. well if you have more than 1 clusters, you need to choose the right one in order to interact with (see here)
let me rephrase it

phoracek · 2022-05-03T11:15:30Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+#### Installing MetalLB on the cluster
+
+There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator). 


Suggested change

There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator).

There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator) and click the Install button.

phoracek · 2022-05-03T11:17:45Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+#### Installing MetalLB on the cluster
+
+There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator). 


The OLM 0.21.1 installation command fails for me on a fresh kind cluster with

The CustomResourceDefinition "clusterserviceversions.operators.coreos.com" is invalid: metadata.annotations: Too long: must have at most 262144 bytes

strange.. lemme try installing again

Yes I see it. OLM has an open issue operator-framework/operator-lifecycle-manager#2767 on this.
well it's not blocking or anything, so I'm on the fence here. Do you think we should install MetalLB a different way?

I stand corrected, it's not really working.. I'll look for another way to install MetalLB

It failed for me on OLM 0.21.1, but 0.20.0 worked ok

Moved to installing via manifest.

phoracek · 2022-05-03T11:50:35Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+### Access the virtual machine from outside the cluster
+
+Finally, we can check that the nginx server is accessible from outside the cluster:


Could you add a note explaining that it may take a while before the VM becomes available? Or better, show how to wait for it to become available

I added a note, though it is not related to the vm being ready. It looks like the url simply takes some time to work after setting the service.
@fedepaol do you know why this happens? I wonder if there is something we can "wait" upon being ready and not simply wait fr the url to work..

phoracek · 2022-05-03T11:51:02Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+title: Load-balancer for virtual machines on bare metal Kubernetes clusters
+description: This post illustrates setting up a virtual machine with MetalLb loadBalance service.
+navbar_active: Blogs
+pub-date: April 03


Let's update this, in case we forget to update it later before publishing

phoracek · 2022-05-04T07:54:38Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+layout: post
+author: Ram Lavi
+title: Load-balancer for virtual machines on bare metal Kubernetes clusters
+description: This post illustrates setting up a virtual machine with MetalLb loadBalance service.


Suggested change

description: This post illustrates setting up a virtual machine with MetalLb loadBalance service.

description: This post illustrates setting up a virtual machine with MetalLB loadBalance service.

phoracek · 2022-05-04T08:03:18Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+##### External IPs on macOS and Windows
+
+This demo runs Docker on Linux, which allows sending traffic directly to the load-balancer's external IP if the IP space is within the docker IP space.
+On macOS and Windows however, docker does not expose the docker network to the host, rendering the external IP unreachable from other kind nodes. In order to workaround this, one could expose pods and services using extra port mappings as shown in the extra port mappings section of kind's [Configuration Guide](https://kind.sigs.k8s.io/docs/user/configuration#extra-port-mappings).


How would I then configure the address range in the MetalLB ConfigMap?

in short - the same.

This section deals not with the MetalLB ConfigMap, but with IP choosing considerations.
It's basically saying, if you're not using docker on linux, then you need to configure yourself the ip range is reachable from your pc console.
But once you made sure that IP is reachable, the configuring MetalLB ConfigMap is the same..

phoracek · 2022-05-04T08:04:19Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+kind create cluster
+```
+
+In order to interact with the specific cluster created:


Sorry, I still don't understand what are you suggesting here and how is calling cluster-info useful in the remainder of the post

I think what's confusing is kind's terminology for clusters.
when using kind, you may have more than 1 clusters to use:

$ kind get clusters kind kind-2

with kubectl cluster-info --context kind-kind you choose to work with the specific cluster you want kubectl to point to..
it's explained here

I see, so you are calling this to make sure that in case somebody changed the kind context before running this demo, we change it back to the implicit default name "kind-kind". Should set a specific context name while calling kind create cluster to make the connection clear?

phoracek · 2022-05-04T08:06:33Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+docker network inspect -f '{{.IPAM.Config}}' kind
+```
+
+You should get the subclass you can set the IP range from. The output should contain a cidr such as 172.18.0.0/16


Suggested change

You should get the subclass you can set the IP range from. The output should contain a cidr such as 172.18.0.0/16

You should get the subclass you can set the IP range from. The output should contain a cidr such as 172.18.0.0/16.

phoracek · 2022-05-04T08:09:30Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+
+#### Installing Kubevirt on the cluster


Could you move this under "Installing components"? That way we would split the post into:

Intro

Prerequisites

Cluster deployment

Components deployment

Network resources configuration

Network utilization

So it would run from the lowest layer to the highest.

not sure I understand you. #### Installing Kubevirt on the cluster is under ### Installing components, isn't it?

Nevermind. Not sure what I meant here

I think I do:

Cluster deployment: this one is creating the kind cluster

Components deployment: this one is installing metalLB and KubeVirt in the cluster

Network resources configuration: this one is configuring the metalLB pool

Network utilization: this one is creating the VM and service

Does it make sense ? I share the concern @phoracek 's mentioned in https://github.com/kubevirt/kubevirt.github.io/pull/837/files#r864557203 ... That is part of the cluster configuration, and probably deserves to live in a separate section.

I like it. Done.
Let me know what you think

phoracek · 2022-05-04T08:09:51Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+kubectl get pods -n metallb-system --watch
+```
+
+#### Setting Address Pool to be used by the LoadBalancer


This should not be a part of component installation

I agree, but the problem is that then the nesting would be ##### Setting Address Pool to be used by the LoadBalancer which is not supported CSS style wise..

phoracek · 2022-05-04T08:12:16Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+Over the last year, Kubevirt and MetalLB have shown to be powerful duo in order to support fault-tolerant access to an application on virtual machines through an external IP address. 
+As a Cluster administrator using an on-prem cluster without a network load-balancer, now it's possible to use MetalLB operator to provide load-balancer capabilities (with Services of type `LoadBalancer`) to virtual machines.
+
+## Introducing MetalLB


Suggested change

## Introducing MetalLB

## MetalLB

The headline above already says "Introduction", saying it again (despite this is not nested) seems redundant

phoracek · 2022-05-04T08:14:21Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+- First install kind on your machine following its [installation guide](https://kind.sigs.k8s.io/docs/user/quick-start/#installation).
+- To use kind, you will also need to [install docker](https://docs.docker.com/install/).
+
+##### External IPs on macOS and Windows


This is so nested that our website does not have a CSS style defined for it. What about removing the prerequirements headline and just keeping its list under the section above. Then you can remove one level of nesting on this one

phoracek

Just some typos. I'm ok with this PR once these are resolved

phoracek · 2022-05-11T08:27:45Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+layout: post
+author: Ram Lavi
+title: Load-balancer for virtual machines on bare metal Kubernetes clusters
+description: This post illustrates setting up a virtual machine with MetalLB loadBalance service.


Suggested change

description: This post illustrates setting up a virtual machine with MetalLB loadBalance service.

description: This post illustrates setting up a virtual machine with MetalLB LoadBalancer service.

phoracek · 2022-05-11T08:29:51Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+## MetalLB
+
+[MetalLB](https://metallb.universe.tf/) allows you to create Kubernetes services of type `LoadBalancer`, and provides network load-balancer implementation in on-prem clusters that don’t run on a cloud provider.
+MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 modes: Layer 2 and BGP:


Suggested change

MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 modes: Layer 2 and BGP:

MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 modes, Layer 2 and BGP:

phoracek · 2022-05-11T08:30:04Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+- Layer 2 mode (ARP/NDP):
+
+  This mode - which actually does not implement real Load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as the service owner. This configuration mode makes the IPs reachable from the local network.  


Suggested change

This mode - which actually does not implement real Load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as the service owner. This configuration mode makes the IPs reachable from the local network.

This mode - which actually does not implement real load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as the service owner. This configuration mode makes the IPs reachable from the local network.

phoracek · 2022-05-11T08:33:59Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+#### Installing MetalLB on the cluster
+
+There are [many ways](https://metallb.universe.tf/installation/) to install MetalLB. For the sake of this example, we will install MetalLB via Manifests. To do this, follow this [guide](https://metallb.universe.tf/installation/#installation-by-manifest). 


Suggested change

There are [many ways](https://metallb.universe.tf/installation/) to install MetalLB. For the sake of this example, we will install MetalLB via Manifests. To do this, follow this [guide](https://metallb.universe.tf/installation/#installation-by-manifest).

There are [many ways](https://metallb.universe.tf/installation/) to install MetalLB. For the sake of this example, we will install MetalLB via manifests. To do this, follow this [guide](https://metallb.universe.tf/installation/#installation-by-manifest).

phoracek · 2022-05-11T08:38:37Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+
+### Expose the virtual machine with a typed `LoaBalancer` service 


Suggested change

### Expose the virtual machine with a typed `LoaBalancer` service

### Expose the virtual machine with a typed `LoadBalancer` service

Good catch! Done

phoracek · 2022-05-11T08:41:10Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+
+### Spin up a Virtual Machine running Nginx
+
+Now it's time to start-up a virtual machine running nginx using this yaml:


Suggested change

Now it's time to start-up a virtual machine running nginx using this yaml:

Now it's time to start-up a virtual machine running nginx using the following yaml.

phoracek · 2022-05-11T08:45:25Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+### Expose the virtual machine with a typed `LoaBalancer` service 
+
+When creating the `LoadBalancer` typed service, we need to remember annotating the address-pool we want to use 
+`addresspool-sample1` and also add the selector `metallb-service: nginx` 


Suggested change

`addresspool-sample1` and also add the selector `metallb-service: nginx`

`addresspool-sample1` and also add the selector `metallb-service: nginx`:

phoracek · 2022-05-11T08:46:50Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+
+Notice that the service got assigned with an external IP from the range assigned by the PoolAddress: 


Suggested change

Notice that the service got assigned with an external IP from the range assigned by the PoolAddress:

Notice that the service got assigned with an external IP from the range assigned by the address pool:

?

Right, now that we are not using the MetalLB operator, PoolAddress has no meaning.
Done

phoracek · 2022-05-11T08:46:59Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+kubectl get service -n default metallb-nginx-svc
+```
+
+Example output


Suggested change

Example output

Example output:

phoracek · 2022-05-11T08:47:18Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+curl -s -o /dev/null 172.18.1.1:5678 && echo "URL exists"
+```
+
+Example output


Suggested change

Example output

Example output:

RamLavi · 2022-05-11T12:19:17Z

@phoracek ready for your re-review

phoracek · 2022-05-12T11:45:07Z

/lgtm

@maiqueb could you please take a final look?

maiqueb

/lgtm
/hold

Holding in case you want to re-work the sections. You can remove the hold if you prefer @RamLavi .

maiqueb · 2022-05-23T08:56:26Z

_posts/2022-04-03-Virtual-Machines-with-MetalLB.markdown

+EOF
+```
+
+#### Installing Kubevirt on the cluster


I think I do:

Cluster deployment: this one is creating the kind cluster

Components deployment: this one is installing metalLB and KubeVirt in the cluster

Network resources configuration: this one is configuring the metalLB pool

Network utilization: this one is creating the VM and service

Does it make sense ? I share the concern @phoracek 's mentioned in https://github.com/kubevirt/kubevirt.github.io/pull/837/files#r864557203 ... That is part of the cluster configuration, and probably deserves to live in a separate section.

Signed-off-by: Ram Lavi <[email protected]>

maiqueb · 2022-05-23T10:03:41Z

/lgtm
/hold cancel

Thanks.

phoracek

/approve

kubevirt-bot · 2022-05-23T10:51:09Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: phoracek

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [phoracek]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

kubevirt-bot added do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. dco-signoff: yes Indicates the PR's author has DCO signed all their commits. size/L labels Apr 3, 2022

kubevirt-bot requested review from jobbler and mazzystr April 3, 2022 14:24

kubevirt-bot added the kind/blog Label for blog entries label Apr 3, 2022

RamLavi force-pushed the vm_with_metal_lb branch from 6a3eb5c to 1ba52f1 Compare April 4, 2022 09:59

RamLavi marked this pull request as ready for review April 4, 2022 09:59

kubevirt-bot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Apr 4, 2022

RamLavi changed the title ~~[DNM] Load-balancer for VirtualMachines on bare metal Kubernetes clusters~~ Load-balancer for VirtualMachines on bare metal Kubernetes clusters Apr 4, 2022

kubevirt-bot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Apr 4, 2022

phoracek suggested changes Apr 4, 2022

View reviewed changes

kubevirt-bot assigned phoracek Apr 4, 2022

RamLavi force-pushed the vm_with_metal_lb branch 2 times, most recently from de22f5a to 48b93b9 Compare April 5, 2022 11:20

RamLavi changed the title ~~Load-balancer for VirtualMachines on bare metal Kubernetes clusters~~ Load-balancer for virtual machines on bare metal Kubernetes clusters Apr 5, 2022

RamLavi force-pushed the vm_with_metal_lb branch from 48b93b9 to f2495f5 Compare April 5, 2022 11:53

kubevirt-bot removed the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Apr 5, 2022

RamLavi force-pushed the vm_with_metal_lb branch from f2495f5 to dea5202 Compare April 5, 2022 12:27

maiqueb suggested changes Apr 5, 2022

View reviewed changes

kubevirt-bot assigned maiqueb Apr 5, 2022

RamLavi force-pushed the vm_with_metal_lb branch 2 times, most recently from 08510b7 to 9a4ede2 Compare April 6, 2022 10:40

maiqueb suggested changes Apr 6, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch from 9a4ede2 to f66c3e9 Compare April 7, 2022 08:19

maiqueb approved these changes Apr 7, 2022

View reviewed changes

kubevirt-bot added the lgtm Indicates that a PR is ready to be merged. label Apr 7, 2022

RamLavi force-pushed the vm_with_metal_lb branch from f66c3e9 to 19ff2b6 Compare April 7, 2022 09:22

kubevirt-bot removed the lgtm Indicates that a PR is ready to be merged. label Apr 7, 2022

phoracek suggested changes Apr 11, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch from 19ff2b6 to e1a8855 Compare May 1, 2022 11:44

phoracek suggested changes May 2, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch 3 times, most recently from 2ada556 to 4ce0e1f Compare May 3, 2022 08:45

phoracek reviewed May 3, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch from 4ce0e1f to afeb7bb Compare May 4, 2022 04:47

phoracek reviewed May 4, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch 2 times, most recently from 50df1f2 to 762f416 Compare May 4, 2022 08:46

phoracek reviewed May 11, 2022

View reviewed changes

RamLavi force-pushed the vm_with_metal_lb branch from 762f416 to 9c84efd Compare May 11, 2022 12:17

kubevirt-bot added the lgtm Indicates that a PR is ready to be merged. label May 12, 2022

maiqueb reviewed May 23, 2022

View reviewed changes

kubevirt-bot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label May 23, 2022

blog: Load-balancer for vms on bare-metal K8s clusters

040c794

Signed-off-by: Ram Lavi <[email protected]>

RamLavi force-pushed the vm_with_metal_lb branch from 9c84efd to 040c794 Compare May 23, 2022 09:51

kubevirt-bot removed the lgtm Indicates that a PR is ready to be merged. label May 23, 2022

kubevirt-bot added lgtm Indicates that a PR is ready to be merged. and removed do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. labels May 23, 2022

phoracek reviewed May 23, 2022

View reviewed changes

kubevirt-bot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label May 23, 2022

kubevirt-bot merged commit d995305 into kubevirt:main May 23, 2022


		## Introduction

		Over the last year, kubevirt has integrated with MetalLB in order to support fault-tolerant access to an application

	the local network. This method announces the IPs in ARP (for IPV4) and NDP (for IPV6) protocols over the network. From a
	the local network. This method announces the IPs in ARP (for IPv4) and NDP (for IPv6) protocols over the network. From a

	To start it up follow this [guide](https://kind.sigs.k8s.io/docs/user/quick-start)
	To start it up follow this [guide](https://kind.sigs.k8s.io/docs/user/quick-start).


		### Demo environment setup

		We are going to use [kind](https://kind.sigs.k8s.io) provider as an ephemeral Kubernetes cluster.


		### Spin up a Virtual Machine running Nginx

		Now it's time to start-up a virtualMachine running nginx, using this yaml:

	### Expose the virtualMachine with a typed LoaBalancer service
	### Expose the VirtualMachine with a typed LoaBalancer service

	MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 working modes: Layer 2 and BGP modes:
	MetalLB is responsible for assigning/unassigning an external IP Address to your service, using IPs from pre-configured pools. In order for the external IPs to be announced externally, MetalLB works in 2 modes: Layer 2 and BGP:

	This mode provides a real load-balancing behavior, By establishing BGP peering sessions with the network routers. They in turn advertise the external IPs of the `LoadBalancer` service, distributing the load over the nodes.
	This mode provides real load-balancing behavior, by establishing BGP peering sessions with the network routers - which advertise the external IPs of the `LoadBalancer` service, distributing the load over the nodes.

-This mode does not implement Load-balancing behavior provides a failover mechanism where one node owns the `LoadBalancer` service, until the node fails and another node is chosen. In a strict sense, this is not a real load-balancing behavior, but instead it makes the IPs reachable to the local network. This method announces the IPs in ARP (for Ipv4) and NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.
+This mode - which does not implement Load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as service owner. This configuration mode makes the IPs reachable from the local network.
+This method announces the IPs using ARP (for Ipv4) or NDP (for Ipv6) protocols over the network. From a network perspective, the node appears to have multiple IP addresses assigned to a network interface on the chosen node.

	As a Cluster administrator using an on-prem cluster without a network load balancer, now it's possible to add MetalLB operator and gain a load-balancer capabilities (with Services of type LoadBalancer) to virtual machines.
	As a Cluster administrator using an on-prem cluster without a network load balancer, now it's possible to use MetalLB operator to provide load-balancer capabilities (with Services of type LoadBalancer) to virtual machines.

		This demo runs Docker on Linux, which allows sending traffic directly to the load-balancer's external IP if the IP space is within the docker IP space.
		On macOS and Windows however, docker does not expose the docker network to the host, rendering the external IP unreachable from other kind nodes. In order to workaround this, one could expose pods and services using extra port mappings as shown in the extra port mappings section of kind's [Configuration Guide](https://kind.sigs.k8s.io/docs/user/configuration#extra-port-mappings).


		#### Setting Address Pool to be used by the LoadBalancer

		In order to complete the Layer 2 mode configuration, we need to set a range of IP Addresses for the LoadBalancer to use. We want this range to be on the docker kind network, so by using this command:

	You should get the subclass you can set the ip range from. The output should contain a cidr such as 172.18.1.0/16
	You should get the subclass you can set the IP range from. The output should contain a cidr such as 172.18.1.0/16


		#### Installing MetalLB on the cluster

		There are many ways to install MetalLB. For the sake of this example, we will install MetalLB operator on the cluster. To do this, please follow this [link](https://operatorhub.io/operator/metallb-operator).


		### Access the virtual machine from outside the cluster

		Finally, we can check that the nginx server is accessible from outside the cluster:

	description: This post illustrates setting up a virtual machine with MetalLb loadBalance service.
	description: This post illustrates setting up a virtual machine with MetalLB loadBalance service.


		- Layer 2 mode (ARP/NDP):

		This mode - which actually does not implement real Load-balancing behavior - provides a failover mechanism where a single node owns the `LoadBalancer` service, until it fails, triggering another node to be chosen as the service owner. This configuration mode makes the IPs reachable from the local network.


		#### Installing MetalLB on the cluster

		There are [many ways](https://metallb.universe.tf/installation/) to install MetalLB. For the sake of this example, we will install MetalLB via Manifests. To do this, follow this [guide](https://metallb.universe.tf/installation/#installation-by-manifest).

	### Expose the virtual machine with a typed `LoaBalancer` service
	### Expose the virtual machine with a typed `LoadBalancer` service


		### Spin up a Virtual Machine running Nginx

		Now it's time to start-up a virtual machine running nginx using this yaml:

Load-balancer for virtual machines on bare metal Kubernetes clusters #837

Load-balancer for virtual machines on bare metal Kubernetes clusters #837

Conversation

RamLavi commented Apr 3, 2022

kubevirt-bot commented Apr 3, 2022

RamLavi commented Apr 4, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi Apr 5, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi commented Apr 5, 2022

maiqueb left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi Apr 6, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi Apr 6, 2022 • edited Loading

Choose a reason for hiding this comment

RamLavi Apr 6, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

maiqueb left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi May 1, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi May 1, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

RamLavi Apr 5, 2022 •

edited

Loading

RamLavi Apr 6, 2022 •

edited

Loading

RamLavi Apr 6, 2022 •

edited

Loading

RamLavi Apr 6, 2022 •

edited

Loading

RamLavi May 1, 2022 •

edited

Loading

RamLavi May 1, 2022 •

edited

Loading

	`addresspool-sample1` and also add the selector `metallb-service: nginx`
	`addresspool-sample1` and also add the selector `metallb-service: nginx`:

	Notice that the service got assigned with an external IP from the range assigned by the PoolAddress:
	Notice that the service got assigned with an external IP from the range assigned by the address pool: