Load balancer is being provisioned with MetalLB - kubernetes

Question

I have k8s cluster v1.24.4+rke2r1, created by Rancher. I have already installed MetalLB, but when I try create pod with nginx and assign LoadBalancer I still have:
Service is ready. Load balancer is being provisioned
This is my pod and service config
apiVersion: v1
kind: Service
metadata:
name: nginx-service
annotations:
metallb.universe.tf/address-pool: public-ips
spec:
selector:
app: nginx
ports:
- port: 80
targetPort: 80
type: LoadBalancer
Pod
apiVersion: v1
kind: Pod
metadata:
name: nginx
labels:
name: nginx
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.23.2-alpine
resources:
limits:
memory: "128Mi"
cpu: "500m"
ports:
- containerPort: 80
My MetalLB online config
configInline:
address-pools:
- addresses:
- 192.168.1.100-192.168.1.200
autoAssign: true
name: public-ips
protocol: layer2
When I describe my nginx-service I got
Name: nginx-service
Namespace: metal-test
Labels: <none>
Annotations: metallb.universe.tf/address-pool: public-ips
Selector: app=nginx
Type: LoadBalancer
IP Family Policy: SingleStack
IP Families: IPv4
IP: 10.43.127.38
IPs: 10.43.127.38
Port: <unset> 80/TCP
TargetPort: 80/TCP
NodePort: <unset> 30010/TCP
Endpoints: 10.42.212.79:80
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
My service list
NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
calico-system calico-kube-controllers-metrics ClusterIP 10.43.208.243 <none> 9094/TCP 21h
calico-system calico-typha ClusterIP 10.43.230.52 <none> 5473/TCP 21h
cattle-system cattle-cluster-agent ClusterIP 10.43.198.73 <none> 80/TCP,443/TCP 21h
default kubernetes ClusterIP 10.43.0.1 <none> 443/TCP 21h
default nginx-service LoadBalancer 10.43.245.80 <pending> 8080:32146/TCP 66m
kube-system rke2-coredns-rke2-coredns ClusterIP 10.43.0.10 <none> 53/UDP,53/TCP 21h
kube-system rke2-metrics-server ClusterIP 10.43.84.128 <none> 443/TCP 21h
longhorn-system csi-attacher ClusterIP 10.43.49.124 <none> 12345/TCP 3h29m
longhorn-system csi-provisioner ClusterIP 10.43.236.132 <none> 12345/TCP 3h29m
longhorn-system csi-resizer ClusterIP 10.43.153.211 <none> 12345/TCP 3h29m
longhorn-system csi-snapshotter ClusterIP 10.43.182.109 <none> 12345/TCP 3h29m
longhorn-system longhorn-admission-webhook ClusterIP 10.43.49.242 <none> 9443/TCP 3h29m
longhorn-system longhorn-backend ClusterIP 10.43.71.124 <none> 9500/TCP 3h29m
longhorn-system longhorn-conversion-webhook ClusterIP 10.43.180.185 <none> 9443/TCP 3h29m
longhorn-system longhorn-engine-manager ClusterIP None <none> <none> 3h29m
longhorn-system longhorn-frontend ClusterIP 10.43.95.1 <none> 80/TCP 3h29m
longhorn-system longhorn-replica-manager ClusterIP None <none> <none> 3h29m
metallb metallb-webhook-service ClusterIP 10.43.211.242 <none> 443/TCP 178m
my metallb pod
NAME READY STATUS RESTARTS AGE
metallb-controller-6776dbc97d-kmkf9 1/1 Running 1 (177m ago) 177m
metallb-speaker-jrnmf 1/1 Running 0
I used this tutorial to install MetalLB http://xybernetics.com/techtalk/how-to-install-metallb-on-rancher-kubernetes-cluster/
I don't have any active firewall and I don't have nginx ingress on my cluster. Any idea what I do wrong? I do this on my local network.