gateway.type is not working in helm upgrade operation

My command is
```
helm upgrade apisix apisix/apisix --set persistence.storageClass=apisix-etcd-nas --set persistence.enabled=true --set gateway.type=LoadBalancer --set ingress-controller.enabled=true --set ingress-controller.config.apisix.serviceNamespace=apisix --namespace apisix
```
The gateway.type is LoadBalancer, But it's not working, still got the following response. It's bacame NODE_PORT.
```
Release "apisix" has been upgraded. Happy Helming!
NAME: apisix
LAST DEPLOYED: Mon Aug 28 14:36:40 2023
NAMESPACE: apisix
STATUS: deployed
REVISION: 2
TEST SUITE: None
NOTES:
1. Get the application URL by running these commands:
export NODE_PORT=$(kubectl get --namespace apisix -o jsonpath="{.spec.ports[0].nodePort}" services apisix-gateway)
export NODE_IP=$(kubectl get nodes --namespace apisix -o jsonpath="{.items[0].status.addresses[0].address}")
echo http://$NODE_IP:$NODE_PORT
```
It caused my loadbalancer IP been changed(assigned by dhcp), how should I do to avoid this problem?

Start by reproducing the shown helm upgrade command and inspect the rendered apisix-gateway Service in Kubernetes, including its service type and assigned address. Compare the chart values with the resulting Service configuration; done means gateway.type=LoadBalancer is applied during upgrade and the load-balancer address is not unexpectedly replaced.

Written by the indexing model from the issue text.