vm ping vpc gw and ping each other (vms inside a vpc subnet) is failed after that kube-ovn-controller set keep-vm-ip as true

[bobz965]

Expected Behavior

vm still works well when kube-ovn-controller set keep-vm-ip = true, and reboot all vm

Actual Behavior

Steps to Reproduce the Problem

use kube-ovn 1.10.0

1. set kube-ovn-controller boot args keep-vm-ip = true
2. wait all kube-ovn-controller update completely
3. reboot all vm to test vm keep ip feature
4. vm keep-vm-ip feature is ok, and vm network (inside vm ) is all set correctly, but vm ping gw failed and ping each other failed
5. kube-ovn-controller updated and update nat-gw-pod （deployment）， so all nat-gw-pod delete and created （should not）
6. gw nat pod ping vpc subnet gw is ok, and ping vms (inside the same vpc subnet) failed, and ping -I net1 external gw failed
7. add some pod, still some pod in the same subnet ping failed
8. after hard reboot vm， vm ping gw ok and ping each other ok

Additional Info

• Kubernetes version:

Client Version: version.Info{Major:"1", Minor:"23", GitVersion:"v1.23.7", GitCommit:"42c05a547468804b2053ecf60a3bd15560362fc2", GitTreeState:"clean", BuildDate:"2022-05-24T12:30:55Z", GoVersion:"go1.17.10", Compiler:"gc", Platform:"linux/amd64"} Server Version: version.Info{Major:"1", Minor:"23", GitVersion:"v1.23.7", GitCommit:"42c05a547468804b2053ecf60a3bd15560362fc2", GitTreeState:"clean", BuildDate:"2022-05-24T12:24:41Z", GoVersion:"go1.17.10", Compiler:"gc", Platform:"linux/amd64"}

• kube-ovn version:

  kubeovn/kube-ovn:v1.10.0
  

CentOS Stream 8 5.4.205-1.el8.elrepo.x86_64

# 

[bobz965]

这边测试逐个硬重启下vm 是可以恢复的，可能是并发重建的VM较多，有49台

[bobz965]

kube-ovn-controller 重启，导致nat gw pod 全部重启，是因为我们手动改了deployment 的taints，但是vpc nat gw 的负载尚未支持taints

[bobz965]

fixed，nat gw pod support tolerations now