Upgrade Version via Helm where install.sh Script Use in Previous Installation

[ikandars]

Feature request

I use install.sh script to install KubeOVN in version 1.11.0. I want to upgrade my cluster into version 1.11.1. It will fail if I use this helm command:

helm install --debug kubeovn ./kubeovn-helm --set MASTER_NODES=${Node0},

Is there any simple step to upgrade from install.sh script to Helm?

Use case

My quick heck to do that is by using this approach:

First, convert the same version into Halm. In my case, I have to convert from v1.11.0 install.sh script base into helm v1.11.0 base. Before I run helm install command, I need to add labels and annotations to existing KubeOVN objects. I use this script to do that:

#!/bin/bash

declare -a arr=("ServiceAccount ovn" 
"CustomResourceDefinition vpc-dnses.kubeovn.io" 
"CustomResourceDefinition switch-lb-rules.kubeovn.io"
"CustomResourceDefinition vpc-nat-gateways.kubeovn.io"
"CustomResourceDefinition iptables-eips.kubeovn.io"
"CustomResourceDefinition iptables-fip-rules.kubeovn.io"
"CustomResourceDefinition iptables-dnat-rules.kubeovn.io"
"CustomResourceDefinition iptables-snat-rules.kubeovn.io"
"CustomResourceDefinition ovn-eips.kubeovn.io"
"CustomResourceDefinition ovn-fips.kubeovn.io"
"CustomResourceDefinition ovn-snat-rules.kubeovn.io"
"CustomResourceDefinition vpcs.kubeovn.io"
"CustomResourceDefinition ips.kubeovn.io"
"CustomResourceDefinition vips.kubeovn.io"
"CustomResourceDefinition subnets.kubeovn.io"
"CustomResourceDefinition vlans.kubeovn.io"
"CustomResourceDefinition provider-networks.kubeovn.io"
"CustomResourceDefinition security-groups.kubeovn.io"
"CustomResourceDefinition htbqoses.kubeovn.io"
"ClusterRole system:ovn"
"ClusterRoleBinding ovn"
"Service kube-ovn-controller"
"Service kube-ovn-monitor"
"Service ovn-nb"
"Service ovn-northd"
"Service kube-ovn-cni"
"Service kube-ovn-pinger"
"Service ovn-sb"
"DaemonSet kube-ovn-cni"
"DaemonSet ovs-ovn"
"DaemonSet kube-ovn-pinger"
"Deployment ovn-central"
"Deployment kube-ovn-controller"
"Deployment kube-ovn-monitor")


for i in "${arr[@]}"
do
   kubectl -n kube-system patch $i --type merge -p '{
    "metadata":{
        "annotations":{
            "meta.helm.sh/release-name": "kubeovn",
            "meta.helm.sh/release-namespace": "default"
        },
        "labels":{
            "app.kubernetes.io/managed-by": "Helm"
        }
    }
}'
done

To get Helm charts, so we need clone the project first.

git clone https://github.com/kubeovn/kube-ovn.git
cd kube-ovn
git checkout v1.11.0

Once it done, then I create a custom helm values.yaml file where the value populated from existing deployment.

replicaCount: 3
MASTER_NODES: "10.255.224.249,10.255.224.246,10.255.224.156"

networking:
  IFACE: "enp1s0"

ipv4:
  POD_CIDR: "172.16.0.0/20"
  POD_GATEWAY: "172.16.0.1"
  SVC_CIDR: "172.16.16.0/20"
  JOIN_CIDR: "100.64.0.0/16"

You can get all those values from kubectl -n kube-system get deploy/kube-ovn-controller -o yaml

kubectl label no -lbeta.kubernetes.io/os=linux kubernetes.io/os=linux --overwrite
kubectl label no -lnode-role.kubernetes.io/control-plane  kube-ovn/role=master --overwrite
kubectl label no -lovn.kubernetes.io/ovs_dp_type!=userspace ovn.kubernetes.io/ovs_dp_type=kernel  --overwrite

helm install --debug kubeovn .kubeovn-helm -f kubeovn-custom-values.yaml

Then check all KubeOVN related pods in running state.

kubectl -n kube-system get pod | grep -i ovn

Now, I run the following steps to do the upgrade via Helm:

git checkout v1.11.1
helm upgrade --debug kubeovn ./kubeovn-helm -f kubeovn-custom-values.yaml

Last, check the KubeOVN related pods status:

kubectl -n kube-system get pod | grep -i ovn

I found that version v1.11.2-v1.11.3 still use container image in version v1.11.1. I put set param in helm upgrade command to correct image tag:

helm upgrade --debug kubeovn ./kubeovn-helm -f kubeovn-custom-values.yaml --set global.images.kubeovn.tag=v1.11.2

[github-actions]

Issues go stale after 60d of inactivity. Please comment or re-open the issue if you are still interested in getting this issue fixed.

[aminmr]

Hello @ikandars
I think if your procedure is working fine and has no issues, we can add it to the kube-ovn documentation. We only need the approval of the main contributors @oilbeater