Upgrade Harvester with bonded NICs on network

• Related issues: #3047 [BUG] migrate_harv_mgmt_to_mgmt_br.sh should remove ClusterNetwork resource

Category:

• Upgrade Harvester

Environment setup from v1.0.3 upgrade to v1.1.1

1. Clone ipxe-example and switch to v1.0 branch
2. Add three additional Network interface in Vagrantfile

     harvester_node.vm.network 'private_network',
       libvirt__network_name: 'harvester',
       mac: @settings['harvester_network_config']['cluster'][node_number]['mac']
     harvester_node.vm.network 'private_network',
       libvirt__network_name: 'harvester'
     harvester_node.vm.network 'private_network',
       libvirt__network_name: 'harvester'
     harvester_node.vm.network 'private_network',
       libvirt__network_name: 'harvester'
   

3. Edit the config-create.yaml.j2 and config-join.yaml.j2 in /ansible/role/harvester/template/
4. Add the cluster_network and defaultPysicalNIC to harvester-mgmt

   cluster_networks:
     vlan:
       enable: true
       description: "some description about this cluster network"
       config:
         defaultPhysicalNIC: harvester-mgmt
   

5. Bond multiple NICs on harvester-mgmt and harvester-vlan networks

   networks:
       harvester-mgmt:
         interfaces:
         - name: {{ settings['harvester_network_config']['cluster'][0]['mgmt_interface'] }}  # The management interface name
         - name: ens9
         method: dhcp
       bond0:
         interfaces:
         - name: {{ settings['harvester_network_config']['cluster'][0]['vagrant_interface'] }}
         method: dhcp
       harvester-vlan:
         interfaces:
         - name: ens7
         - name: ens8
         method: none
   

Verification Steps

1. Provision previous version of Harvester cluster

2. For VLAN 1 testing, enable network on settings, select harvester-mgmt

3. Create a vlan1 network with id 1 on Networks page

4. For other VLAN id testing, enable network on settings, select correct network interface e.g enps0f1

5. Create a vlan91 network with available id 91 on Networks page (id number depends on your available network setting)

6. Create images with different OS distribution

7. Create several virtual machines, set network to management-network or available vlan

8. Create virtual machine on different target node

9. Setup NFS or S3 backup target in settings

10. Backup each virtual machines

11. Shutdown all virtual machines

12. Offline upgrade to target version, refer to https://docs.harvesterhci.io/v1.1/upgrade/automatic

Expected Results

1. Can completely upgrade Harvester to specific version
2. Cluster Network exists
3. Can create Vlan 1 on mgmt and route correctly
4. Can create other valid VLAN on specific NIC and route correctly
5. Check the network connectivity of vlan and management network
6. All pods are running correctly
7. Check can display Monitoring Chart
   • Prometheus dashboard
   • VM metrics
8. Can access dashboard by VIP
9. Can use original password to login
10. Can start VM in running status
11. Image exists without corrupted
12. Volume exists
13. Virtual network exists
14. Backup exists
15. Setting value exists
16. Can restore VM from backup
17. Can create new VMs (test)