There is a scenario that I added a node to cluster which is AOS 18.104.22.168.
After the expanding cluster finished. I need to add the node to the virtual switch.
There are 3 vs in the cluster. So every time I added the node to the virtual switch, all the node in the cluster will be rolling update because there are some workload in the cluster and I need to select the standard mode.
When I just select the nodes in the virtual switch edit page. there is a warning block the operaiton.
Warning information: “When Creating/Editing a virtual switch, all hosts must be selected except storage-only node”
So my question is why all nodes need to be rolling update when I add the node to the virtual switch and other configuraiton keep same. and the operation time will be very long if add a node to a big cluster.
Another strange issue is that when the virtual switch is in the updating process, the VMs will live migrated to the new nodes which is not finished uplink update. So some VMs may lost connection for some time. This is very strange and I think it should not happen in this scenario.