To avoid potential bugs on the DU and for more control for companies and others who would prefer more control over the upgrade process of nodes.
With Kubeadm one would just type one command "kubeadm upgrade node" on each individual node to start the upgrade process for that 1 node. Something like this would be way more efficient method for those with small clusters or for those who would like another option that is more simplified and controlled especially for Worker Nodes.
Is there no way to simplify that further? Where with Kubeadm you go to each Node one by one and run one command to trigger upgrade for that node?
Because if it can simplified one could always use Ansible Playbooks to perform batch upgrades
https://platform9.com/docs/qbert/ref
batchUpgradeNodes
array[string]
List of worker nodes uuids to upgrade in parallel at once during an upgrade
curl --request POST \
--url https://platform9.io/qbert/v4/{project_uuid}/clusters/{uuid}/upgrade \
--header "X-Auth-Token: {X-Auth-Token}" \
--data '{
"batchUpgradePercent": "{number}",
"batchUpgradeNodes": [
"{array[string]...}"
],
"containerRuntime": "{string}"
}'
There is a Node-by-node batch upgrade avaialbe in 5.4 onwards, this needs to be used via the Qbert API