Max Gautier
Max Gautier
I haven't reviewed this super thoroughly, but given this is contrib, I think we can go ahead. /lgtm /approve
Is that reproducible with a setup like this: ```ini [kube_control_plane] node-1 [etcd] node-1 node-2 node-3 [kube_node] node-1 node-2 node-3 node-4 ``` ? (This is the `node-etcd-client` setup which is tested...
Hum it looks like the conditions are: - Separate etcd / master - nodes are etcd clients (eg, calico using etcd store) - maybe node != control plane ? Not...
I was more thinking something like that ```ini [kube_control_plane] host1 [etcd] host2 [kube_node] host3 [all:vars] network_plugin=calico ``` (If HA is not required to trigger the bug, this makes the test...
Great, thanks for testing that ! We'll need to add that to the CI in the PR to fix this so it does not regress again.
I had the ansible-lint error for a while locally, it seems to have weird trigger condition: ansible/ansible-lint#4579 might be the issue, not sure.
Hum, this is a bit weird, I apparently can't reproduce this on master (and I don't see what could have fixed this :thinking: ) https://gitlab.com/kargo-ci/kubernetes-sigs-kubespray/-/jobs/8622090173
Do you still have the issue if you use that inventory with latest master ?
I don't think this will pass CI though, the ansible-lint error seems sensitive to the order of the hooks. The part where is occurs should be cleanup though (it's to...
(Or latest release-2.26 for that matter I can't reproduce it either on the top of the branch :disappointed: