You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
dhclient.conf entries supersede domain-name and supersede domain-search aren't enclosed in quotes. This results in resolvconf enter hooks failing which breaks networking.service on debian. With networking.service broken the dhcp leases cannot be renewed and the machines loses network access after they expire.
What did you expect to happen?
dhclient to have correct format like this:
# Ansible entries BEGIN
supersede domain-name-servers 169.254.25.10,1.1.1.1,8.8.8.8;
supersede domain-name "cluster.local";
supersede domain-search "default.svc.cluster.local,svc.cluster.local";
# Ansible entries END
How can we reproduce it (as minimally and precisely as possible)?
Install kubespray on debian that uses resolvconf. Might be reproducible on other distributions as well, but when resolvconf is not present the issue doesn't happen
What happened?
dhclient.conf entries
supersede domain-name
andsupersede domain-search
aren't enclosed in quotes. This results in resolvconf enter hooks failing which breaks networking.service on debian. With networking.service broken the dhcp leases cannot be renewed and the machines loses network access after they expire.What did you expect to happen?
dhclient to have correct format like this:
How can we reproduce it (as minimally and precisely as possible)?
Install kubespray on debian that uses resolvconf. Might be reproducible on other distributions as well, but when resolvconf is not present the issue doesn't happen
OS
Version of Ansible
Version of Python
Python 3.13.1
Version of Kubespray (commit)
7d14c42
Network plugin used
calico
Full inventory with variables
variables are pretty much default just with cloud_provider: external and kube_oidc_auth enabled
Command used to invoke ansible
ansible-playbook -i inventory.ini -b -u root upgrade-cluster.yaml -e upgrade_node_confirm=true -e system_upgrade=true
Output of ansible run
Ansible runs successfully but nodes start having network issues when the dhcp lease expires
Anything else we need to know
I can submit a PR for this. I have it already fixed locally.
The text was updated successfully, but these errors were encountered: