- Upgrade to latest version of
epfl_si.actions
, thereby gainingdelegate_to:
support in all tasks - Rewrite documentation using the
ansible-doc
standard - Drop
epfl_si.rancher.rke2_access
role, as it was unused and at odds with 0.6.0's strategy of not managing kubeconfig files anymore epfl_si.rancher.rke2-node
roleepfl_si.rancher.rancher_project
lookup pluginepfl_si.rancher.rancher_helm_chart
action plugin
- Remove responsibility for creating / maintaining kubeconfig files.
The
epfl_si.rancher.rancher_login
action now provides a.kubeconfig
field in itsregister
ed structure; it's up to you to write that into a file (should you choose to) - Fix: do not insist on
import yaml
working on the remote side. - Document
epfl_si.rancher.rancher_login
in full i.e. that action plugin is ready for prime-time.
epfl_si.rancher.rke2
connection plugin is gone- Fix bugs and get rid of dead code
rancher_login
can now log in to multiple clusters.- Fix a Python interpreter mis-discovery bug (whence the interpreter discovered over ssh would “leak” over to normal connections)
- Turn off Ansible's Python interpreter discovery when running tasks on the Rancher back-end machine.
rancher_registration
now works before initializing the cluster (DUH!). As a consequence, it now requiresrancher_manager_url
andcluster_name
parameters.
💡 Upload of version 0.2.0 to Galaxy was botched for some reason.
- New
rancher_registration
task