Problem
When proxmox_vm.list.to.jsons.sh returns no data — whether because no VMs are deployed, the Proxmox API is unreachable, or the devkit bug (range42-ansible_roles-debug-devkit#111) is triggered — delete-everything aborts immediately with a fatal error.
This means firewall teardown and known_hosts cleanup are skipped entirely, even though neither depends on VM state. The deployer is left in a dirty state that requires manual intervention.
Fix
Demote "no VM data" from a fatal abort to a warning: VM stop/delete is skipped, but the rest of the nuke sequence (firewall teardown, known_hosts cleanup) always runs to completion.
This also improves resilience when debug-devkit#111 is not yet fixed on the deployer — delete-everything degrades gracefully instead of aborting.
Tracked in PR #208.
Problem
When
proxmox_vm.list.to.jsons.shreturns no data — whether because no VMs are deployed, the Proxmox API is unreachable, or the devkit bug (range42-ansible_roles-debug-devkit#111) is triggered —delete-everythingaborts immediately with a fatal error.This means firewall teardown and
known_hostscleanup are skipped entirely, even though neither depends on VM state. The deployer is left in a dirty state that requires manual intervention.Fix
Demote "no VM data" from a fatal abort to a warning: VM stop/delete is skipped, but the rest of the nuke sequence (firewall teardown,
known_hostscleanup) always runs to completion.This also improves resilience when debug-devkit#111 is not yet fixed on the deployer —
delete-everythingdegrades gracefully instead of aborting.Tracked in PR #208.