guestfs: fix destroy task to handle missing storage volumes gracefully

The destroy task was failing with "Storage volume not found" errors when
trying to undefine VMs. This happened because the 'force: true' flag
attempted to delete storage volumes that may not exist.

Fixed by:
1. Replacing 'force: true' with specific flags (nvram, snapshots-metadata,
   checkpoints-metadata) that don't try to delete storage volumes
2. Adding 'ignore_errors: true' to handle any remaining failures gracefully
3. Adding a separate cleanup task that explicitly attempts to delete
   storage volumes with proper error handling using shell commands

This ensures the destroy process completes successfully even when storage
volumes are missing or already deleted.

Generated-by: Claude AI
Signed-off-by: Luis Chamberlain <[email protected]>

Author: mcgrof

playbooks/roles/guestfs/tasks/destroy.yml

@@ -27,12 +27,23 @@
 - name: Undefine each stopped target node
   community.libvirt.virt:
     command: "undefine"
-    force: true
+    flags:
+      - "nvram"
+      - "snapshots-metadata"
+      - "checkpoints-metadata"
     name: "{{ inventory_hostname }}"
     uri: "{{ libvirt_uri }}"
   changed_when: true
   when:
     - inventory_hostname in shutdown_vms.list_vms
+  ignore_errors: true
+
+- name: Clean up storage volumes for target nodes
+  ansible.builtin.shell: |
+    virsh -c {{ libvirt_uri }} vol-delete --pool {{ kdevops_storage_pool }} {{ inventory_hostname }}/root.raw 2>/dev/null || true
+    virsh -c {{ libvirt_uri }} vol-delete --pool {{ kdevops_storage_pool }} {{ inventory_hostname }}.raw 2>/dev/null || true
+  changed_when: false
+  ignore_errors: true
 
 - name: Remove per-node configuration files
   ansible.builtin.file: