when multus is used cluster playbook run fails on v2.26

[bvujnovac]

What happened?

When multus (kube_network_plugin_multus: true) is used cluster playbook run fails with msg:

TASK [kubernetes-apps/network_plugin/multus : Multus | Start resources] ******** fatal: [ipv46-deploy1]: FAILED! => {"msg": "'ansible.vars.hostvars.HostVarsVars object' has no attribute 'multus_manifest_2.results'. 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'multus_manifest_2.results'"}

The issue seem to be related to change in #10934

What did you expect to happen?

The expectation was for the task to succeed.

How can we reproduce it (as minimally and precisely as possible)?

copy sample inventory
cp -rfp inventory/sample inventory/mycluster

add hosts_ipv46-deploy.yaml to inventory/mycluster

all:
  vars:
    ansible_connection: "ssh"
    ansible_user: "so"
    ansible_ssh_pass: "changeme"
  hosts:
    ipv46-deploy1:
      ansible_host: 172.17.180.50
      ip: 172.17.180.50
      ip6: 2001:2001:362::50
      access_ip: 172.17.180.50
    ipv46-deploy2:
      ansible_host: 172.17.180.51
      ip: 172.17.180.51
      ip6: 2001:2001:362::51
      access_ip: 172.17.180.51
    ipv46-deploy3:
      ansible_host: 172.17.180.52
      ip: 172.17.180.52
      ip6: 2001:2001:362::52
      access_ip: 172.17.180.52
  children:
    kube_control_plane:
      hosts:
        ipv46-deploy1:
    kube_node:
      hosts:
        ipv46-deploy2:
        ipv46-deploy3:
    etcd:
      hosts:
        ipv46-deploy1:
    calico_rr:
      hosts: {}
    k8s_cluster:
      children:
        kube_control_plane:
        kube_node:
        calico_rr:

add cluster variables cluster-variables_dual_stack.yaml to inventory/mycluster

kube_version: v1.30.6
kube_network_plugin: calico
kube_proxy_mode: ipvs
kube_network_plugin_multus: true
macvlan_interface: "net1"
multus_cni_version: "0.3.1"
enable_dual_stack_networks: true

container_manager: crio
etcd_deployment_type: host
download_container: false
skip_downloads: false
crio_insecure_registries:
  - registry-int.com:5000
  - registry-int.com:443

metrics_server_enabled: true
metrics_server_kubelet_insecure_tls: true
metrics_server_metric_resolution: 15s
metrics_server_replicas: 2

enable_nodelocaldns: false
auto_renew_certificates: true

helm_enabled: true

ntp_enabled: true
ntp_manage_config: true

kube_service_addresses: 10.96.0.0/16
kube_pods_subnet: 10.240.0.0/16
kube_service_addresses_ipv6: 2001:db8:42:1::/112
kube_pods_subnet_ipv6: 2001:db8:42:0::/56
kube_network_node_prefix_ipv6: 64

deploy_netchecker: false

calico_vxlan_mode: Always
calico_vxlan_mode_ipv6: Always

nat_outgoing: true
nat_outgoing_ipv6: true

kubeconfig_localhost: false

upstream_dns_servers:
  - 10.14.252.251
  - 10.14.252.248

run playbook
ansible-playbook -i inventory/mycluster/hosts_ipv46-deploy.yaml -e @inventory/mycluster/cluster-variables_dual_stack.yaml --become --become-user=root cluster.yml

OS

Linux 5.14.0-503.15.1.el9_5.x86_64 x86_64
NAME="Rocky Linux"
VERSION="9.5 (Blue Onyx)"
ID="rocky"
ID_LIKE="rhel centos fedora"
VERSION_ID="9.5"
PLATFORM_ID="platform:el9"
PRETTY_NAME="Rocky Linux 9.5 (Blue Onyx)"
ANSI_COLOR="0;32"
LOGO="fedora-logo-icon"
CPE_NAME="cpe:/o:rocky:rocky:9::baseos"
HOME_URL="https://rockylinux.org/"
VENDOR_NAME="RESF"
VENDOR_URL="https://resf.org/"
BUG_REPORT_URL="https://bugs.rockylinux.org/"
SUPPORT_END="2032-05-31"
ROCKY_SUPPORT_PRODUCT="Rocky-Linux-9"
ROCKY_SUPPORT_PRODUCT_VERSION="9.5"
REDHAT_SUPPORT_PRODUCT="Rocky Linux"
REDHAT_SUPPORT_PRODUCT_VERSION="9.5"

Version of Ansible

ansible [core 2.16.14]
config file = /root/k8s-1.30_install/kubespray/ansible.cfg
configured module search path = ['/root/k8s-1.30_install/kubespray/library']
ansible python module location = /root/k8s-1.30_install/kubespray-venv/lib64/python3.11/site-packages/ansible
ansible collection location = /root/.ansible/collections:/usr/share/ansible/collections
executable location = /root/k8s-1.30_install/kubespray-venv/bin/ansible
python version = 3.11.9 (main, Sep 11 2024, 00:00:00) [GCC 11.5.0 20240719 (Red Hat 11.5.0-2)] (/root/k8s-1.30_install/kubespray-venv/bin/python3.11)
jinja version = 3.1.4
libyaml = True

Version of Python

Python 3.11.9

Version of Kubespray (commit)

75e12e8

Network plugin used

calico

Full inventory with variables

inventory with variables:
https://gist.github.com/bvujnovac/a17012e5c3c36a9d0d84e89d1206e217#file-gistfile1-txt

Command used to invoke ansible

ansible-playbook -i inventory/mycluster/hosts_ipv46-deploy.yaml -e @inventory/mycluster/cluster-variables_dual_stack.yaml --become --become-user=root cluster.yml

Output of ansible run

playbook run output:

https://gist.github.com/bvujnovac/ec4400289b25b31c02744c032d8552dd#file-gistfile1-txt

ansible -i inventory/mycluster/hosts_ipv46-deploy.yaml all -m debug -a "var=hostvars[inventory_hostname]"

Anything else we need to know

I've tried to change the affecting line as I don't think multus_manifest_2.results can be directly accessed like that.

the following patch seems to work but I have only tested it for my limited use case.

diff --git a/roles/kubernetes-apps/network_plugin/multus/tasks/main.yml b/roles/kubernetes-apps/network_plugin/multus/tasks/main.yml
index 54dd1ed61..82ff57efb 100644
--- a/roles/kubernetes-apps/network_plugin/multus/tasks/main.yml
+++ b/roles/kubernetes-apps/network_plugin/multus/tasks/main.yml
@@ -9,7 +9,7 @@
     state: "latest"
   delegate_to: "{{ groups['kube_control_plane'][0] }}"
   run_once: true
-  with_items: "{{ (multus_manifest_1.results | default([])) + (multus_nodes_list | map('extract', hostvars, 'multus_manifest_2.results') | default([]) | list) }}"
+  with_items: "{{ (multus_manifest_1.results | default([])) + (multus_nodes_list | map('extract', hostvars, 'multus_manifest_2') | map(attribute='results') | default([]) | list) }}"
   loop_control:
     label: "{{ item.item.name if item != None else 'skipped' }}"
   vars: