Update linux-common role to support Infiniband netdevs (#230)

* Introduce linux-common role, initial inbox support for Infiniband

* Change cloud-init net renderer\ to support IB netdev conf

* Add play for linux-common role

* Replace references to stackhpc where possible (#232)

* Use workflow actions from azimuth-cloud (#233)

* Allow image builds to run on other clouds (#236)

* Allow image builds to run on other clouds

* Update zenith tag

* Add support for specifying a volume type

* Update ubuntu-jammy base image (20240808) (#229)

Co-authored-by: mkjpryor <[email protected]>

* Update Kubernetes patch versions (v1.30.4, v1.27.16, v1.29.8, v1.28.13) (#227)

Co-authored-by: mkjpryor <[email protected]>

* Update ubuntu-jammy base image (20240821) (#237)

Co-authored-by: mkjpryor <[email protected]>

* Bump community.general to 9.3.0 (#225)

Co-authored-by: mkjpryor <[email protected]>

* Bump prometheus.prometheus to 0.17.1 (#219)

Co-authored-by: mkjpryor <[email protected]>

* Bump containers.podman to 1.15.4 (#224)

Co-authored-by: mkjpryor <[email protected]>

* Bump prometheus to v2.54.0 (#234)

Co-authored-by: mkjpryor <[email protected]>

* Stop building images for Kubernetes 1.27 (#238)

* Bump zenith-client to 0.12.0 (#239)

Co-authored-by: mkjpryor <[email protected]>

* Add new Kubernetes version (v1.31.0) (#235)

Co-authored-by: mkjpryor <[email protected]>

* Use linux-ansible-init from azimuth-cloud image_utils collection (#226)

* Update to use linux-ansible-init role from Galaxy and remove in-repo

* Update requirements.yml

add external ansible-init to requirements

* Update main.yml

update ansible-init dependency for linux-webconsole

* Update main.yml

* Update main.yml

* Update main.yml

* Update requirements.yml

* Use ansible-init from image_utils collection

* bump azimuth-image-utils release

* verbose linux-rstudio.pkr.hcl

* verbose linux-rstudio.pkr.hcl

* image-utils 0.2.0

* Update linux-rstudio.pkr.hcl

---------

Co-authored-by: bertie <[email protected]>

* Move common tasks into a shared linux-common role (#242)

* Move common tasks into a shared linux-common role

* Update Trivy action version

* Remove isolate command to prevent ansible-init from running

* Update ubuntu-jammy base image (20240912) (#245)

Co-authored-by: mkjpryor <[email protected]>

* Update Kubernetes patch versions (v1.31.1, v1.29.9, v1.30.5, v1.28.14) (#244)

Co-authored-by: mkjpryor <[email protected]>

* Pass GitHub token for Trivy DB download (#249)

* Attempt 2 at fixing Trivy DB ratelimit errors

* Use Trivy DB mirror to avoid rate-limiting issues

* Use image-builder tags when updating

* Update ubuntu-jammy base image (20241002) (#253)

Co-authored-by: mkjpryor <[email protected]>

* Exclude low severity results from SARIF

* Bump prometheus.prometheus to 0.19.0 (#246)

Co-authored-by: mkjpryor <[email protected]>

* Bump prometheus to v2.54.1 (#240)

Co-authored-by: mkjpryor <[email protected]>

* Bump community.general to 9.5.0 (#241)

Co-authored-by: mkjpryor <[email protected]>

* Bump ansible.posix to 1.6.0 (#243)

Co-authored-by: mkjpryor <[email protected]>

* Bump containers.podman to 1.16.1 (#248)

Co-authored-by: mkjpryor <[email protected]>

* Bump nginx to 1.27.2 (#252)

Co-authored-by: mkjpryor <[email protected]>

* Install Infiniband inbox driver + load kernel modules via linux-common

* Bump default repo2docker volume size to 10GB

* Change default cloud-init network activator to networkd

* Tidy common role

* Tidy common role

* Update package install name

* Bump Linux desktop size to 15GB

---------

Co-authored-by: Matt Pryor <[email protected]>
Co-authored-by: azimuth-ci-bot[bot] <142236172+azimuth-ci-bot[bot]@users.noreply.github.com>
Co-authored-by: mkjpryor <[email protected]>
Co-authored-by: azimuth-automation-bot[bot] <177105362+azimuth-automation-bot[bot]@users.noreply.github.com>
Co-authored-by: bertiethorpe <[email protected]>
Co-authored-by: bertie <[email protected]>

ansible/linux-common.yml

@@ -0,0 +1,4 @@
+---
+- hosts: all
+  roles:
+    - role: linux-common

ansible/roles/linux-common/defaults/main.yml

@@ -0,0 +1,3 @@
+---
+kernel_modules:
+  - ib_ipoib

ansible/roles/linux-common/tasks/main.yml

@@ -1,28 +1,45 @@
 ---
-
-- name: Update apt cache
+# ACL Required for become to an unprivileged user to work
+# Ensure up-to-date CA certificates
+# Use linux-generic to track linux-modules-extra (IB)
+- name: Install common packages
   apt:
     update_cache: true
-  when: ansible_os_family == "Debian"
-
-- name: Ensure up-to-date CA certificates
-  package:
-    name: ca-certificates
+    pkg:
+    - acl
+    - ca-certificates
+    - linux-generic
     state: latest
 
-# Required for become to an unprivileged user to work
-# Using the apt module seems to work more reliably than package :-/
-- name: Install ACL package
-  apt:
-    update_cache: true
-    name: acl
-    state: present
-  when: ansible_os_family == "Debian"
-
 # Enables the ansible-init system, but does not install any default playbooks
 - include_role:
     name: azimuth_cloud.image_utils.linux_ansible_init
 
 # Enables volumes to be referred to by tag for setting up mounts in cloud-config
 - include_role:
     name: linux-volumes-by-tag
+
+- name: Load relevant kernel modules
+  template:
+    src: modules.conf.j2
+    dest: /etc/modules-load.d/modules.conf
+
+# The cloud-init netplan renderer is default for ubuntu
+# but does not support InfiniBand devices
+- name: Switch cloud-init to alternative network renderer
+  lineinfile:
+    path: /etc/cloud/cloud.cfg
+    regexp: "^    renderers:.*"
+    line: "    renderers: ['networkd', 'network-manager']"
+    state: present
+
+- name: Switch cloud-init to alternative network activator
+  lineinfile:
+    path: /etc/cloud/cloud.cfg
+    regexp: "^    activators:.*"
+    line: "    activators: ['networkd', 'eni', 'network-manager']"
+    state: present
+
+- name: Clean up after ourselves
+  ansible.builtin.apt:
+    clean: yes

ansible/roles/linux-common/templates/modules.conf.j2

@@ -0,0 +1,3 @@
+{% for module in kernel_modules %}
+{{ module }}
+{% endfor %}

packer/jupyter-repo2docker.pkr.hcl

@@ -30,7 +30,7 @@ variable "volume_type" {
 
 variable "volume_size" {
   type = number
-  default = 5
+  default = 10
 }
 
 variable "disk_format" {

packer/linux-desktop.pkr.hcl

@@ -30,7 +30,7 @@ variable "volume_type" {
 
 variable "volume_size" {
   type = number
-  default = 10
+  default = 15
 }
 
 variable "disk_format" {