Refactor NRI (Node Resource Interface) activation in CRI-O and
containerd. Introduce a shared variable, nri_enabled, to streamline
the process. Currently, enabling NRI requires a separate update of
defaults for each container runtime independently, without any
verification of NRI support for the specific version of containerd
or CRI-O in use.
With this commit, the previous approach is replaced. Now, a single
variable, nri_enabled, handles this functionality. Also, this commit
separates the responsibility of verifying NRI supported versions of
containerd and CRI-O from cluster administrators, and leaves it to
Ansible.
Signed-off-by: Feruzjon Muyassarov <feruzjon.muyassarov@intel.com>
* [containerd] Add Configuration option for Node Resource Interface
Node Resource Interface (NRI) is a common is a common framework for
plugging domain or vendor-specific custom logic into container
runtime like containerd. With this commit, we introduce the
containerd_disable_nri configuration flag, providing cluster
administrators the flexibility to opt in or out (defaulted to 'out')
of this feature in containerd. In line with containerd's default
configuration, NRI is disabled by default in this containerd role
defaults.
Signed-off-by: Feruzjon Muyassarov <feruzjon.muyassarov@intel.com>
* [cri-o] Add configuration option for Node Resource Interface
Node Resource Interface (NRI) is a common is a common framework for
plugging domain or vendor-specific custom logic into container
runtimes like containerd/crio. With this commit, we introduce the
crio_enable_nri configuration flag, providing cluster
administrators the flexibility to opt in or out (defaulted to 'out')
of this feature in cri-o runtime. In line with crio's default
configuration, NRI is disabled by default in this cri-o role
defaults.
Signed-off-by: Feruzjon Muyassarov <feruzjon.muyassarov@intel.com>
---------
Signed-off-by: Feruzjon Muyassarov <feruzjon.muyassarov@intel.com>
* Fix containerd_registries in config_path for mirrors and remove nerdctl global insecure_registry setting
* Make containerd hosts.toml mode 0640
* Add containerd_registries_mirrors and keep containerd_registries to pass packet_debian11-calico-upgrade
Set owner/group to root/root when unarchiving kata-containers binary to prevent kata-containers binaries/directories and especially / from getting chowned to 1001:123, the file owner specified in the kata-containers archive
* tests: replace fedora35 with fedora37
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* tests: replace fedora36 with fedora38
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* docs: update fedora version in docs
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* molecule: upgrade fedora version
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* tests: upgrade fedora images for vagrant and kubevirt
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* vagrant: workaround to fix private network ip address in fedora
Fedora stop supporting syconfig network script so we added a workaround
here
https://github.com/hashicorp/vagrant/issues/12762#issuecomment-1535957837
to fix it.
* netowrkmanager: do not configure dns if using systemd-resolved
We should not configure dns if we point to systemd-resolved.
Systemd-resolved is using NetworkManager to infer the upstream DNS
server so if we set NetworkManager to 127.0.0.53 it will prevent
systemd-resolved to get the correct network DNS server.
Thus if we are in this case we just don't set this setting.
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* image-builder: update centos7 image
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* gitlab-ci: mark fedora packet jobs as allow failure
Fedora networking is still broken on Packet, let's mark it as allow
failure for now.
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
---------
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix ansible-lint name
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: ignore jinja template error in names
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: capitalize ansible name
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: update notify after name capitalization
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
---------
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix var-spacing ansible rule
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix spacing on the beginning/end of jinja template
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix spacing of default filter
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix spacing between filter arguments
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix double space at beginning/end of jinja
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: fix remaining jinja[spacing] ansible-lint warning
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
---------
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: update all dependencies including ansible
Upgrade to ansible 7.x and ansible-core 2.14.x. There seems to be issue
with ansible 8/ansible-core 2.15 so we remain on those versions for now.
It's quite a big bump already anyway.
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* tests: install aws galaxy collection
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* ansible-lint: disable various rules after ansible upgrade
Temporarily disable a bunch of linting action following ansible upgrade.
Those should be taken care of separately.
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve deprecated-module ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve no-free-form ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve schema[meta] ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve schema[playbook] ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve schema[tasks] ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve risky-file-permissions ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve risky-shell-pipe ansible-lint error
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: remove deprecated warn args
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: use fqcn for non builtin tasks
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: resolve syntax-check[missing-file] for contrib playbook
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
* project: use arithmetic inside jinja to fix ansible 6 upgrade
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
---------
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
Molecule 5.0 require ansible-core 2.12.10.
So this commit we update ansible-core from 2.12.5 to 2.12.10.
We also drop supporting two ansible-core version. Also we now use the "oldest"
still supported ansible-core version as both 2.11 is EOL and not
supported by molecule.
tests/molecule: remove linting in molecule to support molecule 5
tests/molecule: remove role name check for molecule 5 support
Kubespray doesn't use ansible galaxy style naming so we have to disable
that check.
contrib/inventory_builder: fix tox.ini for tox4
tests/molecule: fix get_playbook in testinfra tests
tests: upgrade most tests requirements
Exclude ansible-lint for now, I will do that in a separate PR.
tests/molecule: force kvm driver option
If we don't do this it fallbacks to qemu emulated on our CI for some
reasons.
Signed-off-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@proton.ch>
- Test with new version: 37.20230322.3.0. Both containerd and
cri-o is tested
- bugfix: when we use crio and the var bin_dir is changed,
there will be some error about the new bin dir.
crun_bin_dir was used to specify the destination of the crun binary during the
download process. This path must match with the value provided in the CRI-O
configuration file. So changing its value to bin_dir helps to mismatch errors.
Signed-off-by: Victor Morales <chipahuac@hotmail.com>
* optimize cgroups settings for node reserved
* fix
* set cgroup slice for multi container engine
* set cgroup slice for crio
* add reserved cgroups variables to sample files
* Compatible with cgroup path for different container managers
* add cgroups doc
* fix markdown
The `containerd-common` role is responsible for gathering OS specific variables from the vars directory of the roles that include or import it. `containerd-common` is imported via role dependency by a total of two roles, `container-engine/docker`, and `container-engine/containerd`.
containerd-common is needed by both the docker and containerd roles as a dependency when:
- containerd is selected as the container engine
- a docker install is detected and needs to be removed
- apt is the package manager
However, by default, roles can not be invoked more than once in the same play, unless `allow_duplicates: true` is set for that role. This results in the failure of the `containerd | Remove containerd repository` task, since only the docker vars will be loaded in the play, and `containerd_repo_info.repos`, normally populated by containerd/vars, is left empty.
This change sets `allow_duplicates: true` for `containerd-common` which fixes the currently failing containerd tasks if docker was detected and removed in the same play.
by setting a default runtime spec with a patch for RLIMIT_NOFILE.
- Introduces containerd_base_runtime_spec_rlimit_nofile.
- Generates base_runtime_spec on-the-fly, to use the containerd version
of the node.
For the following configuration
```
containerd_insecure_registries:
docker.io:
- dockerhubcache.example.com
```
the rendered /etc/containerd/config.toml contains
```
[plugins."io.containerd.grpc.v1.cri".registry.configs."docker.io".tls]
insecure_skip_verify = true
```
but it needs to be
```
[plugins."io.containerd.grpc.v1.cri".registry.configs."dockerhubcache.example.com".tls]
insecure_skip_verify = true
```
* Update main.yaml
* remove version in dpkg_selection name
* make lint happy
* Fix typo
* add comment / remove useless contition
* remove dpkg hold in reset tasks
* feat: make kubernetes owner parametrized
* docs: update hardening guide with configuration for CIS 1.1.19
* fix: set etcd data directory permissions to be compliant to CIS 1.1.12
* extra admission controls now don't have a version in their file names
eventratelimit.v1beta2.yaml.j2 -> eventratelimit.yaml.j2
* cri_socket variable includes the unix:// prefix to be conformat with
upstream
When running molecule jobs, we saw the folloing warning message:
[DEPRECATION WARNING]: [defaults]callback_whitelist option, normalizing names
to new standard, use callbacks_enabled instead. This feature will be removed
from ansible-core in version 2.15. Deprecation warnings can be disabled by
setting deprecation_warnings=False in ansible.cfg.
callbacks_enabled has been added since Ansible 2.11 and Kubespray is using
Ansible 2.12 at master branch. So we can use callbacks_enabled safely to
avoid the warning message.
* Force containerd service unmasking
Force systemd to unmask and start service when adding containerd service
* Eliminate restart and move unmasking step
Switch to start instead of restart
Move unmasking to restart handler
* Add unmasking to similar container runtimes
* Add missing service names
When running cluster.yml for new machines what containerd is already
install but Kubernetes cluster were not installed before, the task
"remove-node | List nodes" is failed like
"changed": false,
"cmd": [
"/usr/local/bin/kubectl", "--kubeconfig",
"/etc/kubernetes/admin.conf", "get", "nodes", "-o",
"go-template={{ range .items }}{{ .metadata.name }}
{{ "\n" }}{{ end }}"
],
..
"stderr": "error: stat /etc/kubernetes/admin.conf: no such file or directory",
That was due to lack to check the existing Kubernetes cluster exists
or not before running "kubectl drain" command.
This adds the check to avoid the issue.
* Add epoch to docker-ce and docker-ce-cli packages to ensure docker upgrade
* Split container-engine redhat vars to support legacy RHEL 7 version management
* Support ansible_distribution_major_version when disvering vars with ansible_os_family
Heather Lapointe
Feruzjon Muyassarov
Feruzjon Muyassarov
Florian Ruynat
yun
Arthur Outhenin-Chalandre
ERIK
Nico
Victor Morales
Louis Tu
yun
qlijin
James
Denis
pingrulkin
DRAGON2002
Marijn van der Giesen
rongfu.leng
Samuel BECK
Bas
rongfu.leng
Kay Yan
Shelming.Song
Fredrik Liv
Chad Swenson
Ilya Margolin
lijin-union
Cristian Calin
Kenichi Omichi
Krystian Młynek
Pavel Chekin
Mostafa Ghadimi
rptaylor
Alessio Greggi
Calin Cristian Andrei
orange-llajeanne
Thearas
Chris Ricker
cyril-corbon
Sergey
Unai Arríen
Branko Mijuskovic