All container image versions were defined in download/defaults/main.yml
except containerd.
The inconsistency caused the offline script(generate_list.sh) could not
output the URL of containerd image.
This moves the definition into a valid file.
In addition, this adds host_os to generate_list.sh for downloading
krew from a valid URL.
If trying to pull k8scsi/csi-resizer image from gcr.io, we face the error
like:
$ docker pull gcr.io/k8scsi/csi-resizer:v1.0.0
Error response from daemon: Head https://gcr.io/v2/k8scsi/csi-resizer/
manifests/v1.0.0: unknown: Project 'project:k8scsi' not found or deleted.
$
We can pull the image from quay.io instead.
This fixes the issue.
kube-bench scan outputs warning related to Calico like:
* text: "Ensure that the Container Network Interface file
permissions are set to 644 or more restrictive (Manual)"
* text: "Ensure that the Container Network Interface file
ownership is set to root:root (Manual)"
This fixes these warnings.
The addon-resizer container can reduce resource limits of cpu and
memory of metrics-server container in the pod, and that caused
OOMKilled.
In addition, the original metrics-server manifest doesn't contain
the addon-resizer container as [1].
So this adds metrics_server_resizer option to control the addon-resizer
container deployment and the default value is false to make it stable
for most environments.
[1]: https://github.com/kubernetes-sigs/metrics-server/blob/527679e5e8a103919c935d0575c20741796bc25d/manifests/base/deployment.yaml
The path of kubeconfig should be configurable, and its default value
is /etc/kubernetes/admin.conf. Most paths of the file are configurable
but some were not. This make those configurable.
non-kubeadm mode has been removed since ddffdb63bf
2.5 years ago. The non-kubeadm makes unnecessary confusion today, then
this updates the documentation.
Previously IDs of container images were gotten from tar files of container
images but that way was wrong. If multiple json files are contained in a
tar file, the script got multiple IDs and tried to pass these IDs on
`docker tag` command. Then the command was failed.
This updates the script to get image IDs from `docker image inspect` command
to fix this issue.
In addition, this adds a check a registry container exists already or not
before deploying registry container to avoid a container conflict failure.
To download necessary files in advance for offline deployment,
we can see all file URLs with contrib/offline/generate_list.sh
Most URLs are downloadable, but gvisor's one is not because the
URL is a part of full URLs for gvisor.
To download gvisor's files from the URLs directory, this separates
into two URLs for runsc and the shim.
tf-elax_ubuntu18-calico is so flake today. The test job is failed
due to SSH connectivity check error after deploying virtual machines
which are used for Kubernetes nodes.
This allows failure on the job to see the test situation without
pull request merger failures.
When running the script, I faced the following error but it was
difficult to know the root problem due to lack of error handling.
docker tag" requires exactly 2 arguments.
See 'docker tag --help'.
Usage: docker tag SOURCE_IMAGE[:TAG] TARGET_IMAGE[:TAG]
Create a tag TARGET_IMAGE that refers to SOURCE_IMAGE
To investigate such errors easily, this adds an error handling.
Basically we need to make necessary TCP/UDP ports open.
However the necessary ports are so many, and sometimes it is difficult
to figure out that is due to firewall issues or not if facing deployment
issues.
To distinguish a root problem on such situation, this adds contrib
playbook to disable the service firewall for Kubespray development
and test.
As the official document[1], the parameter keepcache should be
'0' or '1' as string. To avoid the following warning message,
this fixes the parameter value:
[WARNING]: The value False (type bool) in a string field was
converted to u'False' (type string). If this does not look
like what you expect, quote the entire value to ensure it
does not change.
https://docs.ansible.com/ansible/latest/collections/ansible/builtin/yum_repository_module.html
This replaces kube-master with kube_control_plane because of [1]:
The Kubernetes project is moving away from wording that is
considered offensive. A new working group WG Naming was created
to track this work, and the word "master" was declared as offensive.
A proposal was formalized for replacing the word "master" with
"control plane". This means it should be removed from source code,
documentation, and user-facing configuration from Kubernetes and
its sub-projects.
NOTE: The reason why this changes it to kube_control_plane not
kube-control-plane is for valid group names on ansible.
[1]: https://github.com/kubernetes/enhancements/blob/master/keps/sig-cluster-lifecycle/kubeadm/2067-rename-master-label-taint/README.md#motivation
upgrades.md explains how to do upgrade from v1.4.3 to v1.4.6 as an
example. The versions are a little old, and the doc readers would
have a concern the upgrade works fine or not.
This updates versions after verifying the way works fine by hands.
This replaces KUBE_MASTERS with KUBE_CONTROL_HOSTS because of [1]:
```
The Kubernetes project is moving away from wording that is
considered offensive. A new working group WG Naming was created
to track this work, and the word "master" was declared as offensive.
A proposal was formalized for replacing the word "master" with
"control plane". This means it should be removed from source code,
documentation, and user-facing configuration from Kubernetes and
its sub-projects.
```
[1]: https://github.com/kubernetes/enhancements/blob/master/keps/sig-cluster-lifecycle/kubeadm/2067-rename-master-label-taint/README.md#motivation
According to the following recommendation, this moves the directory
to control-plane:
The Kubernetes project is moving away from wording that is considered
offensive. A new working group WG Naming was created to track this work,
and the word "master" was declared as offensive. A proposal was formalized
for replacing the word "master" with "control plane".
TASK [Generate a list of information about the images on a node]
registers list of container images to docker_images.
Then the next TASK [Set pull_required if the desired image is not
yet loaded] does based on expecting images are registered.
However sometimes the first TASK was failed as [1] but the failure
is ignored due to failed_when:false and it makes another issue.
This removes this unnecessary failed_when to detect the failure
at the point.
In addition, this removes no_log:true also because the output doesn't
contain any sensitive data and now it just makes debugging difficult.
[1]: https://gitlab.com/kargo-ci/kubernetes-sigs-kubespray/-/jobs/934714534#L2953
This fixes the following failures:
./contrib/offline/README.md:14:1 MD014/commands-show-output Dollar signs used before commands without showing output [Context: "$ ./manage-offline-container-i..."]
./contrib/offline/README.md:20:1 MD014/commands-show-output Dollar signs used before commands without showing output [Context: "$ ./manage-offline-container-i..."]
Now markdownlint covers ./README.md and md files under ./docs only.
However we have a lot of md files under different directories also.
This enables markdownlint for other md files also.
Just after creating a namespace, the corresponding token could not be
created and sometimes the pod creation might be failed.
This adds check of the token in the new namespace to make this test
case stable.
tests/scripts/ansible-lint.sh was written on the doc, but there was
not such file actually. We can use ansible-lint command to check
ansible yml files without any options.
This updates to use the command.
If a branch name contains '.sh', current shellcheck checks the branch
file under .git/ and outputs error because the format is not shell
script one.
This makes shellcheck exclude files under .git/ to avoid this issue.
This fixes the following warning:
[kubernetes/client : Generate admin kubeconfig with external api endpoint]
[WARNING]: Consider using the file module with state=directory rather than
running 'mkdir'. If you need to use command because file is insufficient
you can
The task outputs the following warning:
TASK [kubernetes/preinstall : Enable ip forwarding]
[WARNING]: The value 1 (type int) in a string field was converted
to u'1' (type string). If this does not look like what you expect,
quote the entire value to ensure it does not change.
After upgrading to newer Kubernetes(v1.17 at least), kubectl command
shows the following warning message:
WARNING: Kubernetes configuration file is group-readable.
This is insecure. Location: /home/foo/.kube/config
The kubeconfig was copied from {{ artifacts_dir }}/admin.conf with
kubeconfig_localhost feature. It is better to set valid file mode
at getting it on Kubespray.
When stopping at the check of "Stop if ip var does not match local ips"
the error message is like:
fatal: [single-k8s]: FAILED! => {
"assertion": "ip in ansible_all_ipv4_addresses",
"changed": false,
"evaluated_to": false,
"msg": "Assertion failed"
}
That doesn't contain actual IP addresses and it is difficult to understand
what was wrong. This adds the error message which contain actual IP addresses
to investigate the issue if happens.
This changes MetalLB contrib to one of addons for deploying MetalLB with
Kubernetes cluster deployment. By the default, Kubespray doesn't deploy
MetalLB addon.
inventory_builder creates hosts.yaml file with hostnames like "node1",
"node2", etc. Even if specifying override_system_hostname=false, the
output of "kubectl get nodes" shows those hostnames ("node1", etc.)
without using actual hostnames.
To solve this issue, this adds an option USE_REAL_HOSTNAME to get
actual hostnames when creating hosts.yaml file instead of "node1", etc.
Now the in-tree cloud provider is deprecated and it is recommended to
the external cloud provider for OpenStack instead.
The doc described how to upgrade from the in-tree cloud provider, but
it is better to describe how to deploy the external cloud provider from
scratch instead for current situation.
This updates the OpenStack doc for this usecase.
We need to specify either external_openstack_tenant_name or
external_openstack_tenant_id. Those values were checked by seeing they
are defined or they have actual values separately.
However those values are always defined because of the following code
of openstack/defaults/main.yml:
external_openstack_tenant_id: "{{ lookup('env','OS_TENANT_ID')| default(lookup('env','OS_PROJECT_ID'),true) }}"
external_openstack_tenant_name: "{{ lookup('env','OS_TENANT_NAME')| default(lookup('env','OS_PROJECT_NAME'),true) }}"
So even if not specifying both values, those checks could not detect
the misconfiguration. This fixes this to detect the misconfiguration.
On OpenStack history, we used to call "tenant" for separeted namespace.
However we use "project" now instead.
Then we have replaced "tenant" with "project". Then all "TENANT" variables
also are renamed to "PROJECT".
This makes Kubespray search "PROJECT" variable also for newer OpenStack
clouds.
If running MetalLB v0.7.3 on k8s v1.18.2, metallb pods output the
following parsing error of v1.ServiceList:
$ kubectl logs controller-dbb46cf84-fw8h8 -n metallb-system
{
"caller":"reflector.go:205",
"level":"error",
"msg":"go.universe.tf/metallb/internal/k8s/k8s.go:231:
Failed to list *v1.Service: v1.ServiceList:
Items: []v1.Service: v1.Service: ObjectMeta:
v1.ObjectMeta: readObjectFieldAsBytes:
expect : after object field, parsing 1605
Then an external IP address is never allocated to the Service of
LoadBalancer type.
By updating MetalLB version to the latest v0.9[1] today, this issue
can be solved.
[1]: https://hub.docker.com/r/metallb/controller/tags
This updates MetalLB README as following
- Remove unnecessary markdown to read it easily on github
- Make words consistency (kubernetes, loadbalancer)
- Add change-required option
Due to lack of requirements installation on Azure README, the error
can happen:
"The ipaddr filter requires python's netaddr be installed on the
ansible controller"
It is nice to add the installation for Azure users.
apply-rg.sh was for Azure command version 1("azure" command) and the
command is old and version 2("az" command) is officially used today.
apply-rg_2.sh was for the version 2. In addition, the README[1] says
we need to run apply-rg.sh for applying templates.
This renames apply-rg_2.sh to apply-rg.sh for common usages of the
version 2.
[1]: https://github.com/kubernetes-sigs/kubespray/tree/master/contrib/azurerm#generating-and-applying
The ansible-playbook needs to ssh-login to Azure virtual machines with
ssh keypair, and users need to specify ssh_public_keys for their own
ssh public key. The change of ssh_public_keys is mandatory.
So this updates contrib/azurerm/README.md to explain that.
In addition, the path of all.yml was wrong. That also is updated with
this.
apply-rg_2.sh uses 'az group deployment' command but the command is
deprecated like the following warning message:
"This command is implicitly deprecated because command group
'group deployment' is deprecated and will be removed in a future release.
Use 'deployment group' instead."
This updates these deprecated commands.
FYI: The command has been deprecated since [1] on azure-cli side.
[1]: https://github.com/Azure/azure-cli/commit/991cb7cc7c9e577d8bec3b78e0316dfb5d39fe10#diff-2057bbb8441166e4910b34b09d22b58cR222
cloud_provider option exists in ./inventory/sample/group_vars/all/all.yml
In addition, the quick start shows to create configuration by copying
./inventory/sample. So this updates path of all.yml for fitting the above.
Kenichi Omichi