Commit Graph

7167 Commits (0ba2e655f4b96bca6316663c91cbd2d66839973b)
 

Author SHA1 Message Date
Olivier Lemasle 507091ec8b
Replace cluster_name by dns_domain (#7923)
`cluster_name` defaults to `dns_domain` value (see [here][1] and [here][2])
but they could have different values.

`dns_domain` should be used here instead of `cluster_name` because the DNS
resolution is configured to use `dns_domain`.

[1]: 0ef7af76bc/roles/kubespray-defaults/defaults/main.yaml (L104)
[2]: 1afdb05ea9/inventory/sample/group_vars/k8s_cluster/k8s-cluster.yml (L196)
2021-09-01 08:18:59 -07:00
Florian Ruynat c7529270ff
Fix CI script for Terraform >0.15 (#7928) 2021-09-01 04:30:59 -07:00
Maciej Wereski 48ceca4919
MetalLB: update to v0.10.2 (#7925)
Signed-off-by: Maciej Wereski <m.wereski@partner.samsung.com>
2021-09-01 03:00:59 -07:00
Florian Ruynat 0171c71de0 Update Terraform 0.14 to .11, remove 0.13 jobs and add 0.15 2021-08-31 16:32:59 -07:00
Florian Ruynat 46d0df394f Add one retry to packet_pr jobs 2021-08-31 16:32:59 -07:00
Florian Ruynat 207d3e7b4e
Add Debian-11 image and CI (#7919) 2021-08-31 14:02:22 -07:00
Cristian Calin 426ad81db0
Calico: replace hashes for latest 3.17 and 3.18 to the .5 minor versions (#7924) 2021-08-31 13:38:21 -07:00
Olivier Lemasle 497d2ca306
Fix Calico's FelixConfiguration when "IP in IP" is disabled (#7926)
When using Calico with:

- `calico_network_backend: vxlan`,
- `calico_ipip_mode: "Never"`,
- `calico_vxlan_mode: "Always"`,

the `FelixConfiguration` object has `ipipEnabled: true`, when it should be false:

This is caused by an error in the `| bool` conversion in the install task:
when `calico_ipip_mode` is `Never`,
`{{ calico_ipip_mode != 'Never' | bool }}` evaluates to `true`:
2021-08-31 13:14:21 -07:00
Calvin Park 9d3888a756
During pre-upgrade add a flag to always cordon (#7892)
* During pre-upgrade add a flag to always cordon

* empty

* empty

* empty

* Better default val
2021-08-30 10:56:09 -07:00
rtsp c8e090c17f
Add preliminary Debian 11 (bullseye) support (#7853)
- Use python3-apt instead because python-apt was removed in Debian 11
- Add gnupg (fix "container-engine/containerd : ensure containerd repository public key is installed" task failed)
- Remove aufs-tools

Signed-off-by: rtsp <git@rtsp.us>
2021-08-30 09:53:06 -07:00
Florian Ruynat 77a74adedd
Bump centos8 CI job memory to 3go and remove mitogen for fedora CI (#7921) 2021-08-30 08:25:13 -07:00
Florian Ruynat 1ccf32e08f
Update docker to 20.10.8 (#7918) 2021-08-30 08:25:06 -07:00
Florian Ruynat b5aced20e1 Update Kubernetes version to 1.21.4 2021-08-30 08:17:05 -07:00
Florian Ruynat 17af348be8 Add bunch of Kubernetes versions missing 2021-08-30 08:17:05 -07:00
Cristian Calin 1afdb05ea9
Fedora and RHEL use etc_t and the convention is <type_name>_t (#7891)
* Fedora and RHEL use etc_t and the convention is <type_name>_t

* Docs: specify all values for preinstall_selinux_state

* CI: Add Fedora 34 with SELinux in enforcing mode
2021-08-27 14:20:53 -07:00
Olivier Clavel 425b6741c6
Fix failed image build on pip installing ansible (#7862)
Related pip bug: https://stackoverflow.com/questions/68687029/unable-to-build-kubespray-container-from-dockerfile
Proposed workaround in comment: https://github.com/pypa/pip/issues/10219#issuecomment-887337037
Setting LANG only prior to launching pip fixes the issue with a successful build
2021-08-26 07:47:23 -07:00
Léopold Jacquot d635961120
Add Infomaniak to compatible public clouds list (#7910) 2021-08-26 06:47:24 -07:00
Hari Hud d5b865da4d
Update etcd without rotating etcd certs (#7907) 2021-08-26 00:21:23 -07:00
Sergey 89993e4833
fix error metrics server capabilities name (#7905) 2021-08-25 12:06:15 -07:00
jayonlau 6b5da84014
Clean up extra spaces last one (#7904)
Although these errors are not important, they affect the code specification.
2021-08-25 12:06:09 -07:00
Cristian Calin 1c3d33e146
Calico: 3.20.0 policy update to allow access to endpointslices (#7899) 2021-08-25 12:06:01 -07:00
Pratik Raj 71af4b4a85
chore : use --no-cache-dir flag to pip in dockerfiles to save space (#7898)
using --no-cache-dir flag in pip install ,make sure downloaded packages
by pip don't cached on system . This is a best practice which make sure
to fetch from repo instead of using local cached one . Further , in case
of Docker Containers , by restricting caching , we can reduce image size.
In term of stats , it depends upon the number of python packages
multiplied by their respective size . e.g for heavy packages with a lot
of dependencies it reduce a lot by don't caching pip packages.

Further , more detail information can be found at

https://medium.com/sciforce/strategies-of-docker-images-optimization-2ca9cc5719b6

Signed-off-by: Pratik Raj <rajpratik71@gmail.com>
2021-08-25 12:05:55 -07:00
Sergey c49dd50ef3
add tags: always to all included sevice playbook (#7906) 2021-08-25 12:01:54 -07:00
Cristian Calin f66c49bf42
Calico: replace version 3.19.1 with 3.19.2 and set as default (#7867)
Bump calico version to 3.19.2 due to adding 3.20.0 earlier
2021-08-25 07:32:41 -07:00
rtsp 4c9d7dedb3
addons/cert_manager: retries until webhook pods has been created (#7850)
Fix task 'Cert Manager | Wait for Webhook pods become ready' failed due to webhook pods don't exist yet by using `retries..until` trick like kubernetes-sigs/kubespray#7842

This fix should be removed in the future if the kubernetes/kubernetes#83242 is resolved.

Signed-off-by: rtsp <git@rtsp.us>
2021-08-25 07:16:41 -07:00
Sergey 5336943a8c
add cilium_operator_api_serve_addr to cilium operator config (#7901) 2021-08-24 03:49:13 -07:00
Baargav 9dfade5641
Update nodes.md (#7902) 2021-08-24 02:43:14 -07:00
Samuel a040e521b4
feat(containerd): auth support (#7868)
* feat(containerd): auth support

* fix(registry-auth): rename variable
2021-08-23 06:40:00 -07:00
Hrittik Roy dad4b26c6f
Update Azure.md (#7880) 2021-08-20 20:23:58 -07:00
Cristian Calin 0ac364dfae
Calico: use --allow-version-mismatch in calicoctl.sh to allow upgrades (#7873) 2021-08-20 14:30:48 -07:00
rtsp dfd35892f2
docs/cert_manager.md: Update docs for K8s v1.22 (#7877) 2021-08-19 18:31:24 -07:00
rtsp 79166496f3
debian: Fix test failed after bullseye release (#7888) 2021-08-19 15:37:24 -07:00
Victor Morales c7d12cddec
Ensure python main function return values (#7860)
The main functions are wrapped by a sys.exit function which expects and
argument. The curent implementation isn't returning values in all cases.
This change ensures main functions return a value in all cases.
2021-08-19 06:51:24 -07:00
Frank Ritchie 1f09229740
Update cilium to 1.9.9 (#7871)
Now that 1.10 is out this is to make 1.9.9 the default. I am running
this version successfully.
2021-08-16 13:34:22 -07:00
Victor Morales c2d4700571
Remove unused python imports (#7859) 2021-08-13 13:35:32 -07:00
Léopold Jacquot c06896a352
Update metrics-server to 0.5.0 (#7864) 2021-08-12 08:19:48 -07:00
Cristian Calin c119620f7c
Calico: add v3.20.0 hashes (#7855) 2021-08-11 07:50:46 -07:00
Daniil Muidinov 7f309bb092
fix parameters for module replace in 0060-resolvconf (#7858) 2021-08-10 17:13:26 -07:00
Eugene Artemenko e2b67b5700
Add suport of Vsphere CSI driver 2.2.X versions (#7848) 2021-08-09 08:19:38 -07:00
rtsp 82a9064d8d
addons/cert_manager: fix kubernetes-sigs#7085 by adding retries..until (#7842)
Fix task 'Cert Manager | Apply ClusterIssuer manifest' failed due to service/endpoints updating delayed even though the wekhook pod status is ready.

Signed-off-by: rtsp <git@rtsp.us>
2021-08-09 08:19:31 -07:00
Victor Morales a70fab2249
Bump crun to 0.21 version (#7854) 2021-08-09 08:11:31 -07:00
Shubham Deshmukh 86b45fce6a
Remove environment variable in remove-node play (#7729) 2021-08-02 04:29:21 -07:00
Smita Srivastava 31a5a4e808
retry to fetch binary if it fails first time (#7839) 2021-07-30 00:17:38 -07:00
Vitaliy D 5db86f4c2b
Update vSphere CPI (#7838)
Changes:
  * ClusterRole updated according to the latest manifests from
    https://github.com/kubernetes/cloud-provider-vsphere
  * vSphere CPI/CSI default versions bumped and
    tested successfully on K8S 1.21.1
  * vSphere documentation updated

Signed-off-by: Vitaliy D <vi7alya@gmail.com>
2021-07-29 18:17:37 -07:00
Kenichi Omichi 20c284c276
doc: Update 'Kubespray vs Kubeadm' (#7834)
non-kubeadm mode has been removed since ddffdb63bf
2.5 years ago. The non-kubeadm makes unnecessary confusion today, then
this updates the documentation.
2021-07-28 03:15:34 -07:00
Markus Opolka befc6cd650
Update MetalLB documentation (#7833)
- Added a hint about the kube_proxy_strict_arp configuration, which is required for MetalLB to work
 - See also https://github.com/kubernetes-sigs/kubespray/pull/5180/files
2021-07-27 08:46:45 -07:00
Maxime Guyot 97d95775a5
Disable OVH CI until voucher situation is cleared up (#7824) 2021-07-26 06:16:33 -07:00
Kenichi Omichi 8f44cd35d8
Fix how to get image ID on offline deployment (#7808)
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.
2021-07-26 00:56:33 -07:00
AnatomicJC 627a06e30d
CRI-O: Install libseccomp2 from backports on Debian 10 (#7816)
* CRI-O: Install libseccomp2 from backports on Debian 10

libseccomp2 is a required dependency of cri-o-runc package

The one provided in Debian 10 repositories is outdated

* 7816: Remove useless when condition

As this condition is handled by block
2021-07-23 07:07:16 -07:00
Samuel bfebcfa2c5
fix(misc): contrib/terraform/aws (#7818)
* fix(misc): terraform/aws

- handles deployment with a single availability zone
- handles deployment with more than two availability zone
- handles etcd collocation with control-plane nodes (`aws_etcd_num=0`)
- allows to set a bastion instances count (`aws_bastion_num`)
- allows to set bastion/etcd/control-plane/workers rootfs volume size
- removes variables from terraform.tfvars that were not re-used
- adds .terraform.lock.hcl to .gitignore
- changes/updates base image from ubuntu-18.03 to debian-10

tested by a few coworkers of mine, and myself: thanks for the outstanding
work, on both those terraform samples and kubespray playbooks.
I did not test ubuntu deployments, I could still swap from buster to
focal. LMK.

* fix(gitlab-ci)

AFAIU, terraform.tfvars indentation should be fixed for / no diff
returned running `terraform fmt -check -diff`

https://gitlab.com/kargo-ci/kubernetes-sigs-kubespray/-/jobs/1445622114
2021-07-23 02:43:16 -07:00