kubespray/docs/ansible.md

215 lines
8.9 KiB
Markdown
Raw Normal View History

2019-12-04 23:22:57 +08:00
# Ansible variables
2016-07-04 20:13:18 +08:00
2019-12-04 23:22:57 +08:00
## Inventory
2016-07-04 20:13:18 +08:00
The inventory is composed of 3 groups:
2016-07-04 20:13:18 +08:00
* **kube_node** : list of kubernetes nodes where the pods will run.
* **kube_control_plane** : list of servers where kubernetes control plane components (apiserver, scheduler, controller) will run.
2017-02-14 18:08:27 +08:00
* **etcd**: list of servers to compose the etcd server. You should have at least 3 servers for failover purpose.
2016-07-04 20:13:18 +08:00
Note: do not modify the children of _k8s_cluster_, like putting
the _etcd_ group into the _k8s_cluster_, unless you are certain
to do that and you have it fully contained in the latter:
2019-12-04 23:22:57 +08:00
```ShellSession
k8s_cluster ⊂ etcd => kube_node ∩ etcd = etcd
```
When _kube_node_ contains _etcd_, you define your etcd cluster to be as well schedulable for Kubernetes workloads.
If you want it a standalone, make sure those groups do not intersect.
If you want the server to act both as control-plane and node, the server must be defined
on both groups _kube_control_plane_ and _kube_node_. If you want a standalone and
unschedulable control plane, the server must be defined only in the _kube_control_plane_ and
not _kube_node_.
There are also two special groups:
* **calico_rr** : explained for [advanced Calico networking cases](calico.md)
* **bastion** : configure a bastion host if your nodes are not directly reachable
Below is a complete inventory example:
2016-07-04 20:13:18 +08:00
2019-12-04 23:22:57 +08:00
```ini
2016-07-04 20:13:18 +08:00
## Configure 'ip' variable to bind kubernetes services on a
## different ip than the default iface
node1 ansible_host=95.54.0.12 ip=10.3.0.1
node2 ansible_host=95.54.0.13 ip=10.3.0.2
node3 ansible_host=95.54.0.14 ip=10.3.0.3
node4 ansible_host=95.54.0.15 ip=10.3.0.4
node5 ansible_host=95.54.0.16 ip=10.3.0.5
node6 ansible_host=95.54.0.17 ip=10.3.0.6
2016-07-04 20:13:18 +08:00
[kube_control_plane]
2016-07-04 20:13:18 +08:00
node1
node2
[etcd]
node1
node2
node3
[kube_node]
2016-07-04 20:13:18 +08:00
node2
node3
node4
node5
node6
[k8s_cluster:children]
kube_node
kube_control_plane
2016-07-04 20:13:18 +08:00
```
2019-12-04 23:22:57 +08:00
## Group vars and overriding variables precedence
The group variables to control main deployment options are located in the directory ``inventory/sample/group_vars``.
Optional variables are located in the `inventory/sample/group_vars/all.yml`.
Mandatory variables that are common for at least one role (or a node group) can be found in the
`inventory/sample/group_vars/k8s_cluster.yml`.
There are also role vars for docker, kubernetes preinstall and control plane roles.
According to the [ansible docs](https://docs.ansible.com/ansible/latest/playbooks_variables.html#variable-precedence-where-should-i-put-a-variable),
2018-10-26 21:49:57 +08:00
those cannot be overridden from the group vars. In order to override, one should use
2019-12-04 23:22:57 +08:00
the `-e` runtime flags (most simple way) or other layers described in the docs.
2017-06-17 01:25:46 +08:00
Kubespray uses only a few layers to override things (or expect them to
2018-10-26 21:49:57 +08:00
be overridden for roles):
Layer | Comment
------|--------
2017-06-17 01:25:46 +08:00
**role defaults** | provides best UX to override things for Kubespray deployments
inventory vars | Unused
**inventory group_vars** | Expects users to use ``all.yml``,``k8s_cluster.yml`` etc. to override things
inventory host_vars | Unused
2017-06-12 19:20:15 +08:00
playbook group_vars | Unused
playbook host_vars | Unused
2017-06-17 01:25:46 +08:00
**host facts** | Kubespray overrides for internal roles' logic, like state flags
play vars | Unused
play vars_prompt | Unused
play vars_files | Unused
registered vars | Unused
2017-06-17 01:25:46 +08:00
set_facts | Kubespray overrides those, for some places
**role and include vars** | Provides bad UX to override things! Use extra vars to enforce
2017-06-17 01:25:46 +08:00
block vars (only for tasks in block) | Kubespray overrides for internal roles' logic
task vars (only for the task) | Unused for roles, but only for helper scripts
**extra vars** (always win precedence) | override with ``ansible-playbook -e @foo.yml``
2019-12-04 23:22:57 +08:00
## Ansible tags
The following tags are defined in playbooks:
| Tag name | Used for
|--------------------------|---------
| apps | K8s apps definitions
| azure | Cloud-provider Azure
2016-12-12 21:53:33 +08:00
| bastion | Setup ssh config for bastion
| bootstrap-os | Anything related to host OS configuration
| calico | Network plugin Calico
| canal | Network plugin Canal
| cloud-provider | Cloud-provider related tasks
| docker | Configuring docker for hosts
| download | Fetching container images to a delegate host
| etcd | Configuring etcd cluster
| etcd-pre-upgrade | Upgrading etcd cluster
| etcd-secrets | Configuring etcd certs/keys
| etchosts | Configuring /etc/hosts entries for hosts
| facts | Gathering facts and misc check results
| flannel | Network plugin flannel
| gce | Cloud-provider GCP
| k8s-pre-upgrade | Upgrading K8s cluster
| k8s-secrets | Configuring K8s certs/keys
| kube-apiserver | Configuring static pod kube-apiserver
| kube-controller-manager | Configuring static pod kube-controller-manager
| kubectl | Installing kubectl and bash completion
| kubelet | Configuring kubelet service
| kube-proxy | Configuring static pod kube-proxy
| kube-scheduler | Configuring static pod kube-scheduler
| localhost | Special steps for the localhost (ansible runner)
| master | Configuring K8s master node role
| netchecker | Installing netchecker K8s app
| network | Configuring networking plugins for K8s
| nginx | Configuring LB for kube-apiserver instances
| node | Configuring K8s minion (compute) node role
| openstack | Cloud-provider OpenStack
| preinstall | Preliminary configuration steps
| resolvconf | Configuring /etc/resolv.conf for hosts/apps
| upgrade | Upgrading, f.e. container images/binaries
| upload | Distributing images/binaries across hosts
| weave | Network plugin Weave
| ingress_alb | AWS ALB Ingress Controller
| ambassador | Ambassador Ingress Controller
Note: Use the ``bash scripts/gen_tags.sh`` command to generate a list of all
tags found in the codebase. New tags will be listed with the empty "Used for"
field.
2019-12-04 23:22:57 +08:00
## Example commands
Example command to filter and apply only DNS configuration tasks and skip
everything else related to host OS configuration and downloading images of containers:
2019-12-04 23:22:57 +08:00
```ShellSession
ansible-playbook -i inventory/sample/hosts.ini cluster.yml --tags preinstall,facts --skip-tags=download,bootstrap-os
```
2019-12-04 23:22:57 +08:00
And this play only removes the K8s cluster DNS resolver IP from hosts' /etc/resolv.conf files:
2019-12-04 23:22:57 +08:00
```ShellSession
ansible-playbook -i inventory/sample/hosts.ini -e dns_mode='none' cluster.yml --tags resolvconf
```
2019-12-04 23:22:57 +08:00
2018-08-22 22:40:17 +08:00
And this prepares all container images locally (at the ansible runner node) without installing
or upgrading related stuff or trying to upload container to K8s cluster nodes:
2019-12-04 23:22:57 +08:00
```ShellSession
ansible-playbook -i inventory/sample/hosts.ini cluster.yml \
-e download_run_once=true -e download_localhost=true \
--tags download --skip-tags upload,upgrade
```
Note: use `--tags` and `--skip-tags` wise and only if you're 100% sure what you're doing.
2016-12-09 17:57:50 +08:00
2019-12-04 23:22:57 +08:00
## Bastion host
2016-12-09 17:57:50 +08:00
If you prefer to not make your nodes publicly accessible (nodes with private IPs only),
you can use a so called *bastion* host to connect to your nodes. To specify and use a bastion,
simply add a line to your inventory, where you have to replace x.x.x.x with the public IP of the
bastion host.
2019-12-04 23:22:57 +08:00
```ShellSession
[bastion]
bastion ansible_host=x.x.x.x
2016-12-09 17:57:50 +08:00
```
For more information about Ansible and bastion hosts, read
[Running Ansible Through an SSH Bastion Host](https://blog.scottlowe.org/2015/12/24/running-ansible-through-ssh-bastion-host/)
## Mitogen
You can use [mitogen](mitogen.md) to speed up kubespray.
## Beyond ansible 2.9
Ansible project has decided, in order to ease their maintenance burden, to split between
two projects which are now joined under the Ansible umbrella.
Ansible-base (2.10.x branch) will contain just the ansible language implementation while
ansible modules that were previously bundled into a single repository will be part of the
ansible 3.x package. Pleasee see [this blog post](https://blog.while-true-do.io/ansible-release-3-0-0/)
that explains in detail the need and the evolution plan.
**Note:** this change means that ansible virtual envs cannot be upgraded with `pip install -U`.
You first need to uninstall your old ansible (pre 2.10) version and install the new one.
```ShellSession
pip uninstall ansible
cd kubespray/
pip install -U .
```
**Note:** some changes needed to support ansible 2.10+ are not backwards compatible with 2.9
Kubespray needs to evolve and keep pace with upstream ansible and will be forced to eventually
drop 2.9 support. Kubespray CIs use only the ansible version specified in the `requirements.txt`
and while the `ansible_version.yml` may allow older versions to be used, these are not
exercised in the CI and compatibility is not guaranteed.