https://github.com/alancoding/playbooks
playbooks for debugging and messing with a live AAP controller cluster
https://github.com/alancoding/playbooks
Last synced: 4 months ago
JSON representation
playbooks for debugging and messing with a live AAP controller cluster
- Host: GitHub
- URL: https://github.com/alancoding/playbooks
- Owner: AlanCoding
- Created: 2021-11-18T20:08:16.000Z (over 3 years ago)
- Default Branch: devel
- Last Pushed: 2024-03-24T02:32:25.000Z (over 1 year ago)
- Last Synced: 2025-01-03T05:38:43.287Z (6 months ago)
- Language: Shell
- Size: 43 KB
- Stars: 0
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
### Live patches for AWX and AAP controller
This has Ansible content to help speed up modifying and testing live servers.
#### Configuration of inventory file
These playbooks require a valid AAP inventory file.
This will have groups like `automationcontroller` and `execution_nodes`.
If using the docker-compose inventory file, there is an inventory script floating
around which will build an inventory of this format from `docker ps -a` data.
Specifying the inventory file is going to get old, so this will pick up a symlink
at the following location `awx_inventory.ini`. For example, to set this up:
```
ln -s ~/repos/tower-qa/playbooks/inventory-docker.py awx_inventory.ini
```
You can still specify a different inventory with `-i` if you need.
All the following commands will assume this is set up.
#### Action - create a superuser
Before anything else, you will need a user. Fill in values at `vars/private_vars.yml`
```yaml
awx_username: admin
awx_password: password
awx_email: "[email protected]"
```
Playbook example:
```
ansible-playbook prod_user.yml
```
#### Action - turn on debugging settings
You want to run a job and have the server preserve the job folders and work unit.
```
ansible-playbook -e debug_state=present debug_on.yml
```
Switch `debug_state=present` to absent to undo this.
In the development environment, this will modify the file on your _host_ machine
at `tools/docker-compose/_sources/local_settings.py`, which is templated by
the docker-compose playbook. This is hard to track, so be warned.
#### Action - replace receptor binary
Before you go changing anything, you probably want to print the current versions
that all of the nodes have installed.
```
ansible-playbook --tags=version receptor_bin_swap.yml
```
Tested on Fedora, go to receptor clone, check out the branch you need
```
sudo dnf install -y golang
make
```
This should populate the receptor binary at the top-level of the receptor checkout.
Assuming checkout is now at `~/repos/receptor/receptor` then
```
ansible-playbook -e receptor_bin=~/repos/receptor/receptor receptor_bin_swap.yml
```
TODO: still need a playbook that will replace the `receptorctl` package.
#### Action - replace the ansible-runner install
This follows a similar pattern to replacing receptor. First look at what is installed:
```
ansible-playbook --tags=version runner_swap.yml
```
A typical pattern is that you would checkout ansible-runner from within
your awx directory and install it in-place.
Assuming you are doing an in-place install with it at `/testing/ansible-runner`
```
ansible-playbook -e new_package=file:///awx_devel/testing/ansible-runner runner_swap.yml
```
To install from a git branch, it needs to be non-editable.
```
ansible-playbook -e new_package=git+https://github.com/ansible/ansible-runner.git -e editable=false runner_swap.yml
```
This will allow you to make changes to ansible-runner and have them take
effect in multiple kinds of containers without re-installing.
#### Action - pre-populate image cache with EE
This allows you to distribute an image to all the nodes in the cluster.
Doing this will avoid a long delay before jobs start due to pulling the image.
Most often, you will want to use this to distribute the control plane EE
or the default EE for jobs / ad hoc commands / inventory updates.
Start by assuring that you have the EE locally (or that it's not old).
```
podman pull quay.io/ansible/awx-ee:latest
```
Check current images:
```
ansible-playbook image_distribute.yml --tags=list
```
Then run the playbook:
```
ansible-playbook image_distribute.yml
```
Now, if you run a job, it should start within a few seconds, without a minute
or more of delay.
If you delete the `.tar` file (image archive) then this will trigger re-exporting the image.
You need this if you need to pick up a change in the awx-ee image, but it's
more common that you recycle the dev environment containers.
Using a 3 node cluster and re-exporting the image, this playbook takes about 1.5 minutes.