These scripts are developed by/for the Fedora Copr maintainers, and are made public mostly for educational purposes. Unless you are one of the Fedora Copr team members, you probably don't want to run those (at least not without in-place modifications).
Prepare the ~/.aws/credentials
file. It needs to contain the following:
[fedora-copr]
aws_access_key_id=XXXXXXXXXXXXXXXXXXXX
aws_secret_access_key=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Prepare the ~/.aws/config
file. It needs to contain the following profile:
[profile fedora-copr]
region = us-east-1
First, it is really important to get familiar with the Fedora Copr docs related to instance upgrades.
These playbooks require you to specify copr_instance
and server_id
variables
explicitly. For example:
$ ansible-playbook <playbook> -e copr_instance=dev -e server_id=keygen
Make sure you understand the structure of host_vars/{{ copr_instance }}.yml
files before you execute it.
play-vm-migration-01-new-box.yml
— starts a new virtual machine, named with the-new
suffix.play-vm-migration-02-migrate-backend-box.yml
— properly evacuates the old Copr Backend instance (stops VGs and RAIDs, unmounts the storage), attaches the data volumes to the new backend instance, moves v4/v6 IPs.play-vm-migration-02-migrate-non-backend-box.yml
— performs similar task to the above playbook, but for non-backend instances (frontend, keygen, distgit, pulp)play-vm-migration-03-rename-instances.yml
— drops the-new
suffix from the new instances'Name
tags, and ads-old
suffix to the old instances.
-
play-backend-snapshot.yml
- create a set of snapshots for all the data volumes currently attached to copr-backend VM -
play-backend-snapshot-recovery-example.yml
- creates a testing VM and attaches a freshly created volumes-from-BE-snapshots to it.
ansible-galaxy collection install amazon.aws