I am currently trying to deploy a maas instance in our lab environment with maas-ansible-playbook. The maas-servers in question are running Ubuntu 22.04. I am using deb installation and have tried with both version 3.3 and 3.4, the deployment works fine but after adding DHCP to one or multiple VLANs the maas-dhcpd service does not start.
This has been observed both here on the forums and been raised as an issue in the github repo linked above. I wrote a comment on that issue with more details
How can we solve this within the playbooks available? I am currently deploying a new instance to see if the case is the same using snap installation, however i want to use deb in our lab since that is what we use in production today.
Deployed with ansible again but using snap instead of deb, now the services is active on both my rack controllers.
Thank you for posting. Unfortunately,
maas-ansible-playbook has been deprecated and it is not actively maintained anymore.
maas-terraform-provider is the solution that the MAAS team is providing as infrastructure-as-code.
Hi and thank you for your reply!
Uh ok, from my point of view the maas-ansible-playbook and maas-terraform-provider has two completely different use cases.
The maas-ansible-playbooks can be used to deploy a new maas instance from scratch. While the maas-terraform-provider requires you to have an up and running and functional maas to even use it (it communicates with the maas API). So how can it be replaced when the use cases are so different?
Correct me if im wrong but the terraform provider is used to manage resources whitin maas and the ansible-playbooks is used to deploy maas itself?
Really sorry for my incomplete reply and confusion. You are right, Terraform provider is used to manage resources, while the ansible playbooks are used to deploy maas. The alternative for the deployment would be ansible-collection that creates an Ansible plugin.
Going back to the issue, it can happen because the rack controller is having problems creating some DHCP files used during the configuration.
Could you please provide some logs (maybe through journalctl) when running the installation using deb? If we see some errors related to a
dhcpd.conf files while running the deb installation, we can narrow down the problem.
Jumping in to the discussion here just to clarify:
The Terraform-provider and Ansible-collection both require a pre-running MAAS instance to communicate with.
We don’t currently have a product that can stand up a MAAS instance with the deprecation of the playbooks.
You’ve mentioned the snap was successful, is it possible to use the MAAS snap alongisde your other deb instances? I know that’s not ideal.
Failing that, continuing with Javier’s troubleshooting might help!
Thanks for clarifying!
I will do some more testing with deb and check the installation logs to see if i can catch any errors. If solvable with me i will add them to git aswell
Since this specific situation is about a lab env i can run with a snap installation there ofc. Very bad timing i guess since i was hoping to use those ansible-playbooks for deployment of new maas instances in production
And for what it’s worth i think there is a lot of value in providing users with some type of automated deployment method for multi-node setups sorry to hear about the deprecation