Fix MLM 5.1 payg AWS deployment issues.

[maximenoel8]

What does this PR change?

Fix AWS MLM 5.1 deployment.

• server -> setup fix
• sles12sp5 -> test repo issue with GPG
• rhel9 -> need to be registered

Changes

• update AMI image selections
• bump minima and jdupe version
• update minimal requirements for testsuite repositories in minima
• exclude controller from client tools repositories
• update combustion to support ssh configuration in sshd configuration file and directory
• add variable to select the correct sshd configuration in main.tf
• move SCC init.sls in salt stages to declare the repositories at the right time (needed for some packages)
• Fix gpg key issue with SLES by forcince the gpg import key
• Change partition for secondary disk

[NamelessOne91]

Thanks Maxime

A couple of comments in the PR and another question: do we need to actually touch the server > additional_disks.sls Salt states?
I'd expect 5.1 to use server_containerized > additional_disks.sls which relies on the provided tools to setup volumes for the containerized server.

[NamelessOne91]

Just care that the sintax for the YAML configuration files consumed by Minima has changed in some places.
It may work for testsuite repositories but break elsewhere.

[maximenoel8]

It works fine with the required yaml.

[NamelessOne91]

I am not so much worried about this yaml, but others yaml files (sometimes inline in Salt states) that we use.

I remember trying to update minima some time ago, and there were a few other things to adapt
See #1832

[maximenoel8]

Thanks Maxime

A couple of comments in the PR and another question: do we need to actually touch the server > additional_disks.sls Salt states? I'd expect 5.1 to use server_containerized > additional_disks.sls which relies on the provided tools to setup volumes for the containerized server.

For the tools to work, you need to specify the disk you want to add. To be honest this one is tricky. I have the feeling the main disk and additional disk name are switching from time to time because of a race conditions. If the issue persist, we will need a smart code to select the unused disk.

[NamelessOne91]

To be honest this one is tricky. I have the feeling the main disk and additional disk name are switching from time to time because of a race conditions.

Yes, it's rare, but I have seen that happen.
Anyway, I was wondering if changing additional_disks.sls for server does anything for 5.1, as I think it will use the one for server_containerized.