Server is crippled. Rip and replace or build new and migrate?

I’ve given up hope on repairing my current homelab install. The pulp database appears to be more or less corrupted which is crippling the server’s functionality.

If i were to power the VM down and spin up a direct replacement with the same hostname, aside from
the server’s SSH keypairs are there components I could selectively backup from the current server and import on the new one that would allow puppet and subscription-manager clients to automatically communicate with the new host, populating themselves in the inventory?

OR would it be better to stand up an additional VM and then reprovision the agent configs on the clients so they are managed by the new server/hostname?

I’m beginning to think the hosed install might have been associated with Epel being present on the EL8 setup as there were some weird errors related to a couple of the pulp core packages. I’m also in the process of rebuilding using 3.1 and 4.3 instead of the nightly now that EL8 support is official.

Hi @gawainxl,

I would highly recommend continuing down the path of using a non-nightly version of Katello. Nightly is highly unstable for anything other than testing.

I can’t speak for backing up the puppet side of things, but I don’t see a way to back up your registered content host DB data over to the new Katello install. If it was a backup between the same versions of Katello then maybe, but not from nightly. You’ll likely save time by starting from scratch and re-registering your hosts.

1 Like