I'm not sure the backups you took with the Central API/CLI when you set up the tunnel were good backups. My guess is that they are corrupt due to no disk space. If I had this problem, I would:
- Make a written plan and have my teammates (and maybe ChatGPT?) critique it before proceeding. I'd also get https://asciinema.org installed so I have a record of all actions taken to go along with notes I took during the process.
- Use rsync to move the docker install from the source machine to a larger destination machine with the same OS and docker version. If rsync doesn't work, use the tunnel method to run pg_dump on the source database and restore it with pg_restore using the scripts I shared earlier.
- Confirm everything works after rsync move, then upgrade Central on the destination machine, version by version. Make sure you can login to every version before continuing. Use snapshots so you can easily rollback if something fails.