Finding the repository to backup into causes staging directory population

What is the problem?
To find the repo to backup into the backup script calls cbbackupmgr info (1, 2). When backing up to a PVC or to the cloud with a persistent staging directory this is fast, if a little indirect. However, when using ephemeral staging this causes a staging directory population for the entire archive, which can take a long time. The backup itself only needs to repo to be staged, which should be much quicker.

What is the solution?
The repo to backup should be persisted somehow, or calculated in a more efficient way