Simple checkpoint backup

[zalberico]

Replaying due to a corrupted checkpoint can take on the order of weeks (and is currently quite difficult because of how the replay code functions which we're trying to fix in: https://github.com/urbit/vere/issues/157)

It'd save a lot of misery if a backup checkpoint file was created - at least once a week? We can work around this in hosting with https://github.com/tloncorp/ylem/issues/784 but it seems like we should be able to do something to make this better for everyone else (even though the real fix is live snapshotting).

[zalberico]

I'd even take just created on upgrade or having a runtime command for it like next, maybe save (that runs on upgrade)? Have it save the last 3 in a checkpoint backups folder or something. This with a functional replay will fix most of the misery with restoring planets that have corrupted checkpoints.