Shutdown, Takeaway and restore

[rustyrussell]

1. A shutdown-bundle (?) command which takes a filename, zip/tar/etc everything up (configs, hsm_secret, database, certificates, etc) after stopping but before exiting. Maybe moves those files out the way, so we can't accidentally start again?
2. A --restore-from (?) parameter to start a blank node by unpacking those files.
3. Plugins should be able to add filenames to include in the bundle (e.g. bookeeper's db, clnrest certificates)

[whitslack]

Doing this cleanly will require implementing #4842 because that's the only way a node can safely go offline. If there are any outstanding HTLCs in any of the node's channels, then it's not safe for the node to go offline for any amount of time — and the longer the node stays offline, the greater the probability that an HTLC will hit its deadline while the node is offline, and the peer will be forced to unilaterally close.