As part of disagg, we're introducing a new recovery settings called precise checkpoint. If this setting is enabled, we should not run RTS on recovery and vice versa. However when recovery, there's currently no way to know if the backup is saved with precise checkpoint or not.

To resolve this, we need to encode what type of checkpoint into metadata. Then as part of recovery, decide to Rollback to stable (if precise_checkpoint is false) or recovery with precise checkpoint.