I started feeling better about it after considering that system install for upgrade is my only test for recovery from drive loss under non-emergency conditions. Two root partitions; install to the spare, them populate it from backups (admittedly very fresh). If anything is missing, copy it on from old root, figure out why backups omitted it, and fix the scripts. Non-essential data e.g. read-only git tree clones goes on a third, large "cache" partition.