Wolf480pl on Nostr: LisPi Normal Mikoto :kagami_shrug: :gekota: akko ok, so what you're saying is if a ...

LisPi (nprofile…2mzm) Normal Mikoto :kagami_shrug: :gekota: (nprofile…nc0q) akko (nprofile…wssg) ok, so what you're saying is if a process does one or more overwrite in-place of part of a file, and either does not fsync, or the power outage / kernel panic happens before it can do fsync, then an arvitrary subset of those writea will have happened, and some of those may've only happened partially?

I think that's something Postgres and other DBs are prepared to live with.