MacOS has some interesting tricks for this. First, UID/GID 99 is special — it always matches the current id.
Second, on mount, is the "noowners" option, which causes all objects to seem to have uid/gid 99. Of course, this was before ACLs, which … complicated things. Sean. > On May 2, 2020, at 7:36 AM, allanj...@freebsd.org wrote: > > I think this idea has been discussed before, but solution that jumps to my > mind is a set of filesystem property (zfs set) to override the > ownership/group/etc of all files in the dataset. So in this case you might > have something like: > > zfs set override_user=sarvi mypool/build/cloned > (or zfs clone -o override_user=sarvi mypool/build@snap mypool/build/cloned) > > It might be more useful to try to support something more like a mapping, but > I don't know if it is worth the complexity: > > zfs set override_users=buildusr:sarvi,nobody:operator mypool/build/cloned > > openzfs / openzfs-developer / see discussions + participants + delivery > options Permalink ------------------------------------------ openzfs: openzfs-developer Permalink: https://openzfs.topicbox.com/groups/developer/T82313bf565b8e09c-Mfd7b3bab7394c103f73aadb4 Delivery options: https://openzfs.topicbox.com/groups/developer/subscription
