Hey, I'm experiencing an issue with postMountHook. I have a zfs pool that looks roughly like this:

{
  type = "zpool";

  options = {
    ashift = "12";
  };

  rootFsOptions = {
    acltype = "posix";
    compression = "on";
    mountpoint = "none";
    overlay = "off";
    xattr = "sa";
  };

  datasets =
    let
      mkDataset = lib.attrsets.recursiveUpdate {
        type = "zfs_fs";
      };
    in
    {
      "rootfs" = mkDataset {
        mountpoint = "/";
      };

      "home" = mkDataset {
        mountpoint = "/home";
      };

      "home/user" = mkDataset {
        name = "home/${config.custom.username}";
        mountpoint = "/home/${config.custom.username}";

        options = {
          mountpoint = "/home/${config.custom.username}";
          encryption = "on";
          keyformat = "passphrase";
          keylocation = "prompt";
          "com.sun:auto-snapshot" = "true";
        };

        postMountHook = ''
          # ...
        '';
      };
    };
}

(other datasets omitted for the sake of brevity)

Initially I did not specify options.mountpoint, and ran into the problem that my dataset was not mounted by the time postMountHook got executed. I saw the zfs example in the disko repo using both mountpoint and options.mountpoint, so I figured I'd give it a shot. Now the problem is that it remains mounted when the main disko script tries to mount the rest, which then fails because /mnt is not empty... The error message:

+ findmnt nixos/rootfs /mnt/
+ mount nixos/rootfs /mnt/ -o X-mount.mkdir -o defaults -o zfsutil -t zfs
zfs_mount_at() failed: directory is not empty+ rm -rf /tmp/tmp.aSJMo4rNbE

I don't know if I'm doing something wrong or if the script generation is just broken?

Hey, I'm experiencing an issue with postMountHook. I have a zfs pool that looks roughly like this:

{
  type = "zpool";

  options = {
    ashift = "12";
  };

  rootFsOptions = {
    acltype = "posix";
    compression = "on";
    mountpoint = "none";
    overlay = "off";
    xattr = "sa";
  };

  datasets =
    let
      mkDataset = lib.attrsets.recursiveUpdate {
        type = "zfs_fs";
      };
    in
    {
      "rootfs" = mkDataset {
        mountpoint = "/";
      };

      "home" = mkDataset {
        mountpoint = "/home";
      };

      "home/${config.custom.username}" = mkDataset {
        mountpoint = "/home/${config.custom.username}";

        options = {
          mountpoint = "/home/${config.custom.username}";
          encryption = "on";
          keyformat = "passphrase";
          keylocation = "prompt";
          "com.sun:auto-snapshot" = "true";
        };

        postMountHook = ''
          # ...
        '';
      };
    };
}

(other datasets omitted for the sake of brevity)

Initially I did not specify options.mountpoint, and ran into the problem that my dataset was not mounted by the time postMountHook got executed. I saw the zfs example in the disko repo using both mountpoint and options.mountpoint, so I figured I'd give it a shot. Now the problem is that it remains mounted when the main disko script tries to mount the rest, which then fails because /mnt is not empty... The error message:

+ findmnt nixos/rootfs /mnt/
+ mount nixos/rootfs /mnt/ -o X-mount.mkdir -o defaults -o zfsutil -t zfs
zfs_mount_at() failed: directory is not empty+ rm -rf /tmp/tmp.aSJMo4rNbE

I don't know if I'm doing something wrong or if the script generation is just broken?