skip lost+found during the ownership sweep

docker/main/rootfs/usr/local/bin/fix-ownership

@@ -10,6 +10,8 @@
 #                on every boot)
 #
 # Only files whose uid OR gid differs are touched, so re-runs are cheap.
+# lost+found is skipped: it belongs to the filesystem, not to Frigate, and
+# fsck puts recovered fragments of arbitrary files there under root-only 0700.
 # Top-level /config additionally grants group frigate-data TRAVERSE ONLY
 # (g+rx) so the separate go2rtc user can reach its pre-created HomeKit file
 # on hosts where /config is mounted 0700. Never g+w: directory write means
@@ -82,7 +84,7 @@ for path in "$@"; do
     # stale mount. Tolerate it rather than aborting under errexit, but never
     # read a failed scan as "nothing to do": that would record the sweep as
     # complete without having looked.
-    if ! count=$(find "$path" \( -not -uid "$target_uid" -o -not -gid "$target_gid" \) -printf '.' 2>/dev/null | wc -c); then
+    if ! count=$(find "$path" -name lost+found -prune -o \( -not -uid "$target_uid" -o -not -gid "$target_gid" \) -printf '.' 2>/dev/null | wc -c); then
         swept_clean=0
         echo "[WARN] fix-ownership: could not scan ${path}; will retry on next boot"
         continue
@@ -109,7 +111,7 @@ for path in "$@"; do
     # -print feeds the progress counter while -exec {} + keeps the chown
     # batched; the scan above is what makes a real percentage possible
     started=$SECONDS
-    if find "$path" \( -not -uid "$target_uid" -o -not -gid "$target_gid" \) \
+    if find "$path" -name lost+found -prune -o \( -not -uid "$target_uid" -o -not -gid "$target_gid" \) \
         -print -exec chown -h "${target_uid}:${target_gid}" {} + \
         | awk -v total="$count" -v path="$path" '
             BEGIN { next_pct = 5 }

docs/docs/configuration/non_root.md

@@ -56,6 +56,10 @@ Pass `PUID` and `PGID` as the third and fourth arguments if you're not using the
 
 Once the volumes are aligned, start Frigate normally. A sentinel at `/config/.permissions_version` records what was done, so later boots skip the sweep entirely unless you change `PUID`/`PGID`.
 
+`lost+found` is left alone. It belongs to the filesystem rather than to Frigate, and `fsck` recovers fragments of arbitrary files into it under root-only permissions, so handing it to the runtime user would expose whatever ends up there. Expect to see it still owned by root afterward, on any volume that's a dedicated mount.
+
+If something under your volumes genuinely can't be chowned, a read-only btrfs snapshot directory for example, the sweep warns and names the path, and it deliberately doesn't write the sentinel. That means it retries on the next boot rather than recording a migration that didn't finish. Either move those paths outside `/media/frigate` or expect the scan to repeat.
+
 ### Network storage
 
 Storing recordings on a NAS is common, and ownership behaves differently there. Check what you have before migrating anything: