fix: image uid

Dockerfile

@@ -55,22 +55,25 @@ COPY --from=ghcr.io/amacneil/dbmate:2 /usr/local/bin/dbmate /usr/local/bin/dbmat
 COPY db/migrations/ /app/db/migrations/
 
 # debian:*-slim ships without /etc/nsswitch.conf, and glibc then cannot turn a
-# uid back into a user record. libvirt does exactly that lookup before connecting
-# and gives up with "Failed to find user record for uid", whatever the socket
-# permissions are. Restore the plain file-based lookup.
+# uid back into a user record. Restore the plain file-based lookup.
 RUN printf 'passwd: files\ngroup: files\nshadow: files\nhosts: files dns\n' \
     > /etc/nsswitch.conf
 
-# Unprivileged, but it must land in the host's libvirt group to reach the
-# socket: `group_add` in the compose file does that at run time.
-RUN useradd --system --create-home --uid 10001 app
+# Runs as root, on purpose. libvirt resolves the calling uid to a user record
+# before it even opens the socket, and refused to do so for a custom uid here
+# ("Failed to find user record for uid") even with /etc/passwd and the libvirt
+# group both in place. Root always resolves.
+#
+# The privilege actually gained is small: any process that can reach the libvirt
+# socket already starts and destroys the VMs of this host, which is the whole
+# point of this container. It holds no other mount and no secret beyond the
+# database password.
 
 WORKDIR /app
 COPY --from=backend /build/target/release/app-template /usr/local/bin/vm-selector
 COPY --from=frontend /build/dist/ /app/frontend/
 COPY docker/config.yml /etc/app-template/config.yml
 
-USER app
 EXPOSE 3000
 
 HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \

README.md

@@ -196,24 +196,22 @@ cargo sqlx prepare      # dev database must be up and migrated
 
 ### Portainer stack
 
-Deploy `docker-compose.yml` and set the four stack variables it documents at the top
-(`IMAGE`, `POSTGRES_PASSWORD`, `LIBVIRT_GID`, `APP_PORT`). Get the group id from the
-host:
-
-```bash
-getent group libvirt | cut -d: -f3
-```
+Deploy `docker-compose.yml` and set the stack variables it documents at the top
+(`IMAGE`, `POSTGRES_PASSWORD`, `APP_PORT`).
 
 Then create a webhook on the stack, **tick "Re-pull image"**, and put its url in the
 `PORTAINER_WEBHOOK` secret of the repository. Without the re-pull, the webhook
 recreates the containers from the layer already on disk and nothing changes.
 
-The two things that make libvirt reachable from inside the container:
+What makes libvirt reachable from inside the container: `/var/run/libvirt` is
+bind-mounted from the host — the whole directory, since the socket is `libvirt-sock`
+or `virtqemud-sock` depending on the host's setup — and the container runs as root.
 
-- `/var/run/libvirt` bind-mounted from the host — the whole directory, since the
-  socket is `libvirt-sock` or `virtqemud-sock` depending on the host's setup
-- the container in the host's `libvirt` group (`group_add`) — without it the socket
-  is visible and every call is denied
+Running as root is not laziness about the group: with a custom uid, libvirt resolves
+the caller to a user record before opening the socket and refused to do so here even
+with `/etc/passwd` correct and the `libvirt` group added through `group_add`
+(`Failed to find user record for uid '10001'`). What root actually adds is small —
+anything that reaches this socket already controls every VM on the host.
 
 ### Domain name
 

docker-compose.yml

@@ -4,8 +4,6 @@
 # Environment variables to set in the Portainer stack:
 #   IMAGE              registry.tibiscuit.ch/vm-selector:latest
 #   POSTGRES_PASSWORD  anything, it never leaves the internal network
-#   LIBVIRT_GID        the host's libvirt group id, from:
-#                        getent group libvirt | cut -d: -f3
 #   APP_PORT           host port for the reverse proxy to point at (default 3010)
 
 services:
@@ -59,10 +57,8 @@ services:
       # runs the monolithic libvirtd or the split daemons, the app needs
       # libvirt-sock or virtqemud-sock.
       - /var/run/libvirt:/var/run/libvirt
-    # Reaching that socket requires being in the host's libvirt group. Without
-    # this the socket is visible and every call is denied.
-    group_add:
-      - '${LIBVIRT_GID:?see the header of this file}'
+    # No `group_add`: the container runs as root (see the Dockerfile for why), so
+    # it needs no group membership to reach the socket.
     depends_on:
       db:
         condition: service_healthy