deploy: find cellular profile by id, no-auto-default NM conf, warn on NM enable failure

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KqhjP7nud5gdBg25zCiRk9

ap/nm-no-auto-default.conf

@@ -0,0 +1,5 @@
+# NetworkManager must never auto-create "Wired connection" DHCP profiles on ports
+# that netplan / ./deploy.sh --apply-network has not handed to it yet; otherwise it
+# races systemd-networkd's DHCP on the Starlink port.
+[main]
+no-auto-default=*

deploy.sh

@@ -13,9 +13,12 @@
 #         ./deploy.sh --wan-plan [conf]    # what the NetworkManager WAN provisioning would do (read-only, no root)
 #         ./deploy.sh render <file>        # preview a templated file on stdout
 #         ./deploy.sh --redact             # filter stdin, masking secret values (used by --diff; for tests)
-# Idempotent. Netplan (wlan0 + Starlink NIC = NM-managed WANs; eth0 is a LAN
+#         VANLINK_NM_DIR=<dir>             # test override: where existing NM *.nmconnection files are looked up (default /etc/NetworkManager/system-connections)
-# port on br0) is NOT deployed here — reference copy in ap/50-van-wan.yaml,
+# Idempotent. Netplan (the NM-managed WANs; eth0 is a LAN port on br0) is never
-# applied once manually (apply flaps uplinks). We do check it for drift below.
+# changed by a routine deploy (applying it flaps the uplinks). ENABLE_NM_WAN=0:
+# applied manually from the reference copy in ap/50-van-wan.yaml, and checked for
+# drift below. ENABLE_NM_WAN=1: a missing/differing WAN netplan is only warned
+# about; it is applied by `./deploy.sh --apply-network` alone.
 set -euo pipefail
 cd "$(dirname "$(readlink -f "$0")")"
 
@@ -98,6 +101,19 @@ nm_wan_state_netplan() {  # missing | differs | matches (the installed WAN netpl
     else echo differs; fi
 }
 
+cellular_profile_exists() {  # true if an NM profile with id=$CELLULAR_CONN_NAME exists (any file name, or known to nmcli)
+    local d=${VANLINK_NM_DIR:-/etc/NetworkManager/system-connections} f names
+    for f in "$d"/*.nmconnection; do
+        [ -f "$f" ] || continue
+        if grep -Fx "id=${CELLULAR_CONN_NAME}" "$f" >/dev/null 2>&1; then return 0; fi
+    done
+    if command -v nmcli >/dev/null 2>&1; then
+        names=$(nmcli -g NAME connection show 2>/dev/null) || names=
+        if printf '%s\n' "$names" | grep -Fx "${CELLULAR_CONN_NAME}" >/dev/null 2>&1; then return 0; fi
+    fi
+    return 1
+}
+
 check_rendered() {  # check_rendered <src> <rendered-file> — fail on a leftover @TOKEN@
     local left
     left=$(grep -oE "$TOKEN_RE" "$2" | sort -u | tr '\n' ' ') || true
@@ -315,7 +331,7 @@ case "${1:-}" in
             if [ -f /etc/netplan/10-dhcp-all-interfaces.yaml ]; then echo "  armbian: /etc/netplan/10-dhcp-all-interfaces.yaml present -> would move aside"; else echo "  armbian: /etc/netplan/10-dhcp-all-interfaces.yaml absent"; fi
             if [ -z "${CELLULAR_APN:-}" ]; then
                 echo "  cellular profile: skipped (no CELLULAR_APN)"
-            elif [ -e "/etc/NetworkManager/system-connections/${CELLULAR_CONN_NAME}.nmconnection" ]; then
+            elif cellular_profile_exists; then
                 echo "  cellular profile: ${CELLULAR_CONN_NAME} apn=${CELLULAR_APN} exists"
             else
                 echo "  cellular profile: ${CELLULAR_CONN_NAME} apn=${CELLULAR_APN} would create"
@@ -643,6 +659,8 @@ install -D -m0644 power/10-vanlink-watchdog.conf /etc/systemd/system.conf.d/10-v
 
 if [ "$ENABLE_NM_WAN" = 1 ]; then
     echo "== networkmanager WANs =="
+    # In place before the package starts NM: no auto-created "Wired connection" profiles.
+    install -D -m0644 ap/nm-no-auto-default.conf /etc/NetworkManager/conf.d/00-van-no-auto-default.conf
     if ! dpkg -s network-manager >/dev/null 2>&1; then
         # Installed only now: the package starts NM, which manages Wi-Fi/gsm devices by
         # default, so van-ap-unmanaged.conf (AP radios, LAN ports) must already be in place.
@@ -650,7 +668,8 @@ if [ "$ENABLE_NM_WAN" = 1 ]; then
         DEBIAN_FRONTEND=noninteractive apt-get install -y network-manager \
             || { echo "deploy: apt-get install network-manager failed" >&2; exit 1; }
     fi
-    systemctl enable --now NetworkManager ModemManager >/dev/null 2>&1 || true
+    systemctl enable --now NetworkManager ModemManager >/dev/null 2>&1 \
+        || warn "failed to enable/start NetworkManager/ModemManager (check: systemctl status NetworkManager ModemManager)"
     nmcli general reload conf 2>/dev/null || true
     for nm_if in "$WIFI_5G_IFACE" "$WIFI_2G_IFACE" "$LAN_IFACE" ${ONBOARD_LAN_IFACE:-}; do
         [ -e "/sys/class/net/$nm_if" ] || continue
@@ -660,7 +679,7 @@ if [ "$ENABLE_NM_WAN" = 1 ]; then
     done
     if [ -n "${CELLULAR_APN:-}" ]; then
         nm_keyfile="/etc/NetworkManager/system-connections/${CELLULAR_CONN_NAME}.nmconnection"
-        if [ -e "$nm_keyfile" ]; then
+        if cellular_profile_exists; then
             echo "   cellular connection ${CELLULAR_CONN_NAME} exists — left unchanged"
         elif [ -z "${CELLULAR_UUID:-}" ]; then
             warn "CELLULAR_UUID is empty — not creating the cellular keyfile $nm_keyfile"
@@ -670,8 +689,9 @@ if [ "$ENABLE_NM_WAN" = 1 ]; then
             echo "   created cellular connection ${CELLULAR_CONN_NAME} (apn ${CELLULAR_APN}) — no modem needed"
         fi
     fi
-    if [ "$(nm_wan_state_netplan)" != matches ]; then
+    nm_np_state=$(nm_wan_state_netplan)
-        warn "WAN netplan is not provisioned/up to date: run ./deploy.sh --apply-network (backs up /etc/netplan and rolls back automatically if the uplink does not come back)"
+    if [ "$nm_np_state" != matches ]; then
+        warn "WAN netplan is $nm_np_state (not provisioned/up to date): run ./deploy.sh --apply-network (backs up /etc/netplan and rolls back automatically if the uplink does not come back)"
     fi
 fi
 

tests/check-templates.sh

@@ -242,4 +242,19 @@ u2=$(VANLINK_CONF=hosts/wan.conf ./deploy.sh render ap/cellular.nmconnection | g
 [ -n "$u1" ] && [ "$u1" != "uuid=" ] && [ "$u1" = "$u2" ] && ok "cellular uuid non-empty and stable across hosts" || bad "cellular uuid: '$u1' vs '$u2'"
 [ "$(grep -v '^[[:space:]]*#' deploy.sh | grep -c uuidgen)" = 0 ] && ok "deploy.sh does not call uuidgen" || bad "deploy.sh calls uuidgen"
 
+# 17b. an existing cellular profile is found by id, whatever its file name
+nmd=$(mktemp -d); nme=$(mktemp -d)
+printf '[connection]\nid=Koodo\ntype=gsm\n' > "$nmd/weird-file-name.nmconnection"
+p=$(VANLINK_NM_DIR=$nmd VANLINK_CONF=hosts/van.conf ./deploy.sh --wan-plan 2>&1)
+echo "$p" | grep -q 'cellular profile: Koodo apn=sp.koodo.com exists' && ok "wan-plan: profile found under another file name" || bad "wan-plan: profile under other file name not found"
+if command -v nmcli >/dev/null 2>&1 && nmcli -g NAME connection show 2>/dev/null | grep -Fxq Koodo; then
+    ok "wan-plan would-create assertion skipped (this machine's NM lists Koodo)"
+else
+    p=$(VANLINK_NM_DIR=$nme VANLINK_CONF=hosts/van.conf ./deploy.sh --wan-plan 2>&1)
+    echo "$p" | grep -q 'cellular profile: Koodo apn=sp.koodo.com would create' && ok "wan-plan: empty NM dir -> would create" || bad "wan-plan: empty NM dir should say would create"
+fi
+rm -rf "$nmd" "$nme"
+# 17c. NM must not auto-create "Wired connection" profiles
+[ "$(grep -v '^[[:space:]]*#' ap/nm-no-auto-default.conf 2>/dev/null | grep -v '^[[:space:]]*$')" = "$(printf '[main]\nno-auto-default=*')" ] && ok "nm-no-auto-default.conf is [main] + no-auto-default=*" || bad "ap/nm-no-auto-default.conf missing or wrong"
+
 exit "$fail"