feat: add CUPS watchdog timers

2026-06-09 14:58:13 +08:00
parent 2a688d2514
commit 00d512f9d9
11 changed files with 668 additions and 0 deletions
--- a/watchdog/print-watchdog.sh
+++ b/watchdog/print-watchdog.sh
@@ -0,0 +1,162 @@
+#!/bin/bash
+#
+# Print service watchdog for the CUPS print server.
+# It checks systemd service state plus quick CUPS command/HTTP probes.
+# When checks fail repeatedly, it restarts CUPS and related services.
+#
+
+set -u
+
+CONFIG_FILE="${CUPS_PRINT_WATCHDOG_CONFIG:-/etc/cups-watchdog/print-watchdog.conf}"
+STATE_DIR="/run/cups-watchdog"
+
+SERVICES="cups avahi-daemon cups-driver-manager"
+CUPS_URL="http://127.0.0.1:631/"
+CHECK_CUPS_HTTP=1
+CHECK_LPSTAT=1
+COMMAND_TIMEOUT=8
+FAIL_THRESHOLD=2
+RESTART_COOLDOWN=60
+LOG_FILE="/var/log/cups-watchdog/print.log"
+
+if [ -f "$CONFIG_FILE" ]; then
+    # shellcheck disable=SC1090
+    . "$CONFIG_FILE"
+fi
+
+mkdir -p "$STATE_DIR"
+mkdir -p "$(dirname "$LOG_FILE")" 2>/dev/null || true
+
+log_msg() {
+    local msg="$1"
+    local line
+    line="$(date '+%Y-%m-%d %H:%M:%S') [print-watchdog] $msg"
+    echo "$line"
+    echo "$line" >> "$LOG_FILE" 2>/dev/null || true
+}
+
+unit_exists() {
+    local service="$1"
+
+    systemctl list-unit-files --no-legend "$service.service" 2>/dev/null | awk '{print $1}' | grep -qx "$service.service" && return 0
+    systemctl list-units --all --no-legend "$service.service" 2>/dev/null | awk '{print $1}' | grep -qx "$service.service"
+}
+
+service_should_check() {
+    local service="$1"
+
+    unit_exists "$service" || return 1
+    if systemctl is-enabled --quiet "$service.service" 2>/dev/null; then
+        return 0
+    fi
+    if systemctl is-active --quiet "$service.service" 2>/dev/null; then
+        return 0
+    fi
+
+    return 1
+}
+
+check_systemd_services() {
+    local service
+    local failed=0
+
+    for service in $SERVICES; do
+        service_should_check "$service" || continue
+        if ! systemctl is-active --quiet "$service.service" 2>/dev/null; then
+            log_msg "service not active: $service"
+            failed=1
+        fi
+    done
+
+    return "$failed"
+}
+
+check_lpstat() {
+    [ "$CHECK_LPSTAT" = "1" ] || return 0
+    command -v lpstat >/dev/null 2>&1 || return 0
+
+    timeout "$COMMAND_TIMEOUT" lpstat -r >/dev/null 2>&1
+}
+
+check_http() {
+    [ "$CHECK_CUPS_HTTP" = "1" ] || return 0
+    command -v curl >/dev/null 2>&1 || return 0
+
+    curl -fsS --max-time "$COMMAND_TIMEOUT" "$CUPS_URL" >/dev/null 2>&1
+}
+
+restart_print_stack() {
+    local now
+    local last_file="$STATE_DIR/print.last_restart"
+    local last=0
+    local service
+
+    now="$(date +%s)"
+    if [ -f "$last_file" ]; then
+        last="$(cat "$last_file" 2>/dev/null || echo 0)"
+    fi
+
+    if [ $((now - last)) -lt "$RESTART_COOLDOWN" ]; then
+        log_msg "restart skipped: cooldown active"
+        return 0
+    fi
+
+    log_msg "restarting print stack"
+    systemctl reset-failed cups.service >/dev/null 2>&1 || true
+    systemctl restart cups.service
+
+    for service in avahi-daemon cups-driver-manager; do
+        service_should_check "$service" || continue
+        systemctl reset-failed "$service.service" >/dev/null 2>&1 || true
+        systemctl restart "$service.service" >/dev/null 2>&1 || true
+    done
+
+    echo "$now" > "$last_file"
+}
+
+check_print_stack() {
+    local count_file="$STATE_DIR/print.fail_count"
+    local fail_count=0
+    local failed=0
+
+    check_systemd_services || failed=1
+    if ! check_lpstat; then
+        log_msg "lpstat check failed or timed out"
+        failed=1
+    fi
+    if ! check_http; then
+        log_msg "CUPS HTTP check failed or timed out"
+        failed=1
+    fi
+
+    if [ "$failed" = "0" ]; then
+        echo 0 > "$count_file"
+        log_msg "print stack ok"
+        return 0
+    fi
+
+    if [ -f "$count_file" ]; then
+        fail_count="$(cat "$count_file" 2>/dev/null || echo 0)"
+    fi
+    fail_count=$((fail_count + 1))
+    echo "$fail_count" > "$count_file"
+
+    log_msg "print stack failed ($fail_count/$FAIL_THRESHOLD)"
+    if [ "$fail_count" -ge "$FAIL_THRESHOLD" ]; then
+        restart_print_stack
+        echo 0 > "$count_file"
+    fi
+}
+
+case "${1:-check}" in
+    check)
+        check_print_stack
+        ;;
+    restart)
+        restart_print_stack
+        ;;
+    *)
+        echo "Usage: $0 [check|restart]"
+        exit 2
+        ;;
+esac