added safe guards to all scripts that need it. all reference comm.sh like it should be

This commit is contained in:
2026-04-20 17:18:25 -04:00
parent 72940ec202
commit e83473ab8e
20 changed files with 263 additions and 32 deletions
+208 -32
View File
@@ -21,38 +21,12 @@
# v1.5 — Full icon set expanded — each operation and state has its own distinct icon
# All function output updated to use correct icon per context
# Icons grouped and commented by category for clarity
# v1.6 — ICON_CONTAINERS added — 📦 anchors all container sections for visual consistency
# ICON_NOT_RUNNING changed to ⭕ — distinct from ICON_STOPPED 🔴
# Section dividers updated from --- to ━━━ for cleaner log readability
# Summary passed/failed lines use ICON_SUCCESS and ICON_ERROR consistently
# v1.7 — ICON_MOVER added for mover operations
# ICON_CONTAINERS replaces ICON_DOCKER for docker/container operations
# validate_int added — reusable integer validation for any script
# v1.8 — ICON_PHP added for PHP-FPM operations
# v1.9 — ICON_REBOOT added for server reboot operations
# v2.0 — ICON_PLUGIN added for User Scripts plugin operations
# v2.1 — ICON_ZFS and ICON_MEM added for ZFS and memory diagnostics
# Diagnostics icon group added to icon block
# v2.2 — ICON_WATCHDOG added for Docker watchdog monitoring operations
# v2.3 — ICON_NOTIFY added for notification operations
# notify() added — shared notification function supporting unRAID native and Discord
# NOTIFY_UNRAID and DISCORD_WEBHOOK configured in Master.conf
# v2.4 — ICON_CLEAN, ICON_TRASH added for media cleaner operations
# ICON_PERMS, ICON_UNLOCKED added for media permissions operations
# ICON_REBOOT_SMART added for smart conditional reboot
# v2.5 — ICON_RAM added for ramdisk operations
# ICON_LINK added for symlink state and management
# Transcode scripts group added to ecosystem
# v2.6 — ICON_FAILOVER added for failover operations
# check_local_array added — verifies local /mnt/user is mounted and healthy
# check_remote_array added — verifies remote /mnt/user is mounted and healthy
# check_remote_docker added — verifies remote Docker daemon is responding
# ping_remote added — non-fatal ping returning status for failover use
# ping_internet added — non-fatal external ping for failover use
# v2.7 — ICON_WEBGUI added for WebGUI watchdog operations
# ICON_DOCKER_NET added for Docker network connect operations
# v2.8 — ICON_CERT added for SSL certificate monitoring operations
# v2.9 — ICON_MONITOR added for monitoring section headers
# v3.0 — Script locking system added — prevents concurrent execution conflicts
# acquire_lock() — create PID lock file, register EXIT trap
# release_lock() — remove lock file on exit
# acquire_rsync_lock() — per-profile lock + global concurrent limit
# release_rsync_lock() — decrement global counter, remove profile lock
# check_api() — pre-flight API reachability check
# ICON_SMART added for drive SMART health operations
# ICON_BANDWIDTH added for bandwidth tracking operations
# ICON_DIGEST added for health digest operations
@@ -579,6 +553,208 @@ get_rsync_opts() {
fi
}
# -----------------------------------------------------------------------------------------------
# SCRIPT LOCKING — v3.0
# Prevents multiple instances of the same script running simultaneously.
# All lock files live in /tmp/unraid_locks/ — auto-cleared on reboot.
#
# Usage in scripts:
# acquire_lock — strict: exit immediately if already running
# acquire_lock "wait" — wait mode: wait briefly then exit if still locked
# acquire_rsync_lock "$profile" — per-profile + global concurrent limit
#
# Stale lock detection — if lock file exists but PID is dead, clears and proceeds.
# Lock age warning — if lock is older than expected, warns but does not override.
# EXIT trap registered automatically — lock always released on exit, crash, or kill.
# -----------------------------------------------------------------------------------------------
LOCK_DIR="/tmp/unraid_locks"
RSYNC_COUNT_FILE="$LOCK_DIR/rsync_active_count"
RSYNC_MAX_CONCURRENT=3
LOCK_WARN_AGE=300 # seconds — warn if lock older than this (5min default)
LOCK_WAIT_TIMEOUT=30 # seconds — how long "wait" mode waits before giving up
# Internal — script name used as lock identifier
_lock_name() {
basename "${BASH_SOURCE[1]:-$0}" .sh
}
# Internal — lock file path for this script
_lock_file() {
echo "$LOCK_DIR/${1:-$(_lock_name)}.lock"
}
# Internal — release lock on exit
_release_on_exit() {
local lockfile="$1"
[[ -f "$lockfile" ]] && rm -f "$lockfile"
}
# -----------------------------------------------------------------------------------------------
# acquire_lock — acquire exclusive lock for this script
# Mode: strict (default) — exit immediately if locked
# wait — wait LOCK_WAIT_TIMEOUT seconds then exit
# Stale lock: if PID in lock file is dead → clear and acquire
# Age warning: if lock older than LOCK_WARN_AGE → warn
# -----------------------------------------------------------------------------------------------
acquire_lock() {
local mode="${1:-strict}"
local script_name
script_name=$(basename "${BASH_SOURCE[1]:-$0}" .sh)
local lockfile
lockfile="$(_lock_file "$script_name")"
mkdir -p "$LOCK_DIR"
# Check for existing lock
if [[ -f "$lockfile" ]]; then
local existing_pid
existing_pid=$(cat "$lockfile" 2>/dev/null)
# Stale lock detection — PID no longer running
if [[ -n "$existing_pid" ]] && ! kill -0 "$existing_pid" 2>/dev/null; then
warn "Stale lock detected for $script_name (PID $existing_pid gone) — clearing"
rm -f "$lockfile"
else
# Lock is active — check age
local lock_age
lock_age=$(( $(date +%s) - $(stat -c %Y "$lockfile" 2>/dev/null || echo 0) ))
if [[ "$lock_age" -gt "$LOCK_WARN_AGE" ]]; then
warn "$script_name has been running for ${lock_age}s — may be stuck (PID $existing_pid)"
fi
if [[ "$mode" == "wait" ]]; then
info "Another instance of $script_name is running — waiting up to ${LOCK_WAIT_TIMEOUT}s"
local waited=0
while [[ -f "$lockfile" ]] && [[ "$waited" -lt "$LOCK_WAIT_TIMEOUT" ]]; do
sleep 1
((waited++))
# Re-check for stale
existing_pid=$(cat "$lockfile" 2>/dev/null)
if [[ -n "$existing_pid" ]] && ! kill -0 "$existing_pid" 2>/dev/null; then
warn "Lock became stale while waiting — clearing"
rm -f "$lockfile"
break
fi
done
if [[ -f "$lockfile" ]]; then
error "$script_name still locked after ${LOCK_WAIT_TIMEOUT}s — exiting"
exit 1
fi
else
error "Another instance of $script_name is already running (PID $existing_pid) — exiting"
# Notify for critical scripts that should rarely overlap
case "$script_name" in
failover|transcode_management|media_management|daily_sync|system_watchdog)
notify "$script_name lock collision on $(hostname) — concurrent instance detected" "$script_name" "warning"
;;
esac
exit 1
fi
fi
fi
# Acquire lock
echo $$ > "$lockfile"
# Register EXIT trap to always release lock
trap "_release_on_exit '$lockfile'" EXIT
log "Lock acquired: $script_name (PID $$)"
}
# -----------------------------------------------------------------------------------------------
# acquire_rsync_lock — per-profile lock + global concurrent limit
# Prevents same profile running twice and limits total concurrent rsync instances
# -----------------------------------------------------------------------------------------------
acquire_rsync_lock() {
local profile="$1"
local profile_lock
profile_lock="$(_lock_file "rsync_${profile}")"
mkdir -p "$LOCK_DIR"
# Per-profile lock — same profile cannot run twice
if [[ -f "$profile_lock" ]]; then
local existing_pid
existing_pid=$(cat "$profile_lock" 2>/dev/null)
if [[ -n "$existing_pid" ]] && kill -0 "$existing_pid" 2>/dev/null; then
error "rsync profile '$profile' is already running (PID $existing_pid) — exiting"
exit 1
else
warn "Stale rsync lock for profile '$profile' — clearing"
rm -f "$profile_lock"
fi
fi
# Global concurrent limit
local current_count=0
if [[ -f "$RSYNC_COUNT_FILE" ]]; then
current_count=$(cat "$RSYNC_COUNT_FILE" 2>/dev/null || echo 0)
# Validate count — clean up if stale
local actual_count=0
for lf in "$LOCK_DIR"/rsync_*.lock; do
[[ -f "$lf" ]] || continue
local lpid
lpid=$(cat "$lf" 2>/dev/null)
kill -0 "$lpid" 2>/dev/null && ((actual_count++))
done
if [[ "$actual_count" -ne "$current_count" ]]; then
log "rsync count corrected: $current_count$actual_count"
current_count=$actual_count
echo "$current_count" > "$RSYNC_COUNT_FILE"
fi
fi
if [[ "$current_count" -ge "$RSYNC_MAX_CONCURRENT" ]]; then
error "Maximum concurrent rsync limit ($RSYNC_MAX_CONCURRENT) reached — exiting"
info "Active rsync locks: $(ls "$LOCK_DIR"/rsync_*.lock 2>/dev/null | xargs -I{} basename {} .lock | tr '\n' ' ')"
exit 1
fi
# Acquire profile lock and increment counter
echo $$ > "$profile_lock"
echo $(( current_count + 1 )) > "$RSYNC_COUNT_FILE"
# Register EXIT trap
trap "_release_rsync_on_exit '$profile_lock'" EXIT
log "rsync lock acquired: profile '$profile' (PID $$, active: $(( current_count + 1 ))/$RSYNC_MAX_CONCURRENT)"
}
# Internal — release rsync lock on exit
_release_rsync_on_exit() {
local profile_lock="$1"
[[ -f "$profile_lock" ]] && rm -f "$profile_lock"
# Decrement global counter
if [[ -f "$RSYNC_COUNT_FILE" ]]; then
local count
count=$(cat "$RSYNC_COUNT_FILE" 2>/dev/null || echo 1)
count=$(( count - 1 ))
[[ "$count" -lt 0 ]] && count=0
echo "$count" > "$RSYNC_COUNT_FILE"
fi
}
# -----------------------------------------------------------------------------------------------
# check_api — pre-flight API reachability check
# Verifies API endpoint is reachable before attempting operations
# Usage: check_api "http://localhost:8989" "Sonarr" || exit 1
# -----------------------------------------------------------------------------------------------
check_api() {
local url="$1"
local service="${2:-API}"
local timeout="${3:-10}"
if curl -sf --max-time "$timeout" "$url" >/dev/null 2>&1; then
log "$service API reachable: $url"
return 0
else
error "$service API not reachable: $url"
return 1
fi
}
# -----------------------------------------------------------------------------------------------
# STATUS DISPLAY
# Prints current runtime configuration — triggered by --status flag.