TrueNAS Scale scripts to (a) handle waking a backup server from sleep, (b) shut down the backup server after a grace period has passed AND all replication / scrub activity is complete. `Install wake_backup_server_for_replication.sh` as a cron job on the main server, triggered once 30m before your replication task is scheduled. Install `shutdown_…

shutdown_after_replication.sh

#!/bin/bash

# -----------------------------------------------
# shutdown_after_replication.sh – Overview
#
# This script safely powers off the server when nightly ZFS replication
# has completed (or has not begun within a configured time window).
# Intended to be run via cron after a remote server wakes for replication.
#
# Workflow:
# 1. Script runs only if the /tmp/replication_mode FLAG is present.
#    (Flag indicates an automated boot for replication. Without it, exit.)
#
# 2. Prevents simultaneous executions using a /tmp/shutdown.lock file.
#
# 3. Records the first time the script notices the FLAG via START_TIME_FILE.
#    Exits early while waiting for replication to start.
#
# 4. Waits up to WAIT_WINDOW (default 1h) for replication activity.
#    If still waiting, logs progress and exits.
#
# 5. Checks for active replication or scrub work by scanning system state:
#    - Uses loginctl / job checks to detect zfs send/receive or netcat.
#    - Checks `zpool status tank` for scrubbing.
#    If active, exits and tries again on next cron run.
#
# 6. When replication looks finished or never started, enters a 5-minute
#    grace period and re-checks in case a job resumed.
#
# 7. If still idle, removes FLAG and timing files and shuts down host.
#
# Purpose:
# Ensures automated shutdown only happens when replication is complete,
# avoiding cutting power during ZFS activity or while a scrub is running.
#
# -----------------------------------------------
# SCRIPT CONFIGURATION

WAIT_WINDOW=3600   # seconds to wait for replication to begin (1 hour)

FLAG="/tmp/replication_mode"
LOG="/var/log/shutdown-after-replication.log"
LOCK_FILE="/tmp/shutdown.lock"
START_TIME_FILE="/tmp/replication_wait_start"

# -----------------------------------------------

if [ ! -f "$FLAG" ]; then
  echo "$(date) - No flag present, manual boot assumed. Exiting."
  exit 0
fi

# Prevent multiple simultaneous executions
if [ -f "$LOCK_FILE" ]; then
  echo "$(date) - Shutdown lock active, exiting." | tee -a "$LOG"
  exit 0
fi

# Record first time script sees the flag
if [ ! -f "$START_TIME_FILE" ]; then
  date +%s > "$START_TIME_FILE"
  echo "$(date) - Waiting for replication to begin..." | tee -a "$LOG"
  exit 0
fi

START_TIME=$(cat "$START_TIME_FILE")
NOW=$(date +%s)
ELAPSED=$((NOW - START_TIME))

# If not running yet, but within wait window, hold
if [ "$ELAPSED" -lt "$WAIT_WINDOW" ]; then
  echo "$(date) - Waiting ($ELAPSED / $WAIT_WINDOW seconds) for replication to start..." | tee -a "$LOG"
  exit 0
fi

# Loop while zfs send or netcat are running
if loginctl | awk '{print $2}' | grep -q '^0$' || \
      zpool status tank | grep -q 'scrub in progress'; then
  echo "$(date) - Replication or Scrub in progress..." | tee -a "$LOG"
  exit 0
fi

# If flag exists and replication is finished or never started, safe shutdown
echo "$(date) - Replication window expired or completed. Shutting down after 5m..." | tee -a "$LOG"

touch "$LOCK_FILE"
sleep 300

# final check
if loginctl | awk '{print $2}' | grep -q '^0$' || \
      zpool status tank | grep -q 'scrub in progress'; then
  echo "$(date) - Replication or scrub restarted during grace period!" | tee -a "$LOG"
  rm -f "$LOCK_FILE"
  exit 0
fi

# all good, cul8r
echo "$(date) - TTFN..." | tee -a "$LOG"
rm -f "$LOCK_FILE" "$FLAG" "$START_TIME_FILE"
shutdown -P now

wake_backup_server_for_replication.sh

#!/bin/bash

# -----------------------------------------------
# SCRIPT CONFIGURATION

BACKUP_MAC="d0:xx:xx:xx:xx:fc"      # <-- Backup Server NIC MAC address TO WAKE
HOST_WOL_IFACE="enp2s0"             # <-- Host NIC MAC address to send magic packet FROM
BACKUP_HOST="192.168.1.11"          # <-- hostname or IP address of backup server
SSH_USER="root"
LOG="/var/log/wake_backup_server_for_replication.log"

# -----------------------------------------------

echo "$(date) - Starting replication wake script" | tee -a "$LOG"

# 1) Wake Backup Server using etherwake
echo "$(date) - Sending Wake-on-LAN magic packet to $BACKUP_MAC" | tee -a "$LOG"
/root/sbin/etherwake -i "$HOST_WOL_IFACE" "$BACKUP_MAC"

# 2) Wait until Backup Host responds to ping
echo -n "$(date) - Waiting for $BACKUP_HOST to respond " | tee -a "$LOG"
while ! ping -c1 -W1 "$BACKUP_HOST" >/dev/null 2>&1 ; do
  echo -n "." | tee -a "$LOG"
  sleep 30
done
echo " UP" | tee -a "$LOG"

# 3) Allow time for SSH to become ready
echo "$(date) - Waiting for Backup Host SSH readiness..." | tee -a "$LOG"
sleep 30

# 4) Create flag file indicating auto-shutdown is allowed
echo "$(date) - Setting replication_mode flag on $BACKUP_HOST" | tee -a "$LOG"
if ! ssh "${SSH_USER}@${BACKUP_HOST}" 'date > /tmp/replication_mode'; then
  echo "$(date) - ERROR: Failed to set replication_mode flag on $BACKUP_HOST" | tee -a "$LOG" >&2
  exit 1
fi

echo "$(date) - replication_mode flag created successfully" | tee -a "$LOG"