Updated update script.

[KentuckyFriedBlyat]

I ran into an issue where one of my update script examples (because the user didnt follow my directions precisely) was ending up with the script updating and making a backup and then making backups of the backups recursively..

Naturally, their server imploded when the hard drive filled up.

This should stop that from happening ever again.

#!/bin/bash
# =============================================
# Haven Safe Updater 1.9
# =============================================

set -euo pipefail

export PATH="/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:${PATH:-}"

# ====================== CONFIGURATION ======================
HAVEN_UPDATE_BACKUP_KEEP="${HAVEN_UPDATE_BACKUP_KEEP:-2}"
HEALTH_URL="${HEALTH_URL:-https://localhost:3000/api/health}"
HEALTH_TIMEOUT=60
HEALTH_RETRIES=3
LOGFILE="${LOGFILE:-/var/log/haven_update.log}"

# Disk Space Configuration (in GB)
DISK_WARNING_GB=5
DISK_CRITICAL_GB=3

# Log Configuration
LOG_RETENTION_DAYS=30

# ====================== DIRECTORY DETECTION ======================
if [[ -n "${HAVEN_DIR:-}" ]]; then
    HAVEN_DIR="$(cd "$HAVEN_DIR" && pwd)"
elif [[ -f "docker-compose.yml" || -f "compose.yml" ]]; then
    HAVEN_DIR="$(pwd)"
elif [[ -f "$(dirname "${BASH_SOURCE[0]}")/docker-compose.yml" || -f "$(dirname "${BASH_SOURCE[0]}")/compose.yml" ]]; then
    SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
    HAVEN_DIR="$(dirname "$SCRIPT_DIR")"
else
    HAVEN_DIR=$(docker inspect haven --format '{{ index .Config.Labels "com.docker.compose.project.working_dir" }}' 2>/dev/null || echo "")
fi

if [[ -z "${HAVEN_DIR:-}" || ! -d "$HAVEN_DIR" ]]; then
    echo "ERROR: Could not auto-detect Haven directory. Please set HAVEN_DIR env var." >&2
    exit 1
fi

BACKUP_DIR="$HAVEN_DIR/hotfixes/haven-auto-update-backups"
mkdir -p "$BACKUP_DIR" "$(dirname "$LOGFILE")"

log() {
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1" | tee -a "$LOGFILE"
}

# ====================== LOG ROTATION (30 days) ======================
if [ -f "$LOGFILE" ]; then
    log "Performing log rotation (keeping last $LOG_RETENTION_DAYS days)..."
    temp_log=$(mktemp)
    awk '
        /=== Haven Docker update STARTED at/ {
            match($0, /[A-Za-z]+ [A-Za-z]+ [0-9]+ [0-9:]+ [A-Z]+ [0-9]+/)
            if (RSTART) {
                datestr = substr($0, RSTART, RLENGTH)
                cmd = "date -d \"" datestr "\" +%s 2>/dev/null"
                cmd | getline ts
                close(cmd)
                if (ts != "" && systime() - ts <= '"$LOG_RETENTION_DAYS"' * 86400) {
                    print_line = 1
                } else {
                    print_line = 0
                }
            }
        }
        print_line { print }
    ' "$LOGFILE" > "$temp_log" && mv "$temp_log" "$LOGFILE"
fi

log "=== Haven Docker update STARTED at $(date) ==="
log "Haven directory: $HAVEN_DIR"
log "Health check URL: $HEALTH_URL"

cd "$HAVEN_DIR"

# ====================== DISK SPACE CHECK ======================
FREE_GB=$(df -BG . | awk 'NR==2 {gsub(/[^0-9]/,"",$4); print $4}')

if [ "${FREE_GB:-0}" -lt "$DISK_WARNING_GB" ]; then
    log "!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
    log "CRITICAL WARNING: Low disk space detected (${FREE_GB}GB available)."
    log "Warning threshold is ${DISK_WARNING_GB}GB."
    log "Reducing backup retention to 1."
    log "Please free up disk space IMMEDIATELY."
    log "!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
    HAVEN_UPDATE_BACKUP_KEEP=1
fi

if [ "${FREE_GB:-0}" -lt "$DISK_CRITICAL_GB" ]; then
    log "ERROR: Critical disk space limit reached (<${DISK_CRITICAL_GB}GB). Aborting."
    exit 1
fi

# ====================== CONFIG BACKUP ======================
BACKUP_NAME="pre-update_$(date +%Y%m%dT%H%M%SZ)"
log "Creating config backup..."
shopt -s nullglob
tar -czf "$BACKUP_DIR/$BACKUP_NAME.tar.gz" \
    docker-compose.yml compose.yml .env* *.env 2>/dev/null || true
shopt -u nullglob

find "$BACKUP_DIR" -name "pre-update_*.tar.gz" -type f | sort | head -n -${HAVEN_UPDATE_BACKUP_KEEP} | xargs -r rm -f

# ====================== MAIN UPDATE ======================
log "Pulling latest images..."
docker compose pull --policy=always 2>&1 | tee -a "$LOGFILE"

log "Recreating services..."
docker compose up -d --force-recreate 2>&1 | tee -a "$LOGFILE"

# ====================== HEALTH CHECK ======================
log "Performing health check ($HEALTH_RETRIES attempts)..."
for i in $(seq 1 $HEALTH_RETRIES); do
    sleep 8
    if curl -sSf -m $HEALTH_TIMEOUT -k -o /dev/null "$HEALTH_URL"; then
        log "✅ Health check passed on attempt $i"
        break
    elif [ $i -eq $HEALTH_RETRIES ]; then
        log "❌ Health check failed after $HEALTH_RETRIES attempts - Rolling back"
        docker compose down 2>&1 | tee -a "$LOGFILE"
        docker compose up -d 2>&1 | tee -a "$LOGFILE"
        exit 1
    else
        log "Health check failed (attempt $i/$HEALTH_RETRIES) - retrying..."
    fi
done

# ====================== CLEANUP ======================
log "Cleaning up old images..."
docker image prune -f --filter "until=168h" 2>&1 | tee -a "$LOGFILE"

log "✅ Haven Docker update COMPLETED successfully at $(date)"
echo "=== End ===" | tee -a "$LOGFILE"
echo "" | tee -a "$LOGFILE"

[ancsemi]

Appreciate you catching the recursive-backup blowup and posting the fix. The dynamic disk-space gate plus the hard 5GB abort is a nice touch. Nothing for me to change on the Haven side, but I'll point folks running into update-script issues at this updated 1.7 version. Thanks for sharing it 🙏