#!/usr/bin/env bash # Periodic pg_dump of the roboco DB (interim backup — no PITR/WAL archiving). # Runs once on start, then every BACKUP_INTERVAL_SECONDS (default 24h). A # failed dump logs and is retried next cycle; the loop itself never exits. # ponytail: a plain sleep loop, not a cron daemon — good enough for a fixed # 24h cadence, swap for a scheduler only if finer granularity is ever needed. set -u BACKUP_DIR="${BACKUP_DIR:-/backups}" KEEP="${BACKUP_KEEP:-14}" INTERVAL_SECONDS="${BACKUP_INTERVAL_SECONDS:-86400}" # pg_dump reads these natively — no need to pass -h/-p/-U/-d by hand. export PGHOST="${POSTGRES_HOST:-roboco-postgres}" export PGPORT="${POSTGRES_PORT:-5432}" export PGUSER="${POSTGRES_USER:-roboco}" export PGPASSWORD="${POSTGRES_PASSWORD:-roboco}" export PGDATABASE="${POSTGRES_DB:-roboco}" mkdir -p "$BACKUP_DIR" run_backup() { local ts dest ts="$(date -u +%Y%m%dT%H%M%SZ)" dest="${BACKUP_DIR}/roboco-${ts}.dump" echo "[backup] $(date -u -Iseconds) starting pg_dump -> ${dest}" if pg_dump -Fc -f "${dest}.tmp"; then mv "${dest}.tmp" "${dest}" echo "[backup] $(date -u -Iseconds) OK: ${dest}" else rm -f "${dest}.tmp" echo "[backup] $(date -u -Iseconds) FAILED — will retry next cycle" >&2 fi # Prune to the newest $KEEP dumps. Filenames are our own fixed timestamp # format (no spaces/globs to worry about), so plain `ls -t` is enough. # shellcheck disable=SC2012 ls -1t "${BACKUP_DIR}"/roboco-*.dump 2>/dev/null | tail -n "+$((KEEP + 1))" | while IFS= read -r old; do rm -f "$old" done } while true; do run_backup sleep "$INTERVAL_SECONDS" done