feat: extend backup script and add restore script

backup.sh now covers all data:
- SQLite via podman exec into server container (fallback to host path)
- Postgres via pg_dumpall inside postgres container
- Forgejo data volume via podman volume export
- Caddy TLS certificates via podman volume export
- .env file (plaintext secrets — store archive securely)

restore.sh reverses each step: imports volumes, restores Postgres,
restores SQLite, optionally restores .env (--force to overwrite).

Both scripts find containers dynamically via compose service labels
so they work regardless of the container name podman-compose assigns.

.env.example documents HIY_BACKUP_DIR, HIY_BACKUP_REMOTE,
HIY_BACKUP_RETAIN_DAYS.

https://claude.ai/code/session_01FKCW3FDjNFj6jve4niMFXH

infra/.env.example

@@ -15,3 +15,12 @@ POSTGRES_PASSWORD=changeme
 # Forgejo (optional — only needed if you add the forgejo service to docker-compose.yml).
 FORGEJO_DB_PASSWORD=changeme
 FORGEJO_DOMAIN=git.yourdomain.com
+
+# ── Backup (infra/backup.sh) ──────────────────────────────────────────────────
+# Local directory to store backup archives.
+HIY_BACKUP_DIR=/mnt/usb/hiy-backups
+# Optional rclone remote (e.g. "b2:mybucket/hiy", "s3:mybucket/hiy").
+# Requires rclone installed and configured. Leave blank to skip remote upload.
+HIY_BACKUP_REMOTE=
+# How many days to keep local archives (default 30).
+HIY_BACKUP_RETAIN_DAYS=30

infra/backup.sh

@@ -3,17 +3,21 @@
 #
 # What is backed up:
 #   1. SQLite database (hiy.db)        — apps, deploys, env vars, users
-#   2. Env files directory       — decrypted env files written per deploy
+#   2. Env files                        — per-deploy decrypted env files
 #   3. Git repos                        — bare repos for git-push deploys
+#   4. Postgres                         — pg_dumpall (hiy + forgejo databases)
+#   5. Forgejo data volume              — repositories, avatars, LFS objects
+#   6. Caddy TLS certificates           — caddy-data volume
+#   7. .env file                        — secrets (handle the archive with care)
 #
 # Destination options (mutually exclusive; set one):
 #   HIY_BACKUP_DIR    — local path (e.g. /mnt/usb/hiy-backups, default /tmp/hiy-backups)
 #   HIY_BACKUP_REMOTE — rclone remote:path (e.g. "b2:mybucket/hiy")
 #                       requires rclone installed and configured
 #
-# Retention: 30 days (local only; remote retention is managed by the storage provider)
+# Retention: 30 days local (remote retention managed by the storage provider).
 #
-# Suggested cron (run as the same user as hiy-server):
+# Suggested cron (run as the same user that owns the containers):
 #   0 3 * * * /path/to/infra/backup.sh >> /var/log/hiy-backup.log 2>&1
 
 set -euo pipefail
@@ -24,6 +28,11 @@ BACKUP_DIR="${HIY_BACKUP_DIR:-/tmp/hiy-backups}"
 BACKUP_REMOTE="${HIY_BACKUP_REMOTE:-}"
 RETAIN_DAYS="${HIY_BACKUP_RETAIN_DAYS:-30}"
 
+# Load .env from the repo root (one level up from infra/) so the backup cron
+# can find HIY_DATA_DIR, container names, etc. without extra shell setup.
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+ENV_FILE="${SCRIPT_DIR}/../.env"
+
 TIMESTAMP=$(date +%Y%m%d-%H%M%S)
 ARCHIVE_NAME="hiy-backup-${TIMESTAMP}.tar.gz"
 STAGING="${BACKUP_DIR}/staging-${TIMESTAMP}"
@@ -34,30 +43,88 @@ log "=== HIY Backup ==="
 log "Data dir : ${HIY_DATA_DIR}"
 log "Staging  : ${STAGING}"
 
-# ── 1. Stage files ─────────────────────────────────────────────────────────────
 mkdir -p "${STAGING}"
 
-# SQLite: use the .dump command to produce a portable SQL text dump.
+# ── Helper: find a running container by compose service label ──────────────────
-if [ -f "${HIY_DATA_DIR}/hiy.db" ]; then
+find_container() {
-    log "Dumping SQLite database…"
+    local service="$1"
+    podman ps --filter "label=com.docker.compose.service=${service}" \
+              --format '{{.Names}}' | head -1
+}
+
+# ── 1. SQLite ──────────────────────────────────────────────────────────────────
+log "--- SQLite ---"
+SERVER_CTR=$(find_container server)
+if [ -n "${SERVER_CTR}" ]; then
+    log "Dumping hiy.db via container ${SERVER_CTR}…"
+    podman exec "${SERVER_CTR}" sqlite3 "${HIY_DATA_DIR}/hiy.db" .dump \
+        > "${STAGING}/hiy.sql"
+elif [ -f "${HIY_DATA_DIR}/hiy.db" ]; then
+    log "Server container not running — dumping from host path…"
     sqlite3 "${HIY_DATA_DIR}/hiy.db" .dump > "${STAGING}/hiy.sql"
 else
-    log "WARNING: ${HIY_DATA_DIR}/hiy.db not found — skipping SQLite dump"
+    log "WARNING: hiy.db not found — skipping SQLite dump"
 fi
 
-# Env files (contain decrypted secrets — handle with care).
+# ── 2. Env files ───────────────────────────────────────────────────────────────
-if [ -d "${HIY_DATA_DIR}/envs" ]; then
+log "--- Env files ---"
-    log "Copying env files…"
+if [ -n "${SERVER_CTR}" ]; then
-    cp -r "${HIY_DATA_DIR}/envs" "${STAGING}/envs"
+    podman exec "${SERVER_CTR}" sh -c \
+        "[ -d ${HIY_DATA_DIR}/envs ] && tar -C ${HIY_DATA_DIR} -czf - envs" \
+        > "${STAGING}/envs.tar.gz" 2>/dev/null || true
+elif [ -d "${HIY_DATA_DIR}/envs" ]; then
+    tar -czf "${STAGING}/envs.tar.gz" -C "${HIY_DATA_DIR}" envs
 fi
 
-# Bare git repos.
+# ── 3. Git repos ───────────────────────────────────────────────────────────────
-if [ -d "${HIY_DATA_DIR}/repos" ]; then
+log "--- Git repos ---"
-    log "Copying git repos…"
+if [ -n "${SERVER_CTR}" ]; then
-    cp -r "${HIY_DATA_DIR}/repos" "${STAGING}/repos"
+    podman exec "${SERVER_CTR}" sh -c \
+        "[ -d ${HIY_DATA_DIR}/repos ] && tar -C ${HIY_DATA_DIR} -czf - repos" \
+        > "${STAGING}/repos.tar.gz" 2>/dev/null || true
+elif [ -d "${HIY_DATA_DIR}/repos" ]; then
+    tar -czf "${STAGING}/repos.tar.gz" -C "${HIY_DATA_DIR}" repos
 fi
 
-# ── 2. Create archive ──────────────────────────────────────────────────────────
+# ── 4. Postgres ────────────────────────────────────────────────────────────────
+log "--- Postgres ---"
+PG_CTR=$(find_container postgres)
+if [ -n "${PG_CTR}" ]; then
+    log "Running pg_dumpall via container ${PG_CTR}…"
+    podman exec "${PG_CTR}" pg_dumpall -U hiy_admin \
+        > "${STAGING}/postgres.sql"
+else
+    log "WARNING: postgres container not running — skipping Postgres dump"
+fi
+
+# ── 5. Forgejo data volume ─────────────────────────────────────────────────────
+log "--- Forgejo volume ---"
+if podman volume exists forgejo-data 2>/dev/null; then
+    log "Exporting forgejo-data volume…"
+    podman volume export forgejo-data > "${STAGING}/forgejo-data.tar"
+else
+    log "forgejo-data volume not found — skipping"
+fi
+
+# ── 6. Caddy TLS certificates ──────────────────────────────────────────────────
+log "--- Caddy volume ---"
+if podman volume exists caddy-data 2>/dev/null; then
+    log "Exporting caddy-data volume…"
+    podman volume export caddy-data > "${STAGING}/caddy-data.tar"
+else
+    log "caddy-data volume not found — skipping"
+fi
+
+# ── 7. .env file ───────────────────────────────────────────────────────────────
+log "--- .env ---"
+if [ -f "${ENV_FILE}" ]; then
+    cp "${ENV_FILE}" "${STAGING}/dot-env"
+    log "WARNING: archive contains plaintext secrets — store it securely"
+else
+    log ".env not found at ${ENV_FILE} — skipping"
+fi
+
+# ── Create archive ─────────────────────────────────────────────────────────────
 mkdir -p "${BACKUP_DIR}"
 ARCHIVE_PATH="${BACKUP_DIR}/${ARCHIVE_NAME}"
 log "Creating archive: ${ARCHIVE_PATH}"
@@ -67,19 +134,19 @@ rm -rf "${STAGING}"
 ARCHIVE_SIZE=$(du -sh "${ARCHIVE_PATH}" | cut -f1)
 log "Archive size: ${ARCHIVE_SIZE}"
 
-# ── 3. Upload to remote (optional) ────────────────────────────────────────────
+# ── Upload to remote (optional) ────────────────────────────────────────────────
 if [ -n "${BACKUP_REMOTE}" ]; then
     if command -v rclone &>/dev/null; then
-        log "Uploading to remote: ${BACKUP_REMOTE}"
+        log "Uploading to ${BACKUP_REMOTE}…"
         rclone copy "${ARCHIVE_PATH}" "${BACKUP_REMOTE}/"
         log "Upload complete."
     else
-        log "WARNING: HIY_BACKUP_REMOTE is set but rclone is not installed — skipping upload"
+        log "WARNING: HIY_BACKUP_REMOTE is set but rclone is not installed — skipping"
-        log "Install rclone: https://rclone.org/install/"
+        log "Install: https://rclone.org/install/"
     fi
 fi
 
-# ── 4. Rotate old local backups ────────────────────────────────────────────────
+# ── Rotate old local backups ───────────────────────────────────────────────────
 log "Removing local backups older than ${RETAIN_DAYS} days…"
 find "${BACKUP_DIR}" -maxdepth 1 -name 'hiy-backup-*.tar.gz' \
     -mtime "+${RETAIN_DAYS}" -delete

infra/restore.sh

@@ -0,0 +1,143 @@
+#!/usr/bin/env bash
+# HIY restore script
+#
+# Restores a backup archive produced by infra/backup.sh.
+#
+# Usage:
+#   ./infra/restore.sh /path/to/hiy-backup-20260101-030000.tar.gz
+#
+# What is restored:
+#   1. SQLite database (hiy.db)
+#   2. Env files and git repos
+#   3. Postgres databases (pg_dumpall dump)
+#   4. Forgejo data volume
+#   5. Caddy TLS certificates
+#   6. .env file  (optional — skipped if already present unless --force is passed)
+#
+# ⚠  Run this with the stack STOPPED, then bring it back up afterwards:
+#      podman compose -f infra/docker-compose.yml down
+#      ./infra/restore.sh hiy-backup-*.tar.gz
+#      podman compose -f infra/docker-compose.yml up -d
+
+set -euo pipefail
+
+ARCHIVE="${1:-}"
+FORCE="${2:-}"
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+ENV_FILE="${SCRIPT_DIR}/../.env"
+HIY_DATA_DIR="${HIY_DATA_DIR:-/data}"
+
+log()  { echo "[hiy-restore] $(date '+%H:%M:%S') $*"; }
+die()  { log "ERROR: $*"; exit 1; }
+
+# ── Validate ───────────────────────────────────────────────────────────────────
+[ -z "${ARCHIVE}" ] && die "Usage: $0 <archive.tar.gz> [--force]"
+[ -f "${ARCHIVE}" ] || die "Archive not found: ${ARCHIVE}"
+
+WORK_DIR=$(mktemp -d)
+trap 'rm -rf "${WORK_DIR}"' EXIT
+
+log "=== HIY Restore ==="
+log "Archive : ${ARCHIVE}"
+log "Work dir: ${WORK_DIR}"
+
+log "Extracting archive…"
+tar -xzf "${ARCHIVE}" -C "${WORK_DIR}"
+
+# ── Helper: find a running container by compose service label ──────────────────
+find_container() {
+    local service="$1"
+    podman ps --filter "label=com.docker.compose.service=${service}" \
+              --format '{{.Names}}' | head -1
+}
+
+# ── 1. .env file ───────────────────────────────────────────────────────────────
+log "--- .env ---"
+if [ -f "${WORK_DIR}/dot-env" ]; then
+    if [ -f "${ENV_FILE}" ] && [ "${FORCE}" != "--force" ]; then
+        log "SKIP: ${ENV_FILE} already exists (pass --force to overwrite)"
+    else
+        cp "${WORK_DIR}/dot-env" "${ENV_FILE}"
+        log "Restored .env to ${ENV_FILE}"
+    fi
+else
+    log "No .env in archive — skipping"
+fi
+
+# ── 2. SQLite ──────────────────────────────────────────────────────────────────
+log "--- SQLite ---"
+if [ -f "${WORK_DIR}/hiy.sql" ]; then
+    DB_PATH="${HIY_DATA_DIR}/hiy.db"
+    mkdir -p "$(dirname "${DB_PATH}")"
+    if [ -f "${DB_PATH}" ]; then
+        log "Moving existing hiy.db to hiy.db.bak…"
+        mv "${DB_PATH}" "${DB_PATH}.bak"
+    fi
+    log "Restoring hiy.db…"
+    sqlite3 "${DB_PATH}" < "${WORK_DIR}/hiy.sql"
+    log "SQLite restored."
+else
+    log "No hiy.sql in archive — skipping"
+fi
+
+# ── 3. Env files & git repos ───────────────────────────────────────────────────
+log "--- Env files ---"
+if [ -f "${WORK_DIR}/envs.tar.gz" ]; then
+    log "Restoring envs/…"
+    tar -xzf "${WORK_DIR}/envs.tar.gz" -C "${HIY_DATA_DIR}"
+fi
+
+log "--- Git repos ---"
+if [ -f "${WORK_DIR}/repos.tar.gz" ]; then
+    log "Restoring repos/…"
+    tar -xzf "${WORK_DIR}/repos.tar.gz" -C "${HIY_DATA_DIR}"
+fi
+
+# ── 4. Postgres ────────────────────────────────────────────────────────────────
+log "--- Postgres ---"
+if [ -f "${WORK_DIR}/postgres.sql" ]; then
+    PG_CTR=$(find_container postgres)
+    if [ -n "${PG_CTR}" ]; then
+        log "Restoring Postgres via container ${PG_CTR}…"
+        # Drop existing connections then restore.
+        podman exec -i "${PG_CTR}" psql -U hiy_admin -d postgres \
+            -c "SELECT pg_terminate_backend(pid) FROM pg_stat_activity WHERE datname IN ('hiy','forgejo') AND pid <> pg_backend_pid();" \
+            > /dev/null 2>&1 || true
+        podman exec -i "${PG_CTR}" psql -U hiy_admin -d postgres \
+            < "${WORK_DIR}/postgres.sql"
+        log "Postgres restored."
+    else
+        log "WARNING: postgres container not running"
+        log "  Start Postgres first, then run:"
+        log "  podman exec -i <postgres_container> psql -U hiy_admin -d postgres < ${WORK_DIR}/postgres.sql"
+    fi
+else
+    log "No postgres.sql in archive — skipping"
+fi
+
+# ── 5. Forgejo data volume ─────────────────────────────────────────────────────
+log "--- Forgejo volume ---"
+if [ -f "${WORK_DIR}/forgejo-data.tar" ]; then
+    log "Importing forgejo-data volume…"
+    podman volume exists forgejo-data 2>/dev/null || podman volume create forgejo-data
+    podman volume import forgejo-data "${WORK_DIR}/forgejo-data.tar"
+    log "forgejo-data restored."
+else
+    log "No forgejo-data.tar in archive — skipping"
+fi
+
+# ── 6. Caddy TLS certificates ──────────────────────────────────────────────────
+log "--- Caddy volume ---"
+if [ -f "${WORK_DIR}/caddy-data.tar" ]; then
+    log "Importing caddy-data volume…"
+    podman volume exists caddy-data 2>/dev/null || podman volume create caddy-data
+    podman volume import caddy-data "${WORK_DIR}/caddy-data.tar"
+    log "caddy-data restored."
+else
+    log "No caddy-data.tar in archive — skipping"
+fi
+
+log "=== Restore complete ==="
+log "Bring the stack back up with:"
+log "  podman compose -f ${SCRIPT_DIR}/docker-compose.yml up -d"