backup-now script only contains last database when multiple databases configured

[Noiden]

Summary

When configuring multiple databases (DB01, DB02, etc.), the /usr/bin/backup-now script only contains the last database backup command instead of all configured databases. This causes manual backups to only execute the final database instead of all configured databases.

Steps to reproduce

1. Configure multiple databases in docker-compose.yml:

environment:
  - DB01_TYPE=postgres
  - DB01_HOST=postgresql
  - DB01_NAME=ALL
  - DB01_USER=testuser
  - DB01_PASS=testpass
  
  - DB02_TYPE=mysql
  - DB02_HOST=mariadb
  - DB02_NAME=ALL
  - DB02_USER=root
  - DB02_PASS=testpass

2. Start the container and verify individual scripts exist:

docker exec container-name ls -la /usr/bin/backup*

3. Check the content of the backup-now script:

docker exec container-name cat /usr/bin/backup-now

What is the expected correct behavior?

The /usr/bin/backup-now script should contain all configured database backup commands:

#!/bin/bash
/usr/bin/backup01-now now
/usr/bin/backup02-now now

When executing backup-now, all configured databases should be backed up sequentially.

Relevant logs and/or screenshots

Actual output of backup-now (incorrect):

#!/bin/bash
/usr/bin/backup02-now now

Expected output of backup-now (correct):

#!/bin/bash
/usr/bin/backup01-now now
/usr/bin/backup02-now now

Directory listing shows all individual scripts exist:

-rwxr-xr-x 1 root root  26 Jul  3 00:10 backup-now
-rwxr-xr-x 1 root root 221 Jul  3 00:10 backup01-now
-rwxr-xr-x 1 root root 221 Jul  3 00:10 backup02-now

Environment

• Image version / tag: tiredofit/db-backup:4.1.19
• Host OS: Docker on Linux (DietPi)

docker-compose.yml

services:
  db-backup:
    container_name: db-backup
    image: tiredofit/db-backup
    environment:
      - CONTAINER_NAME=db-backup
      - CONTAINER_ENABLE_MONITORING=FALSE
      
      # Backup settings
      - DEFAULT_BACKUP_LOCATION=FILESYSTEM
      - BACKUP_JOB_CONCURRENCY=1
      - DEFAULT_CHECKSUM=MD5
      - DEFAULT_COMPRESSION=ZSTD
      - DEFAULT_BACKUP_INTERVAL=1440
      - DEFAULT_BACKUP_BEGIN=0100
      - DEFAULT_CLEANUP_TIME=8640

      # PostgreSQL backup
      - DB01_TYPE=postgres
      - DB01_HOST=postgresql
      - DB01_NAME=ALL
      - DB01_USER=${POSTGRES_USER}
      - DB01_PASS=${POSTGRES_PASSWORD}
      - DB01_AUTH=postgres

      # MariaDB backup
      - DB02_TYPE=mysql
      - DB02_HOST=mariadb
      - DB02_NAME=ALL
      - DB02_USER=root
      - DB02_PASS=${MYSQL_ROOT_PASSWORD}

    restart: always
    volumes:
      - /mnt/backup/db:/backup
    networks:
      - postgres_network
      - mariadb_default

networks:
  postgres_network:
    external: true
  mariadb_default:
    external: true

Possible fixes

The issue is in /assets/functions/10-db-backup around line 1270. The script generation logic uses > (overwrite) instead of >> (append) for database instances other than "01":

Current code:

if [ "${instance}" = "01" ] ; then
    touch /usr/bin/backup-now
    chmod +x /usr/bin/backup-now
    cat <<EOF > /usr/bin/backup-now
#!/bin/bash
/usr/bin/backup${instance}-now now
EOF
else
    echo "/usr/bin/backup${instance}-now now" > /usr/bin/backup-now  # BUG: Should be >>
fi

Fix: Change line ~1270 from:

echo "/usr/bin/backup${instance}-now now" > /usr/bin/backup-now

To:

echo "/usr/bin/backup${instance}-now now" >> /usr/bin/backup-now

This will append subsequent database commands to the backup-now script instead of overwriting the entire file.

Workaround: Users can manually fix the script after container startup:

docker exec container-name sh -c 'cat > /usr/bin/backup-now << EOF
#!/bin/bash
/usr/bin/backup01-now now
/usr/bin/backup02-now now
EOF'

[MCP-LTS]

i confirm this bug and a hint is , for container images until and this version was working tiredofit/db-backup:4.1.18 if that helps. for now i use this version 4.1.18 to keep multiple dbs backups working as before