BillionMail Docker Deployment: Web UI Accessible but Login Fails ("Invalid request! Data not saved.")
Hello!
I am currently trying to deploy BillionMail on my NAS device using your provided Docker Compose configuration. I have encountered some issues during the deployment and hope to receive your assistance.
Environment Overview:
Operating System: Debian 12 (running on a NAS device) Docker Version: Docker Engine 26.1.3, Docker Compose v2.27.0 BillionMail Version: Image versions are billionmail/core:1.4, billionmail/dovecot:1.2, billionmail/postfix:1.2, billionmail/rspamd:1.1, and Roundcube is roundcube/roundcubemail:1.6.10-fpm-alpine. NAS IP Address: 192.168.3.222 Deployment Directory: /opt/BillionMail Problem Description:
I have successfully started all services using docker compose up -d. The docker-proxy on the host machine is also correctly listening on the mapped ports. My browser can now access the BillionMail Roundcube login interface (e.g., http://192.168.3.222:8080/roundcube or https://192.168.3.222:8443/roundcube).
Main Issue: When I attempt to log in to Roundcube using the administrator credentials defined in the .env file (ADMIN_USERNAME=dfxFMwVF, ADMIN_PASSWORD=x1z5YRES), the login fails with the page displaying: "Invalid request! Data not saved."
Detailed Diagnostic Process and Confirmed Information:
- Initial Network Connectivity Issue Resolved:
- Initially, due to a mismatch in port mapping for the core-billionmail service in docker-compose.yml (mapped to container internal 80/443, but the internal process was listening on 8080/8443), external access resulted in "Connection refused."
- Solution: I modified the ports configuration for the core-billionmail service in docker-compose.yml to match the actual listening ports observed by netstat -tulnp inside the billionmail-core-billionmail-1 container: yaml ports: - "${HTTP_PORT:-8080}:8080" - "${HTTPS_PORT:-8443}:8443"
- Result: After the modification, sudo docker compose ps shows the correct port mapping (0.0.0.0:8080->8080/tcp, 0.0.0.0:8443->8443/tcp). I can now successfully access the Roundcube login page from my browser.
- Inter-container Network Connectivity is Normal:
- Executing ping dovecot-billionmail -c 3 from within the billionmail-webmail-billionmail-1 container showed zero packet loss and very low latency. This confirms that the Roundcube container can resolve and access the Dovecot container successfully. roundcube:/var/www/html# ping dovecot-billionmail -c 3 PING dovecot-billionmail (172.66.1.6): 56 data bytes 64 bytes from 172.66.1.6: seq=0 ttl=64 time=0.169 ms ... --- dovecot-billionmail ping statistics --- 3 packets transmitted, 3 packets received, 0% packet loss
- .env File Content: I have confirmed that the administrator credentials are defined in the .env file:
# Default BillionMail Username password
ADMIN_USERNAME=dfxFMwVF
ADMIN_PASSWORD=x1z5YRES
# Manage Safe entrance
SafePath=v5zVwQwu
I attempted to log in to Roundcube using dfxFMwVF and x1z5YRES. I also tried accessing the management panel path http://192.168.3.222:8080/v5zVwQwu (using the same credentials), but encountered the "Invalid request! Data not saved." error on both.
4. Log Check Results (immediately after attempting login):
- billionmail-dovecot-billionmail-1 container logs: Showed no error messages related to authentication failure, only container startup success logs.
- billionmail-postfix-billionmail-1 container logs: Also showed no error messages related to authentication failure, only container startup success logs.
- billionmail-core-billionmail-1 container logs: (I have not provided these to you, but I have checked them and found no obvious errors. I can provide them again if needed.)
- billionmail-webmail-billionmail-1 container logs: (I have not provided these to you, but I have checked them and found no obvious errors. I can provide them again if needed.)
- billionmail-pgsql-billionmail-1 container logs: (I have not provided these to you, but I have checked them and found no obvious errors. I can provide them again if needed.)
My Suspicions and Areas for Further Investigation:
The "Invalid request! Data not saved." message is highly likely related to Roundcube's CSRF protection mechanism, session management, or its interaction with the PostgreSQL database (pgsql-billionmail), rather than a direct username/password error. It's possible that the web server (Nginx/Apache) configuration inside the billionmail-core-billionmail-1 container is not correctly processing Roundcube's POST requests, leading to data not being fully forwarded or validated. The Roundcube container (webmail-billionmail-1) might be unable to write session data to its designated location, or unable to manage sessions with the pgsql-billionmail database. I kindly request your assistance with the following:
Please confirm whether the ADMIN_USERNAME and ADMIN_PASSWORD in the .env file are indeed the initial login credentials for Roundcube or the BillionMail management panel. Please guide me on how to further diagnose the "Invalid request! Data not saved." error, specifically within the context of your BillionMail container architecture. For example, whether I need to inspect the Roundcube config.inc.php file inside the webmail-billionmail-1 container, or if there are specific initialization scripts that need to be run. Please let me know if you require any other logs or configuration file contents. Thank you for your time and help!
Yes, ADMIN_USERNAME and ADMIN_PASSWORD is login credentials for BillionMail Console Panel. Can you show me a screenshot?
I mean, take a screenshot of this page that says "Invalid request", and I'd like to pinpoint where this error is occurring.
Hey, please update to latest release. Maybe it fixed in the latest release.
Thanks to the author for fixing this issue. I can now log in without any error messages.