Hi Everyone,
If you’ve been wanting to integrate Rspamd into your Virtualmin installation as a replacement for Spamassassin, over the past few days I’ve begun the process of doing this with one of our customer’s servers.
The spam detection is night and day, and really should become the standard within Virtualmin IMHO, as out of the box works amazingly.
Getting the software installed and integrated with Postfix was relatively easy, and I even managed to configure a globally procmail rule to send detected messages to the Spam folder in the user’s mailbox, though the process did require a little creativity and hopefully won’t get overwritten during an upgrade in the future.
There are still a few todo items to work through as Virtualmin is very Spamassassin designed by default so somethings either had to be disabled or simply don’t work yet.
That being said, reports from the customer have already shown a significant reduction in spam, and messages being filtered into their Spam folder as intended.
If you’ve ever considered making the change, and would like help, please reach out.
2026-09-20:
I was able to get the “Report Spam” button in Usermin to submit message to Rspamd successfully. It required a bit of hack, but it works.
I’m also working on a full stack indepent of Virtualmin as a proof of concept. It utilizes MariaDB for user management, Rspam for spam, CrowdSec for IDS, Postfix/Dovecot for SMTP/IMAP, Sieve for message filtering, RoundCube for Webmail, and a custom Admin Panel for user management and user facing admin. Each of these elements are isolated in their own Incus Container environment but talk to each other internally, and only those containers requiring public access are setup with appropriate port forwarding from the host system. CrowdSec also runs the firewall bouncer on the host allowing it to write firewall rules protecting the full stack at the host level.