RSPAMD - ERROR http://fuzzy.mailcow.email/bad-subject-regex.txt

stefan21

RSPAMD reports: error reading http://fuzzy.mailcow.email/bad-subject-regex.txt([2a01:4f9:c010:687d::1]:80): connection with http server terminated incorrectly: IO write error: Network is unreachable

From a PC in the LAN, called from a browser:

http://fuzzy.mailcow.email/bad-subject-regex.txt is giving back:

/[0-9]+.+Weine.+Delikatessen.+[0-9]+.+Teller.+[0-9]+.+Euro/i
/12x.+[0-9]+.+Euro.+[0-9]+.+anmelden/i
/8x.+Rosé.+Gläser.+geschenkt.+schmeck.+Süden/i
/Account.+Details.+met.+name.+Kimberly.+want.+talk.+profile.+https.+tone.+affomelody.+com.+vWlFWs.+050UEwEyE7sOU.+Ctstrading/i
/Action.+Required.+Verify.+Wallet.+Avoid.+Access.+Issues/i
etc…

From mailcow cli:

dig fuzzy.mailcow.email

; <<>> DiG 9.18.33-1_deb12u2-Debian <<>> fuzzy.mailcow.email
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 20216
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;fuzzy.mailcow.email. IN A

;; ANSWER SECTION:
fuzzy.mailcow.email. 41078 IN A 95.217.129.125

;; Query time: 88 msec
;; SERVER: 192.168.30.100#53(192.168.30.100) (UDP)
;; WHEN: Thu Aug 28 09:12:17 CEST 2025
;; MSG SIZE rcvd: 64

and:

wget http://fuzzy.mailcow.email/bad-subject-regex.txt
–2025-08-28 09:24:01– http://fuzzy.mailcow.email/bad-subject-regex.txt
Auflösen des Hostnamens fuzzy.mailcow.email (fuzzy.mailcow.email)… 95.217.129.125, 2a01:4f9:c010:687d::1
Verbindungsaufbau zu fuzzy.mailcow.email (fuzzy.mailcow.email)|95.217.129.125|:80 … verbunden.
HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 OK
Länge: 8841 (8,6K) [text/plain]
Wird in »bad-subject-regex.txt« gespeichert.

bad-subject-regex.txt 100%[==================================>] 8,63K –.-KB/s in 0s

2025-08-28 09:24:01 (61,5 MB/s) - »bad-subject-regex.txt« gespeichert [8841/8841]

Why/Where does this error come from and how to solve it?

More info:

docker compose exec rspamd-mailcow /bin/bash
From inside the container:

root@rspamd:/# dig fuzzy.mailcow.email

; <<>> DiG 9.18.33-1_deb12u2-Debian <<>> fuzzy.mailcow.email
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 43902
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;fuzzy.mailcow.email. IN A

;; ANSWER SECTION:
fuzzy.mailcow.email. 41626 IN A 95.217.129.125

;; Query time: 0 msec
;; SERVER: 127.0.0.11#53(127.0.0.11) (UDP)
;; WHEN: Thu Aug 28 09:44:56 CEST 2025
;; MSG SIZE rcvd: 64

esackbauer

Could be a problem with IPv6 Stack at the host:
rspamd/rspamd448

stefan21

esackbauer

Thank’s for pointing in the right direction. Following all the way down disabling ip6 solved this. @devs: would nice to disable ip6 with one click. IMVHO not every installation needs ip6.

Must have been a change at docker. Didn’t see this error before the last docker update.

stefan21

More info:

cat /etc/debian_version
12.11

Mailcow is running on a proxmox as a VM.

System is up-to-date. No ip6 at all at the host, neither at the OPNsense (ip6 disabled), to be precise - no ip6 at all. Not at the WAN or any other interface. We do have only static ip4 addresses. Business contracts. Right now, we don’t need it.

docker version
Client: Docker Engine - Community
Version: 28.3.3
API version: 1.51
Go version: go1.24.5
Git commit: 980b856
Built: Fri Jul 25 11:34:00 2025
OS/Arch: linux/amd64
Context: default

Server: Docker Engine - Community
Engine:
Version: 28.3.3
API version: 1.51 (minimum version 1.24)
Go version: go1.24.5
Git commit: bea959c
Built: Fri Jul 25 11:34:00 2025
OS/Arch: linux/amd64
Experimental: true
containerd:
Version: 1.7.27
GitCommit: 05044ec0a9a75232cad458027ca83437aae3f4da
runc:
Version: 1.2.5
GitCommit: v1.2.5-0-g59923ef
docker-init:
Version: 0.19.0
GitCommit: de40ad0

For a test I’ll follow https://docs.mailcow.email/post_installation/firststeps-disable_ipv6/

I’ll report.

stefan21

Didin’t expect the spanish inquisition - bunch of errors are occurring. So hold it down…

I’ll try to track this down with:

https://docs.docker.com/engine/daemon/ipv6/#use-ipv6-for-the-default-bridge-network

and

Modify docker-compose.yml

Change enable_ipv6: true to enable_ipv6: false and comment out the IPv6 subnet:

networks:
mailcow-network:
[…]
enable_ipv6: true # <<< set to false
ipam:
driver: default
config:
- subnet: ${IPV4_NETWORK:-172.22.1}.0/24
- subnet: ${IPV6_NETWORK:-fd4d:6169:6c63:6f77::/64} # <<< comment out with #
[…]

Reverted the rest back…

Will report.

Not good. Leads even to more errors.

Stuck for the moment.

stefan21

While reading from the docs:
"
In installations using a Docker version between 25.0.0 and 25.0.2 (to check, use docker version) the behavior of IPv6 address allocation has changed due to a bug. Simply using enable_ipv6: false is NO LONGER sufficient to completely disable IPv6 in the stack.
This was a bug in the Docker Daemon, which has been fixed with version 25.0.3.
"

Not quite clear to me. While reading this again, now I understand using docker 28.3.3 it should be sufficient to modify the docker-compose.yml and change:

followed by docker compose down, systemctl restart docker and docker compose up.

Did that.

For the moment docker compose logs -f rspamd-mailcow is showing no errors for fuzzy.mailcow.email.

I’ll report.

stefan21

The original error didn’t show up again, since I made the changes to the docker-compose.yml.

BUT: now another issue is popping up:

lua 7803B1 signing failure: cannot make request to load DKIM selector for domain mail.xxx.de: nil

While the first error comes form the mailcow devs (the query goes to fuzzy.mailcow.email), this one seems to be more disturbing. I do not really need fuzzy.mailcow.email/bad-subject-regex.txt. The signing failure is not acceptable.

Right now I’d like to say with the o.m. docker version right now there’s no solution (at least not in the docs) to turn ipv6 off without any disfunction(s) or side effects in (to) mailcow.

Does anybody know where/what the query to http://fuzzy.mailcow.email/bad-subject-regex.txt([2a01:4f9:c010:687d::1]:80): is invoked? Can I turn it off? Can this changed in ipv4 only?

For now I’ll revert back and try to ignore the bad-subject-regex.txt error.

Not good. I don’t like errors on a mail server.

stefan21

New approach.

cd /opt/mailcow-dockerized
docker compose down
systemctl stop docker
Check/change/add/create if in host is /etc/sysctl.d/60-disable-ipv6.conf is:

net.ipv6.conf.all.disable_ipv6 = 1
net.ipv6.conf.default.disable_ipv6 = 1
net.ipv6.conf.lo.disable_ipv6 = 1
mv in /etc/docker the daemon.json to another name
sysctl -p
reboot the host.
Follow all the way down this: https://docs.mailcow.email/post_installation/firststeps-disable_ipv6/
systemctl restart docker

10 cd /opt/mailcow-dockerized
docker compose up -d

I’ll report.

stefan21

Odd.

lua 370901 signing failure: cannot make request to load DKIM selector for domain XXX.de: nil

I’d really like to know, someone out there running a mailcow only on ipv4 without errors in rspamd?

debian 12.11
docker 28.3.3.

Obviously following the docs disabling ipv6 does not work. It’s not soooo special to run a network with ipv4 only.

Hello?

Still investigating.

One error is from an email which was rejected from a local domain (backup server) without a valid DKIM signing. Should be o.k.

The other one was first a few times soft rejected, delivered to the original email receipient, and forwarded with an add header to another local email address.

Hmm? I’d like to understand this, better to have an explaination of it.

stefan21

I was told, that the lua error has nothing to do with disabling ipv6.

Following the last approach disables ipv6 and seems to work. No more resolving errors with ipv4 only for any domains in rspamd so far.