rspamd always MX_MISSING

Hi,
I noticed that my rspamd is always reporting MX_MISSING on every incoming email. I don't know when it started this, but sure not from the startup of the server.
DNS resolution on the host works well, so I cannot explain this.
ISPConfig is 3.2.11p2 running on debian 11
any idea where I can point?
thanks

 

Well, _always_ is bad - anything not working properly?
Keep in mind rspamd is written in python. I'm not up to speed but even if
DNS resolution on the host works, doesn't mean it works for python
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=877785

How did you try DNS resolution?
ping somehost or $dig google.com MX
?

 

I tried resolution with nslookup -> set q=mx -> any domains works
Also dig:

Code:

dig google.com MX

; <<>> DiG 9.18.24-1-Debian <<>> google.com MX
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 34788
;; flags: qr rd ra cd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
; PAD: (404 bytes)
;; QUESTION SECTION:
;google.com.                    IN      MX

;; ANSWER SECTION:
google.com.             157     IN      MX      10 smtp.google.com.

;; Query time: 11 msec
;; SERVER: 10.77.1.64#53(10.77.1.64) (UDP)
;; WHEN: Mon Apr 22 11:12:56 CEST 2024
;; MSG SIZE  rcvd: 468

 

Rspamd is written in C/C++ and uses LUA for some sort of extensibility or additional features that are supposed to be not part of the Rspamd C/C++ core.

Regarding the issue, I would also check if the OP may have accidently (mis-)configured dedicated DNS servers in Rspamd, more info here: https://rspamd.com/doc/configuration/options.html#dns-options

 

uh! you got it!

Code:

cat options.inc
# Addrs local to this server.
local_addrs = [
        "127.0.0.0/8",
        "::1",
        "10.77.1.61",
];

# This list is generated by ISPConfig, place custom addresses/networks in local_networks.inc.
local_networks = "/etc/rspamd/local.d/local_networks.inc";

dns {
        nameserver = ["127.0.0.1:53:10"];
}

Is this a bug? A standard settings? I'm not using local bind servers this i really strange as fixed setting

 

indeed, my head was thinking modules. In fact we are both wrong, it comes from
src/plugins/lua/mx_check.lua:33 =)

topogigio, well usually you have a local nameserver ( systemd ) accessible, so it is fine. It adds the caching layer and reduces the 9ms you'd need to lookup using external dns, however the :10 not sure

 

Yeh, it seems to be a bug.
This line should not exist on ISPConfig servers that have no DNS server installed.

Either we need to remove this line on such servers or we need to install something like unbound as a lightweight substitute for Bind/PowerDNS.

 

u thinking of this special service
systemctl status systemd-resolved
? ^^ well, one should not expect it to be active and if someone does, should add a status/monitoring/log something somewhere
or is it some bind config of ispconfig?

 

which again I could quote

add a "pay attention we expect something to run on that ip/port anyway" ?

 

thank you for the clarification, should've done my research guess it's me who needs to buy you the coffee now

 

I agree, that if one uses a mail system, that it is definitely recommended, that a local name resolver exists.
However, I don't use the auto-installer and I didn't have a look at its the source code yet, but I think that it should show a warning message if one tries to install a full mail system, that a local DNS resolver should exist/is necessary.

 

especially when you get to the point where you just use your own dns-resolver as it handles some company specific rules or lookups and resides next to the mx VM, not much latency there.
is that rspamd-nameserver configureable?

 

ok, so not for the average joe but its a var
fair enough though, ispconfig is meant to be in control