Nameserver error?

maiyannah

I have a problem with some queued messages.
They report, for instance:

ping@tools.mxtoolbox.com (Host or domain name not found. Name service error for name=tools.mxtoolbox.com type=MX: Host not found, try again)

However, on the same machine, I can dig it just fine:

# dig tools.mxtoolbox.com

; <<>> DiG 9.20.15-1~deb13u1-Debian <<>> tools.mxtoolbox.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 5898
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 512
;; QUESTION SECTION:
;tools.mxtoolbox.com.           IN      A

;; AUTHORITY SECTION:
mxtoolbox.com.          900     IN      SOA     ns-1320.awsdns-37.org. awsdns-hostmaster.amazon.com. 1 7200 900 1209600 86400

;; Query time: 16 msec
;; SERVER: 8.8.8.8#53(8.8.8.8) (UDP)
;; WHEN: Wed Oct 29 18:06:31 EDT 2025
;; MSG SIZE  rcvd: 130

I don’t see anything untoward in the unbound log.

Any ideas where to look?

maiyannah

Monitoring the box a bit, it seems to just … randomly lose the ability to contact the outside world.

No idea why, still sifting.

maiyannah

Looks like its docker itself losing networking but not sure why it is. It will work for a period and then seem to trash its own internal iptables chains.

esackbauer

Are you on a supported platform? Which hoster are you on?
For such connections problems often MTU is the culprit:
https://docs.mailcow.email/getstarted/prerequisite-system/?h=mtu#mtu
https://docs.mailcow.email/getstarted/install/#important-notes
There are also some threads in the forum regarding this.

maiyannah

This is on the latest Debian Trixie in a proxmox VM, on a dedicated server I lease.

Rather than a Mailcow specific issue, this appears to be Docker related. Creating an alternative bridge network seems to have resolved it for now, but given the intermittent nature I’m going to have to keep an eye on.
Might be something with how they package Docker, if that fixes it.

maiyannah

Having had it going fine for a couple days now, sending/receiving fine, without losing the network, I’m going to go with the default docker network bridge being the failure state here.

For future people who come across this on Google: if you have intermittent network loss on Mailcow, consider creating a new network bridge in Docker, and running mailcow using it. (The default one, at time of writing, is immutable)