Mailcow-dockerized: Can receive mails, but not send

[Snack-X]

I can receive mails from external services like Gmail, but I can not send any mails.
I also checked Common Problems' corresponding section, my IP is not blacklisted, and my server's mail ports are not blocked at all.

---

From the postfix's log, I can see postfix has some problem.

$ docker-compose logs --tail=10 -f postfix-mailcow
(...)
postfix-mailcow_1    | May 17 02:40:52 mail postfix/qmgr[332]: 58FFE1C1F90: from=<[email protected]
>, size=683, nrcpt=1 (queue active)
postfix-mailcow_1    | May 17 02:41:12 mail postfix/smtp[363]: 58FFE1C1F90: to=<[email protected]
>, relay=none, delay=502, delays=482/0.08/20/0, dsn=4.4.3, status=deferred (Host or domain name not
found. Name service error for name=korsnack.kr type=MX: Host not found, try again)

snack.studio is the domain I want to set my servers up, and korsnack.kr is my other domain which is using Google's mail server.

$ dig MX snack.studio
(...)
snack.studio.           300     IN      MX      10 mail.snack.studio.

$ dig mail.snack.studio
(...)
mail.snack.studio.      300     IN      A       150.95.131.26

$ dig MX korsnack.kr
(...)
korsnack.kr.            60      IN      MX      10 aspmx.l.google.com.

Currently, my docker is behind Ubuntu's UFW, so it's an unusual setup and I think this is where the problem comes from.

# cat /etc/docker/daemon.json
{ "iptables": false }

$ docker exec -i -t mailcowdockerized_bind9-mailcow_1 /bin/ash
/ # ping 8.8.8.8
PING 8.8.8.8 (8.8.8.8): 56 data bytes
^C
--- 8.8.8.8 ping statistics ---
3 packets transmitted, 0 packets received, 100% packet loss
/ # ping korsnack.kr
ping: bad address 'korsnack.kr'

[Snack-X]

...and just after submitting this issue, I managed to solve this problem.
I'm just leaving what I found and what I did for the record.

---

$ iptables -t nat -L -n -v
(...)
Chain POSTROUTING (policy ACCEPT 192 packets, 15178 bytes)
 pkts bytes target     prot opt in     out     source               destination
    0     0 MASQUERADE  all  --  *      !docker0  172.17.0.0/16        0.0.0.0/0

By default, docker creates network interface docker0 and it routes every network packets from the docker containers(172.17.0.0/16) to other interface except docker0 interface itself.

But mailcow creates another network interface, so network packets from mailcow containers cannot go out.

So I modified docker-compose.yml a little bit to assign fixed network interface name, and manually added another rule.

$ git --no-pager diff docker-compose.yml
diff --git a/docker-compose.yml b/docker-compose.yml
index d5c4ead..a1ecd01 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -277,6 +277,8 @@ services:
 networks:
   mailcow-network:
     driver: bridge
+    driver_opts:
+      com.docker.network.bridge.name: br-mailcow
     enable_ipv6: true
     ipam:
       driver: default

$ iptables -t nat -A POSTROUTING -s 172.22.0.0/16 ! -o br-mailcow -j MASQUERADE
$ iptables -t nat -L -n -v
(...)
Chain POSTROUTING (policy ACCEPT 471 packets, 37597 bytes)
 pkts bytes target     prot opt in     out     source               destination
    0     0 MASQUERADE  all  --  *      !docker0  172.17.0.0/16        0.0.0.0/0
  466 34847 MASQUERADE  all  --  *      !br-mailcow  172.22.0.0/16        0.0.0.0/0

And, everything is working, for now.