pmg8to9 FAIL: "Resolved node IP not configured or active" — expected for NAT'd gateway with public HELO hostname?

[gaaike]

Running pmg8to9 ahead of an 8.2→9.1 upgrade and getting one FAIL:

FAIL: Resolved node IP 'x.x.x.x' not configured or active for 'mail'

Setup: PMG VM behind pfSense NAT. Internal IP 10.5.10.204 is what's actually bound to the NIC. /etc/hosts intentionally maps the short hostname (mail) to the public IP (x.x.x.x), separate from the FQDN entries which map to the internal IP:
10.5.10.204 mail.example.com
x.x.x.x mail.example.com mail
This was set up deliberately so the SMTP HELO banner presents the publicly-resolvable, PTR-matching hostname rather than an internal RFC1918 address — standard practice for a gateway behind NAT. Mail flow, TLS, and HELO have all worked correctly under this setup for months.

Before running the actual 8→9 dist-upgrade: is this FAIL a known non-issue for NAT'd gateways with this pattern, or does something in the trixie transition (postfix rework, network handling changes, etc.) actually depend on the short hostname resolving to a locally-bound address? Want to understand whether to fix this pre-upgrade or if it's safe to proceed with it as-is.


Thanks in advance.

 

[Stoiko Ivanov]

This was set up deliberately so the SMTP HELO banner presents the publicly-resolvable, PTR-matching hostname rather than an internal RFC1918 address

Where does the HELO/EHLO banner mention the IP-address for you?
connecting to port 25, and issuing a EHLO testhost - to a PMG does not show any IP-addresses in the output coming from the server for me.

EDIT: without testing this explicitly - I'd assume that the upgrade should work even with your /etc/hosts as it is currently.