I had Openfire 4.8.1 installed on a server on my local LAN on 172.10.1.53 , with port forwarding setup to allow connecting from the internet. Users could connect to user@mydomain.tld. and mydomain.tld pointed to my fiber IP address, which then routed to Openfire on 172.10.1.53
Last night, I setup a Debian server on the internet with a public IP address, 156.38.y.z, and updated the DNS zone record to point mydomain.tld to 156.38.y.z.
In openfire settings, both “XMPP Domain Name” and “Server Host Name (FQDN)” is setup as mydomain.tld
This is likely due to propagation of DNS records not yet being completed.
It takes time for your registrar to distribute the new IP address to its authoritative DNS servers, and then recursive DNS servers (such as your ISP) cache IP addresses for a period of time, this cache needs to expire before it will fetch the new IP address. (Set a lower TTL (Time To Live) to speed this up)
If you give me the domain name I can attempt a s2s connection to your server to see if I can reach it from my server?
Its been 20 hours, DNS propagation should have finished.
No. It’s not DNS propagation. The DNS TTL is set to 600 seconds, and have propagated long ago. Even if I add the public and hostname to my local /etc/hosts folder, and connect to the correct IP address, Openfire still doesn’t work.
This looks like a networking error, not an Openfire configuration error. You could try to verify this by doing packet inspection on the server that’s running Openfire, to see if it receives any data at all.
My guess is that there’s something wrong with your DNS records. Maybe you have added A records, but have lingering old records?
Note that by default, Openfire 4.8.0 and later won’t let you connect to the admin console (9090/9091) from anywhere but the localhost. That limitation obviously does not exist for the regular client ports (5222 et al).