In a self-hosting support thread titled ‘Nginx not binding to ipv6 address using the Discourse Docker image,’ robbyoconnor reports that IPv6 access to their Discourse instance suddenly stopped working on January 7, while IPv4 remains functional. They confirm no recent configuration changes and note the instance runs on Rackspace’s OpenStack, soon to be decommissioned. Falco questions whether something was altered, noting their own Digital Ocean install works fine with IPv6 without custom Docker config. robbyoconnor clarifies they made no changes and asks if Falco modified any Docker files — Falco confirms a vanilla install and asks if the droplet has IPv6 enabled. robbyoconnor reiterates that IPv6 was working and then failed abruptly, and rebuilding the container didn’t help. Frustrated, they post again saying they can’t resolve the issue. Falco suggests debugging by running another service on an IPv6 port to test external connectivity — robbyoconnor confirms OpenSSH is successfully listening on IPv6. Finally, downey recommends using netstat and lsof to inspect whether Nginx or any other process is actually binding to IPv6 on ports 80/443, linking to a Linux guide for reference. The issue remains unresolved, with suspicion falling on infrastructure or networking rather than Discourse config.
I am not sure if this is a bug or if I did something, I don’t think I did and I checked the templates and they are configured to listen on the ipv6 address. This has been happening since January 7 at 4:51 AM UTC according to our monitoring. IPv4 access is still okay.