The user alxndr experienced issues with their Discourse site (lot.almost-dead.net) after restarting the VM on Google Cloud. The site was unreachable, with JS assets and admin pages failing to load, and Redis connection errors appearing in logs. Despite nginx and other services running, the site remained down. RGJ noted the site was completely unreachable, suggesting a network or nginx issue. The user checked firewall rules and found them correct, but later discovered the VM had a new IP address after restart. After updating the A record to point to the new IP, the site was restored. Benjamin_D and IAmGav offered troubleshooting suggestions, including checking SSL certificates and nginx logs, but the root cause was a changed IP address.
Background info: site is lot.almost-dead.net, version 2.8.0.beta4, hosted on Google Cloud/Compute; I followed the official docker-based guide to set it up. (I’m proficient with browser front-end tech and understand the broad strokes of cloud hosting, but am familiar with AWS and not Google.)
Preliminary cause: I stopped the VM instance and then restarted it (attempting to adjust the env vars exposed to Discourse).
When the VM came back and the site was back up, I noticed that JS assets were timing out, along with some of the user avatars. On the admin panel, the Community Health area doesn’t load; the spinner continues to spin, and in Chrome Dev Tools the Network tab shows the /message-bus/.../poll requests all failing. The Upgrade page (/admin/upgrade) fails almost immediately and Chrome shows the ERR_FAILED code. When browsing topics, I see POST requests failing with ERR_CONNECTION_REFUSED and JS-initiated GET requests failing with ERR_FAILED. (This is from a browser which has logged-in cookies set for my admin account.)
Loading the site from a fresh browser instance shows Chrome’s ERR_CONNECTION_REFUSED error.
What I’ve tried:
server-side rebuild — sudo ./launcher rebuild app seems to succeed, but there is no change in the site’s behavior
also tried commenting out plugins and rebuilding, still no change
No, your site is not back up, it’s completely down. You’re partly looking at cache. For someone who never visited your site, it doesn’t respond at all. This is likely something on the network/firewall level, or nginx is not starting / crashing.
Once I sudo ./launcher enter app it appears that nginx is running…
root@adn-prod-app:/var/www/discourse# ps aux | grep nginx
root 548 0.0 0.0 2156 64 ? Ss 07:04 0:00 runsv nginx
root 558 0.0 0.1 55236 2524 ? S 07:04 0:00 nginx: master process /usr/sbin/nginx
www-data 567 0.0 0.2 55996 5068 ? S 07:04 0:00 nginx: worker process
www-data 568 0.0 0.0 55996 1628 ? S 07:04 0:00 nginx: worker process
www-data 569 0.0 0.0 55792 1680 ? S 07:04 0:00 nginx: cache manager process
root 23179 0.0 0.0 6140 884 pts/1 S+ 21:23 0:00 grep nginx
I am not familiar enough with Google Cloud to know what to look for in their network/firewall settings… I do note that the VM Instance has the tags “http-server” and “https-server” and that their firewall system appears to use those tags to apply “default-allow-http” and “default-allow-https” built-in rules to instances with the tags. This sounds correct to me, and nslookup shows that the subdomain I’m using is resolving to the external IP address listed in Google’s UX, however the outside world still can’t access the instance.
Hi,
long shot, but last time I saw some redis error it was somehow related to (well let’s say in the same topic as) the ssl certificates (missing?).
Maybe ./launcher logs app could help ?