A forum experiencing issues with a small percentage of users (about 1%) encountering Chrome: ERR_SSL_PROTOCOL_ERROR and FF: SSL_ERROR_RX_RECORD_TOO_LONG errors. After a recent upgrade, affected users from different ISPs reported the issue despite clearing DNS cache and changing browsers. The forum administrator sixty4bit suspects an issue between the users’ ISPs and the server. Troubleshooting suggestions from pfaffman included checking the SSL certificate and server logs, but the issue seems to be related to the users’ network configurations. A temporary fix was found by rebooting home routers, and a message was posted on the main site advising users to try this solution. The server is hosted on Digital Ocean.
The forum has been alive for a couple of years with no issues. We get about 2k pageviews per day from about 100 unique users.
After an upgrade a couple of weeks ago, I started to get emails from users saying they can’t access the site. I have many users that are still able to access the site. The users with the issues are coming from different ISPs. They’ve cleared DNS cache. They’ve changed browsers.
What questions should I be asking to help troubleshoot the issue?
Is it a standard install or do you have a reverse proxy in front?
What’s the url so we can look at the certificate?
What is the the logs (/var/discourse/shared/standalone/logs/rails/production.log and the nginx log, which you’ll have to search for)
Did you do a rebuild or upgrade from the web interface? You should do a rebuild.
Standard install. No reverse proxy.
We are keeping the forums private, I’ll PM you the URL if you like, but I’d rather not have it on the open Internet. What part of the certificate is of interest?
It is Let’s Encrypt that is valid between Jul 29, 2021 and Oct 27, 2021.
Key size is 256bits
DNS name matches
The last time I upgraded I did a rebuild. Upgraded the server’s OS and everything.
A couple of users have relayed that they can’t get to the forum through their home routers, and when they try on their phones they get nothing as well. I asked them to turn wifi off on their phones and they suddenly have access. I then asked them to reboot their routers and that fixed the issue for their PCs.
What would the ISPs have pushed to the routers that is causing this issue?
Then it’s an issue between their ISP and your server. Whether the problem is with your server’s networking or their ISPs is complicated to discern and unless you can convince whoever’s network is broken to change something there is literally nothing you can do other than move your server to a different provider or network.
Who provides your server?
Digital Ocean. Again, this is only affecting like 1% of my users. It is very random. Some are frequent visitors, some rarely visit.
I’ve just posted a “reboot your router” message on the main site. Hopefully that will fix everyone’s issue.
Thank you for your help!