In a discussion titled ‘Fresh install issue’ under the Installation category, bondbenz reports being unable to access their Discourse site after a fresh Ubuntu VPS install using Cloud Document installation. Running discourse-doctor returns ‘NOT FOUND’ for both the domain and localhost. pfaffman asks if discourse-setup was run, if ./launcher rebuild app was executed multiple times, and confirms DNS and port settings. bondbenz confirms all checks passed — Ubuntu 20.04, DNS A records correct (subdomain), ports open, and no Cloudflare proxy (gray cloud). Despite this, the site remains offline. pfaffman suggests disabling Cloudflare proxy (orange cloud), but bondbenz confirms it was already off. The issue is resolved when bondbenz switches to a different subdomain. pfaffman theorizes that repeated rebuilds with broken DNS triggered Let’s Encrypt rate limiting. When bondbenz asks for a solution, pfaffman explains that adding a second subdomain to the Let’s Encrypt request could bypass the limit — though bondbenz already fixed the problem by changing subdomains.
Hello,
I have a fresh ubuntu vps , tried to installed using Cloud Document installation after finishing i can’t reach the website (Offline) , when i run discourse-doctor i get this :
Discourse version at community.xxxxxx.org: NOT FOUND
Discourse version at localhost: NOT FOUND
Thanks 
Did you run discourse-setup? Did it give any warnings when it ran?
Did you run
./launcher rebuild app
Many times out just one?
Are you sure that you have the ip address for your domain pointing at the vm?
What kind of vm?
Are you sure that ports are open (the setup script would check that)?
Yes everything is fine i tried to rebuild the app , no errors clocking!
Ubuntu 20.04 LTS Fresh install , i even reinstalled same thing, A records are fine in cloudflare domain is linked (subdomain) ports are open.
If you’re using the orange cloud you need to change it to gray cloud.
Oh yeah definetly I m using no proxied DNS (gray cloud), same problem too, it is so weird.
I have fixed it by changing the subdomain to another one.
My guess is that you ran enough rebuilds with your DNS broken that let’s encrypt rate limited you.
What is the solution for that in case?
Another solution if changing subdomains isn’t a possibility is to “cheat” by adding a second subdomain to the let’s encrypt request. But you’ve already solved your problem. I was just explaining why it worked.