-
Notifications
You must be signed in to change notification settings - Fork 3k
Cannot generate SSL certificate #881
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
I've also started getting the Internal Error messages when creating new Proxy Hosts. I'm using Cloudflare and follow the same procedure as always when issuing. It might be related to the new 2.8.0 version as I did not have the issues before that. I tried reverting to 2.7.3 but now experience the same error on this version too. I also tried removing the XX.conf files associated with old entries, but it seems something old is trailing too (database seems empty when checking with adminer?) |
Same issue for me with the 2.8.0 or 2.7.3 version in SQLite mode running into docker container. I have this king of log:
|
Update from my end – it seems to be automatically resolved. I went back to use the :latest image (2.8.0) and booted up NPM and now it seems all is working again. Issuing certs and attaching existing SSL certs to new hosts all seem to work. The issues I had removing proxy hosts is also gone. Phew. Hope it works for you guys too/ |
Same for me, yesterday I tryed to renew all my certificates and all was done with success... |
I'm not having much luck:
and log shows:
Any ideas? edit: more logs at startup of container:
|
I am having the same issue. I can't even get to the HTTP request at all. I uninstalled and reinstalled many time: |
Bump? I've temporarily moved to the addon in home assistant which seems to be working fine for me (but now I'm at my limit for renews, so i have to wait a week). |
I'm having the same issue, kept thinking it's me and how i'm registering via Cloudflare but i'm thinking something' up with Cloudflare. I'm on Unraid with docker version:
Namecheap is where i have my DNS hosted and use Cloudflare to proxy. CF uses the zone edit API now instead of global api to do the acme-challenge. The last time I did this, it was with the global api and worked fine. Made sure this time around after I figured it out to use the Zone API. We can see the acme-challenge with the TXT records being temporarily created and then removed but for some reason, it still failed.
|
This feels like it is some kind of race condition. So I was messing with this in a rather slow staging VM running in 1 CPU core on a 2014 macbook air, and was unable to generate any SSL certificates UNLESS after every internal server error I just immediately spammed a retry. I think what is happening is that the web server that hosts the challenge file that letsencrypt is looking for is not coming up fast enough for the challenge. So what the NGINX Proxy Manager is doing in the background is it is taking down whatever it is that you have occupying port 80 at that domain name, replacing the config with a new server that will host the challenge file at port https://domain.com/.well-known/acme-challenge/ Before this host is actually up, running, and reachable, it is telling the certbot to go ahead and continue the challenge request, and then of course since it can't reach that destination the cert process fails. You can see the residual effect of this when it is done, it takes about a minute after a cert request for the Congratulations! default landing page for nginx proxy to come back online on my test setup. Basically certbot is beating the webservice coming up with the challenge file and winning the race condition to your failure. Hammering on renew before the webservice comes back to the congratulations page did it for me, it also fills up the letsencrypt folder with tons of garbage certs =/ This happens both with this image and the newest image from here: https://github.com/jlesage/docker-nginx-proxy-manager/compare. I Don't know if this setup is using the nginx instance to host the challenge file, or using the auto host built into certbot, but either port 80 isn't unbinding fast enough or the switch to the certbot challenge file host isn't happening quick enough. |
This is an issue in certbot nginx, see: certbot/certbot#8163 they implemented a feature to wait for nginx to start back up after changing it's configuration, it's a hard timer. certbot now has a command line parameter --nginx-sleep-seconds for doing it's job on slower machines where nginx may not start up fast enough. I found this as I was trying to use letsencrypt with a plain nginx system after I could not get a certificate using nginxproxymanager. |
Any idea when this fix will be implemented in Proxy Manager? |
From the documentation, From the code path I can't see how the reloading of nginx wouldn't be completed prior to requesting a cert via certbot however an additional check for nginx being up could be added rather easily. |
Hi, I have the same error: is there a solution in the meantime? |
As failing to create a certificate and the "internal error" is a very generic error, this issue has ended up containing a huge mix off entirely different issues, which are not connected. Hence I will go ahead and close this issue, to prevent it from becoming a graveyard for different problems. You can go ahead an open a new issue with specifics to your issue. #1271 Will help you get started. |
I managed to configure NPM as a reverse proxy for 2 internal services, but i don't know why i cant generate the certificate. It gives no real error i could work with...
I'am using the DNS from namecheap that is not in the list, is it necessary to get a certificate?
The text was updated successfully, but these errors were encountered: