# Using custom domain on AWS Ec2 **URL:** \ **Category:** Issues developing apps\ **Tags:** v2\ **Created:** [May 14, 2020, 12:26pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790 "2020-05-14T12:26:51Z")\ **Posts on this page:** 8\ **Page:** 1
**Author:** ![netamity](https://avatars.discourse-cdn.com/v4/letter/n/eb8c5e/32.png) [@netamity](https://openvidu.discourse.group/u/netamity)\ **Post date:** [May 14, 2020, 12:26pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/1 "2020-05-14T12:26:51Z")
I spun up a server using the cloud formation approach. I’ve set up an external app and successfully connected to it through the dns host name of the EC2 instance and a self sign certificate. With Michael’s help, I’ve managed to disable the call app. Now I need to connect it to a custom domain and get a lets encrypt certificate. I have: 1. Set a cname for openvidu.fitzip.ca to point to [https://ec2-52-14-206-60.us-east-2.compute.amazonaws.com](https://ec2-52-14-206-60.us-east-2.compute.amazonaws.com) which is my EC2 instance 2. Set OPENVIDU\_DOMAIN\_OR\_PUBLIC\_IP=openvidu.fitzip.ca and CERTIFICATE\_TYPE=letsencrypt and [LETSENCRYPT\_EMAIL=andy@netamity.com](mailto:LETSENCRYPT_EMAIL=andy@netamity.com) in .env 3. Restarted openvidu using ./openvidu restart I can still access the dashboard if accept the security warning at [https://ec2-52-14-206-60.us-east-2.compute.amazonaws.com/dashboard/](https://ec2-52-14-206-60.us-east-2.compute.amazonaws.com/dashboard/) Trying to access [https://openvidu.fitzip.ca/dashboard/](https://openvidu.fitzip.ca/dashboard/) gives me a 310 error- too many redirects. I can see it comes up a SSL ok from a certificate i set up using AWS Certificate Manager (not lets encrypt- I haven’t received an email). I also can’t connect from my external app, getting ‘WebSocket connection to ‘wss://openvidu.fitzip.ca/openvidu’ failed: Error during WebSocket handshake: Unexpected response code: 301’ I did previously try adding a load balancer to the openvidu EC2 instance and pointing openvidu.fitzip.ca at the load balancer, and i was getting the same too many redirects error. If I can access openvidu through the raw EC2 public DNS (or IP) and I have the DNS for openvidu.fitzip.ca pointing at that address, why am I not getting the same result as using the EC2 public dns. And what might I have missed re the letsencrypt email. Nothing received. I don’t think this is the main issue, but its something else to overcome. Thank you for your assistance so far and you patience with this newby. ---
**Author:** ![Oscar.Soto](https://yyz2.discourse-cdn.com/free1/user_avatar/openvidu.discourse.group/oscar.soto/32/292_2.png) [@Oscar.Soto](https://openvidu.discourse.group/u/Oscar.Soto)\ **Post date:** [May 14, 2020, 12:47pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/2 "2020-05-14T12:47:01Z")
Hello @netamity , You don’t happen to be using Cloudflare DNS provider? This same problem was reported by another user with this provider and the solution can be seen here [AWS Cloudformation domain issue](https://openvidu.discourse.group/t/aws-cloudformation-domain-issue/563/4) If you’re not using Cloudflare, tell me and we’ll keep investigating. Regards ---
**Author:** ![netamity](https://avatars.discourse-cdn.com/v4/letter/n/eb8c5e/32.png) [@netamity](https://openvidu.discourse.group/u/netamity)\ **Post date:** [May 14, 2020, 1:01pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/3 "2020-05-14T13:01:44Z")
Hi Oscar, no, not using cloudflare. Just a EC2 instance with the domain openvidu.fitzip.ca pointed at it. The ‘too many redirects’ thing could be an attempt to switch between http and https though, that would explain a loop, but not sure why and how to stop it. Your help is appreciated! ---
**Author:** ![Oscar.Soto](https://yyz2.discourse-cdn.com/free1/user_avatar/openvidu.discourse.group/oscar.soto/32/292_2.png) [@Oscar.Soto](https://openvidu.discourse.group/u/Oscar.Soto)\ **Post date:** [May 14, 2020, 1:14pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/4 "2020-05-14T13:14:45Z")
Hi, @netamity, What version of Openvidu are you using? and what type of deployment On Premise or Cloud Formation? Which dns provider provides you with the domain openvidu.fitzip.ca? Regards ---
**Author:** ![netamity](https://avatars.discourse-cdn.com/v4/letter/n/eb8c5e/32.png) [@netamity](https://openvidu.discourse.group/u/netamity)\ **Post date:** [May 14, 2020, 2:23pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/5 "2020-05-14T14:23:36Z")
It was spun up with the cloud formation script (with AWS) - version 2.13 . The nameservers for fitzip.ca are ns1.justlink.info. I think it’s a US/Canada based web host. I set a cname record like so: [openvidu.fitzip.co](http://openvidu.fitzip.co)[ec2-52-14-206-60.us-east-2.compute.amazonaws.com](http://ec2-52-14-206-60.us-east-2.compute.amazonaws.com) and two other records are now seen in whois for openvidu.fitzip.ca which are different to the root fitzip.ca records openvidu.fitzip.ca SOA [dns-external-master.amazon.com](http://dns-external-master.amazon.com) [root@amazon.com](mailto:root@amazon.com) 49158 28800 900 604800 900 and openvidu.fitzip.ca A 52.14.206.60 which both seemed to auto generate The A record is indeed the IP address of the EC2 instance The SOA record was also generated when I added the cname record (or perhaps when I changed it from the load balancer dns previously Hope this all makes more sense to you then me 🙂 nb. The SSL padlock I get when I go to [https://openvidu.fitzip.ca](https://openvidu.fitzip.ca) is issued by amazon, so it looks like its the one issued when I set up the load balancer. The certificate for [https://fitzip.ca](https://fitzip.ca) (where the external app is hosted) is issued by another provider. I have set .env in openvidu to use letsencrpt. Could this suggest the issue? ---
**Author:** ![Oscar.Soto](https://yyz2.discourse-cdn.com/free1/user_avatar/openvidu.discourse.group/oscar.soto/32/292_2.png) [@Oscar.Soto](https://openvidu.discourse.group/u/Oscar.Soto)\ **Post date:** [May 14, 2020, 2:41pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/6 "2020-05-14T14:41:05Z")
Hello @netamity, What you are doing is not right because the IPs Amazon assigns are volatile and will end up changing. Also using a CNAME is not right for this and may be what is going into the redirection loop. Please try creating an elastic IP in amazon and assign it in the Cloud Formation when you generate the cluster. Also assign this elastic IP to your domain openvidu.fitzip.ca with a type A record. Regards ---
**Author:** ![netamity](https://avatars.discourse-cdn.com/v4/letter/n/eb8c5e/32.png) [@netamity](https://openvidu.discourse.group/u/netamity)\ **Post date:** [May 14, 2020, 2:48pm UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/7 "2020-05-14T14:48:26Z")
Thanks Oscar, I’ll give it a go. I have set up a few load balanced applications (not openvidu) using a cname pointing at the public DNS of the load balancer with success. But I’ll try your approach and reply back. Many thanks for your assistance. ---
**Author:** ![netamity](https://avatars.discourse-cdn.com/v4/letter/n/eb8c5e/32.png) [@netamity](https://openvidu.discourse.group/u/netamity)\ **Post date:** [May 15, 2020, 9:24am UTC](https://openvidu.discourse.group/t/using-custom-domain-on-aws-ec2/790/8 "2020-05-15T09:24:49Z")
Just to follow up. I used this approach and have it working. Thank you for your assistance.