Hello
We've been following this article to setup the WCS servers on AWS behind a Load Balancer and enable Auto Scaling: AWS load balancer with auto scale quick setup
Following the steps above we managed to successfully setup an Application Load Balancer and have AWS deploy 2 WCS servers under it.
When testing we're able to log onto the WCS server via the Load Balancer DNS, and under Stream Recording can establish a WSS connection with the server.
However after establishing a connection it's unable to successfully start publishing the video stream, and within 20 seconds we receive the fail status "Failed by ICE timeout".
The WCS servers do not have public IPs as they should only sit behind the Load Balancer.
When we tried assigning public IPs to the instances and restart the WCS servers, we are able to successfully stream videos and play them back without issue, and even able to confirm when multiple streams were recording that they went to separate servers behind the Load Balancer. But as soon as the public IPs are removed and we restart the servers once more, we get the ICE timeout error.
It would be greatly appreciated if you can help us figure out why we are receiving the ICE timeout error.
On top of that there is also another question we would like to ask:
The steps we took with creating the Auto Scaling Group involved us making a custom AMI based on the Flashphoner AMI:
We double-checked AWS Costs and Usage and can confirm it is charging Web Call Server 5 correctly and proportionally to the number of scaled instances. It's just that the documentation above seemed to imply that if we used the Flashphoner AMI we needed to use that directly in the Launch Template instead of being able to make a custom image instead.
Cheers
We've been following this article to setup the WCS servers on AWS behind a Load Balancer and enable Auto Scaling: AWS load balancer with auto scale quick setup
Following the steps above we managed to successfully setup an Application Load Balancer and have AWS deploy 2 WCS servers under it.
When testing we're able to log onto the WCS server via the Load Balancer DNS, and under Stream Recording can establish a WSS connection with the server.
However after establishing a connection it's unable to successfully start publishing the video stream, and within 20 seconds we receive the fail status "Failed by ICE timeout".
The WCS servers do not have public IPs as they should only sit behind the Load Balancer.
When we tried assigning public IPs to the instances and restart the WCS servers, we are able to successfully stream videos and play them back without issue, and even able to confirm when multiple streams were recording that they went to separate servers behind the Load Balancer. But as soon as the public IPs are removed and we restart the servers once more, we get the ICE timeout error.
It would be greatly appreciated if you can help us figure out why we are receiving the ICE timeout error.
On top of that there is also another question we would like to ask:
The steps we took with creating the Auto Scaling Group involved us making a custom AMI based on the Flashphoner AMI:
- We launched an EC2 instance using the Flashphoner AMI.
- Updated the configuration and installed the necessary packages.
- Then created our own custom AMI from it to use in the AWS Launch Template.
We double-checked AWS Costs and Usage and can confirm it is charging Web Call Server 5 correctly and proportionally to the number of scaled instances. It's just that the documentation above seemed to imply that if we used the Flashphoner AMI we needed to use that directly in the Launch Template instead of being able to make a custom image instead.
Cheers
