Hello!
For development purposes we have created a separate server, where we launched Openvidu according to On Premises instructions: On premises - OpenVidu Docs
The problem is that when we are using it for a longer time, AWS EC2 container fully crashes and then we are unable to even connect to it via SSH anymore. It has all the symptoms of the memory leak, as we have to fully recreate the server and we already had this happening several times.
Could you explain what is happening here?
Hi, why are you launching an AWS deloyment using the on premises steps instead of the AWS deployment guide AWS - OpenVidu Docs?
We used it as a quick workaround, so our frontend guys could use Openvidu for development, but they won’t have to launch any of the server stuff themselves, so this seemed like a quick solution. But after a week or so running on t2.micro it crashes due to unknown reasons and we have to reboot whole environment.
For production environments we are definitely planning to use AWS deployment plan with hopes, that autoscaling will work properly there. But currently we are not sure why it crashes - is it because we are doing something wrong, or is this some kind of an expected behaviour in small instances? We want to build up some experience to know what happens, so we are prepared in case something happens in the future.
t2.micro is just too small. 1 CPU a 1GB of memory won’t do it. Try using at least a 2 CPU and 8GB memory machine.
So with the machine that small it is expected to clog up.
Of course. I have just added to the documentation of CE on premises deployment a new Prerequisite point with the minimum hardware requirements, as I realised that it was only being stated in the PRO on premises deployment documentation.