Hello.
I’ve tried to follow instructions on OpenviduPro AWS deployment several times with different parameters, but AWS stack creation process never completes.
I’ve also tried to disable rollback as suggested in the troubleshooting section of OpenviduPro AWS deployment. But I still can’t access my server using ssh in order to collect log details from failed stack creation.
Can someone, please help me?
Did you followed this guide? https://docs.openvidu.io/en/2.16.0/openvidu-pro/deployment/aws/
Also be sure of checking some SSL examples:
The stack works perfectly it should be something in the parameters. Can you show me what parameters are you using?
yes, that guide. Here are the parameters:
| AwsInstanceTypeKMS | c5.xlarge | - |
|---|---|---|
| AwsInstanceTypeOV | c5.large | - |
| ElasticsearchPassword | **** | - |
| ElasticsearchUser | elasticadmin | - |
| KeyName | kviz-nemanjici-openvidu | - |
| LetsEncryptEmail | organizator@nemanjici.net | - |
| MediaNodesStartNumber | 1 | - |
| MyDomainName | video.nemanjici.net | - |
| OpenViduLicense | **** | - |
| OpenViduSecret | **** | - |
| OpenViduSubnet | subnet-9274e4f8 | - |
| OpenViduVPC | vpc-68f15902 | - |
| OwnCertCRT | - | - |
| OwnCertKEY | - | - |
| PublicElasticIP | 18.184.236.77 | - |
| Recording | s3 | - |
| S3RecordingsBucketName | kviz-nemanjici-s3 | - |
| WantToDeployDemos | true | - |
| WhichCert | letsencrypt | - |
Hello @Vitalije_Milosevic
I checked with c5.large, and actually it is not working. I’ve researched what the problem could be and now I know it:
We normally test or environment with c5.xlarge which has 8GB of RAM because ELK takes a lot of RAM to works properly. c5.large has 4GB of RAM which is not enough, in fact, the minimum requirements of OpenVidu to run properly are 2 cpus and 8GB of RAM: https://docs.openvidu.io/en/2.16.0/openvidu-pro/deployment/on-premises/#1-prerequisites.
You have 2 options to solve your problem:
Option 1:
Run your OpenVidu Instance with a t3.large, which is cheaper than c5.large and have 2 cpus and 8GB of RAM. Don’t use t2.large because it has some network restrictions that could affect OpenVidu.
Option 2:
Deploy Elasticsearch and Kibana outside of OpenVidu before deploying. You can check this section which guide you on how to do it: https://docs.openvidu.io/en/2.16.0/openvidu-pro/monitoring-elastic-stack/#configuring-an-external-elastic-stack
Thanks for reporting the problem.
PD: As an additional note, if you deployed a lot of times using the domain video.nemanjici.net and Letsencrypt option, maybe because of a lot of failed attempts to deploy the domain is blocked by Letsencrypt because of a lot of requests to generate the certificates. I recommend you to configure a new subdomain and follow Option 1 or Option 2.
I hope this helps,
Best Regards,
Carlos
Thanks for the detailed explanation. I think I am going to use c5.xlarge and a new subdomain.
It would be good to mention this in the deployment tutorial.
Vitalije
