aws-samples/aws-healthimaging-samples

How to restart TLM proxy

Closed this issue · 1 comments

terry-wilson commented

@yyao84 I have been playing around with the TLM proxy and for the most part it has been working great, however, I did run into a problem the other day where it seemed to randomly not respond to requests. Given that it had been working fine, I figured I would just try to restart it but was not exactly sure how I should do that. I played around with the services in the AWS console and eventually was able to get new instances running after trying a number of different things ... sure enough the TLM proxy started working fine again after this restart. However, assuming I am going to need to do this again at some point, I am hoping to understand the "proper" way to restart the TLM proxy?

yyao84 commented

Hi @terry-wilson - the easiest way is to force a new deployment, which starts up new containers, drains connections from the old containers and then shuts them down.

In the console:

  • Navigate to ECS (double check the region, it takes me to a random region half the time if I don't have a preferred one set in my settings)
  • Select the TileLevelMarkerProxyStack-ecscluster cluster
  • Select the TileLevelMarkerProxyStack-ecsfargateservice service
  • Select the Update service button
  • In the Deployment configuration section, check Force new deployment
  • Select Update
  • This takes you to the Deployments tab for the service, where it shows you the status of the new deployment

To use the CLI:
aws ecs update-service --cluster ${cluster_name} --service ${service_name} --force-new-deployment

I've also seen issues with the TLM proxy, and am working on an update for it to not use Elasticache but rather sit behind CloudFront.