Error from "control-tower info": bosh orchestrator down, unable to diagnose about control-tower HOT 3 CLOSED

After poking at things for ages, I noticed that the RDS psql instance had run out of disk space. I increased that in the AWS console and the system recovered.

So I have a fix. I would appreciate any comments on how I can get shell on the BOSH server or how to diagnose similar things in future, if anyone has any. I'll leave this issue open for a couple of days.

from control-tower.

It's a "feature" of BOSH that you can't SSH to the director by default. The way around it is to add a special user to the vm which you can then SSH to. The official docs on this can be found here.

For historical reasons that I can't remember we actually include this feature on GCP deployments but not on AWS. We can look at adding it to AWS as well.

I think the workaround is to add the ops in the GCP jumpbox-user.yml to the AWS custom-ops.yml then build a local version of Control Tower and deploy with it. Then you should be able to pull the private key from the config.json in S3 and ssh with ssh -i key.pem jumpbox@$BOSH_ENVIRONMENT.

from control-tower.

Thanks, I'll try that next time

from control-tower.