Sandbox Starter Tool



After making the unfortunate error of of terminating my 'Sandbox Starter' AMIs via the AWS console, instead of using the Sandbox 'Destroy' menu option, I have just spent some considerable time uninstalling my Desktop Docker, deleting Aviatrix policies, deleting/generating keys, etc in my efforts to get the  Sandbox Terraform script to successfully complete it's setup of the Sandbox again.
It has proven to be a painful learning and time consuming process.

So - I have some questions:

1 -  In the the event that I mistakenly terminated a SandBox AMI via the AWS console, is there a script/command available that can be used to rebuild that particular AMI again?

2 - In the event that no such 'rebuild' script/command exists, will using the 'Destroy' option remove all remaining elements of the existing Sandbox, so that a new Sandbox can be successfully installed by running 'Sandbox Starter' again? - or is it not that simple?

3 - Does creation of a new Sandbox also require creation of new a access and secret key? (I read something in the Starter notes that seem to indicate this) or can the existing ones be reused?




Best answer by Jeffrey_Hazel 5 May 2021, 21:27

View original

4 replies

Userlevel 2
Badge +1

I'll try to answer your questions based on my interpretation of the state you're currently in. And, to clarify terminology, when I refer to sandbox-starter (or starter) I'm referring to the container running locally on your desktop. Any ami is either the Aviatrix controller or gateways deployed to Aws or Azure.

  1. Assuming you're talking about controller and gateway instances here. Once, launched by the starter there's no way to re-invoke its code piecemeal.

  2. Sandbox-starter uses terraform to provision the controller and gateways in Aws and (optionally) Azure. Since you've removed docker from your system I presume you've reinstalled it and are running an entirely new instance of the starter. In this case the terraform state is blank and from its perspective there's nothing to destroy. You could, however, start over again and have the starter deploy new infrastructure. However, you'll need to remove the two roles and policies the starter deploys or I'd expect it to conflict.

  3. No. You're access and secret key are just your credentials for aws and can be reused for subsequent executions of the starter.

Hope this helps!


Yes - it does indeed help.

I did remove and re-install Docker and then started over again - only for the starter to keep failing. So I  leant the 'hard way' that I needed to remove the the 2 roles and policies!

Thanks for your assistance.

Jeffrey Hazel Are you saying after you removed the Aviatrix roles and policies in AWS, the Sandbox Starter worked properly in Docker? So scenario would be:

1. SS failing after manual deletion from AWS

2. Delete Aviatrix roles and policies in AWS

3. Re-run SS


Joel West  Not quite -  after looking more closely at the SS debug messages (following SS still failing after I re-installed Docker), I deleted the Aviatrix roles and policies and then did an SS 'Destroy'. 

SS then completed it's process without failing.

Moral of the story - Avoid using  AWS console to terminate the Sandbox AMIs.  🙂

Thanks for the info.