How to keep persistent files across deployments?

tl;dr
I need to host large files that can’t be regenerated which each deployment.

Context
I’m working on an image recognition server that uses openCV. In order to match an image upload against a large image archive, I have feature files that need to be stored on the server. These files amount to multiple gigabytes and can’t be recreated with every deployment, nor can they be hosted elsewhere because of speed.

My questions:

  1. Does render allow me to keep certain files persistent? I didn’t find any documentation so I presume not. So;
  2. If I host my feature files on a private service instead, would they load as fast as they would if they were stored together with the rest of the code?
  3. Could there be other approaches to this problem I didn’t think of?

Hi Moenen,

It sounds like disks would be a fit to persist your archive here. However, please be sure to review the application considerations when using disks, for example, services with disks can only have a single instance currently and they also prevent zero downtime deploys.

Hope that helps

Alan

1 Like