Why is it required to provide a s3, azure blob or github blob storage for migrations? (bbs2gh)

[ryaugusta]

I'm working on a migration where we are already exporting the .tar file from the BitBucket repo via the API. I want to use this repo tar file using gh bbs2gh migrate-repo --archive-path .... I'm providing the tar file locally so that it can be uploaded during the migration step of the GEI. When I give the --archive-path I still run into an error saying I must specify external storage to continue.

Why do I need to provide a storage option? Can the GEI just use the local tar from the machine where the GEI is running without needing to upload to storage?

[dylan-smith]

The migration happens on the GitHub backend, this CLI is just a wrapper for various API calls. When the backend migration logic runs it needs access to the migration archive, and can’t reach out to your local machine to get it.

[ryaugusta]

Thanks @dylan-smith ; I assume then that the reason we need to provide BBS_USERNAME, BBS_PASSWORD and other bitbucket parameters after we already extract the tar is that the GEI will reach into bitbucket and extract the repository metadata info like PRs, Issues, Settings etc?

[dylan-smith]

No, you shouldn’t need to provide username and password if you provide the archive directly. In fact the CLI will error if you attempt to do so via args (I think).