Open queglay opened 3 years ago
Running destroy
on a Vault cluster on a regular basis is not a normal use case, so I think the default as-is works fine. That said, a PR to allow providing an external S3 bucket is a reasonable alternative use case, and a PR to add that is welcome!
In using the S3 backend option, I found that it was problematic to create the S3 bucket as part of this module.
Once using a backend like S3, it becomes possible to destroy and deploy vault on a daily basis, but I don't want to destroy the bucket. Would you accept a PR to use an existing bucket instead of creating one as the default workflow?
I'm suggesting it as a default just because I think it defeats the purpose of using the S3 backend and putting it in this module, since a terraform destroy on this module would be just like having no S3 backend at all.
Alternatively if this current implementation does have value the way it is for reasons I haven't considered... then could we use an optional var like 'use_existing_bucket=true' to utilise a data module to aquire an existing bucket instead of creating one?