S3 Deletions

Q: How does the system delete files from S3 endpoint?

A: It doesn't, intentionally, by design. It doesn't even use credentials capable of deletion. Test buckets are cleared out automatically by S3 policy. Production buckets shift objects to Glacier similarly, by policy (the metadata remains available via S3).

Q: Ok then, but how do I delete from S3?

A: Deleting requires credentials from Ops/Julian. You will be issued credentials for one S3 endpoint at a time; the previous credentials will be revoked when the new ones are issued.

Deletion can be performed using the prescat:purge_zips rake task using the supplied credentials.