Using AWS S3 Object Lifecycle, how can I remove deleted objects

Question

Is there a way I can use s3's object lifecycle to prune through current versioned objects and delete those objects premantely that have a deleted marker on them in lets say a month or a week? If there isn't, how is house cleaning performed on versioned buckets?

Answer 1

This page describes object lifecycle management generically. This page describes lifecycle configuration more specifically.

You cannot do precisely what you want with S3's built-in lifecycle. With the versioning feature, there are two types of objects: the current version, and old non-current versions. The current version is the latest, most recently uploaded version of your object. For now, if the latest thing you've done to an object is cover it with a delete-marker, treat the delete-marker as the current version.

With S3's lifecycle, you can set up a rule to permanently delete non-current versions after X days. (You can also set up rules to move to glacier after X days.) But it sounds like what you're asking is for a rule where you can permanently delete non-current-but-only-if-the-current-is-a-delete-marker. This isn't natively supported.

If you want to only delete objects whose current version is a delete marker, you'll have to write your own listing agent to walk your bucket, enumerate these objects, and delete them yourself. There might be existing tools for this already; I haven't checked.