Arvados

Currently we explain the collection lifecycle which helps users understand the distinctions between live / deleted / trashed / scheduled-to-be-trashed collections.

Behind the scenes, there is also a lifecycle for the underlying data blocks stored on backend volumes. This is invisible/irrelevant to regular users, but admins need to understand it so they can:

• predict how much backend storage capacity they will need
• evaluate whether the usage level they're seeing is normal, or indicates a problem
• delete data to reduce backend usage
• detect/troubleshoot/recover from data loss due to backend failure or bugs