Monday, October 12, 2015

[maskwpvj] Managing data quantity on versioned storage

Given a data storage system that stores versions of files (or generically objects), e.g., Git, these seem messy problems:

What things are taking up the most space?
Delete some things to save space.
Move some things to cheaper archival storage, keeping placeholders and metadata.
Restore some things from archival storage.

A "thing" might not a file, but a particular version of a file, perhaps a large insertion that was later deleted in a subsequent revision.

No comments :