Patent 10540095 was granted and assigned to Pure Storage on January, 2020 by the United States Patent and Trademark Office.
A storage controller coupled to a storage array comprising one or more storage devices performs a uniform sampling of data storage items in an append-only file system to identify a sample set and identifies at least two data storage item groups based on an age of the data storage items. The storage controller performs a garbage collection process on the append-only file system to identify stale data storage items in a first logical storage segment and active data storage items in the first logical storage segment and writes a first active data storage item associated with a first group of the two data storage item groups from the first logical storage segment to a second logical storage segment, the second logic storage segment comprising other active data storage items associated with the first group.