Efficient garbage collection for stable data

A system comprising:

• a storage array comprising a plurality of storage volumes;
• and
• a storage controller coupled to the storage array, the storage controller comprising a processing device, the processing device to:
• perform a uniform sampling of data storage items in an append-only file system distributed across the plurality of storage volumes to identify a sample set;
• identify a plurality of data age clusters from the sample set;
• perform a garbage collection process on the append-only file system to identify a first data storage item in a first logical storage segment that is no longer in use, a second data storage item in the first logical storage segment that remains in use, and a third data storage item in the first logical storage segment that remains in use;
• write the second data storage item from the first logical storage segment to a second logical storage segment based on a data age of the second data storage item, wherein the second logical storage segment is associated with a first data age cluster of the plurality of data age clusters;
• and
• write the third data storage item from the first logical storage segment to a third logical storage segment based on an age of the third data storage item, wherein the third logical storage segment is associated with a second data age cluster of the plurality of data age clusters.