| 摘要 |
In one embodiment, a layered file system of a storage input/output (I/O) stack executes on one or more nodes of a cluster. The layered file system includes a flash-optimized, log-structured layer configured to provide sequential storage of data and metadata (i.e., a log-structured layout) on solid state drives (SSDs) of storage arrays in the cluster to reduce write amplification, while leveraging a data de-duplication feature of the storage I/O stack. An extent store layer of the file system performs and maintains mappings of the extent keys to SSD storage locations, while a volume layer of the file system performs and maintains mappings of the LUN offset ranges to the extent keys. Separation of the mapping functions between the volume and extent store layers enables different volumes with different offset ranges to reference a same extent key (and thus a same extent). |
| 主权项 |
1. A method comprising:
receiving a first write request directed towards a first logical unit (LUN), the first write request having write data and having metadata that includes a first logical block address (LBA), the first write request processed at a node of a storage system, the storage system attached to a storage array of solid state drives (SSDs); applying a hash function to the write data to generate a first hash value; selecting an extent store from a plurality of extent stores based on the first hash value; storing the write data in a segment of the selected extent store, wherein the selected extent store spans a set of SSDs of the storage array, wherein a key is formed from the first hash value, storing the key in a first volume metadata structure associated with the first LUN; receiving a second write request directed towards a second LUN, the second write request having the write data and having metadata that includes a second LBA; applying the hash function to the write data to generate a second hash value, wherein the first hash value is equal to the second hash value; and storing the key associated with the write data in a second volume metadata structure associated with the second LUN, without writing the write data again to the storage array, to de-duplicate storage of the write data. |
