Using multi-stream storage devices in log structured storage system

Patent metadata

Field	Value
Publication number	US-11954346-B2
Application number	US-202117238613-A
Country	US
Kind code	B2
Filing date	Apr 23, 2021
Priority date	Apr 23, 2021
Publication date	Apr 9, 2024
Grant date	Apr 9, 2024

Abstract

A method is provided for use in a storage processor, the method comprising: receiving a write request, the write request including a request to store user data in an array that includes a plurality of solid-state drives (SSD); executing the write request by: identifying metadata that is associated with the write request, and writing the user data and the metadata to different data streams that are opened on the plurality of SSDs; wherein writing the user data and the metadata to different data streams causes: (i) the user data to be stored in one or more first erase units of any of the plurality of SSDs, and (ii) the metadata to be stored in one or more second erase units of any of the plurality of SSDs, such that no part of the metadata is stored on any of the one or more first erase units, and no part of the user data is stored on any of the one or more second erase units.

First claim

The invention claimed is:

1. A method for use in a storage processor, the method comprising:
receiving, by the storage processor, a write request, the write request including a request to store user data in a storage array that includes a plurality of solid-state drives (SSDs), the storage array the storage processor being part of a storage system;
reserving, by using an allocation layer that is executed by the storage processor, one or more physical logical blocks (PLBs), each of the reserved PLBs being part of a respective one of a plurality of stripes;
executing the write request by: identifying metadata that is associated with the write request, and writing, by the storage processor, the user data and the metadata to different data streams that are opened on the plurality of SSDs, wherein writing the user data and the metadata to different data streams causes: (i) the user data to be stored in one or more first erase units that are part of any of the plurality of SSDs, and (ii) the metadata to be stored in one or more second erase units that are part of any of the plurality of SSDs, such that no part of the metadata is stored on any of the first erase units, and no part of the user data is stored on any of the second erase units, the first and second erase units being part of a plurality of erase units that are mapped to the reserved PLBs;
wherein the storage array is a Redundant Array of Independent Disks (RAID) array,
wherein the allocation layer is configured to allocate only PLBs that are part of active stripes, and
wherein a maximum number of stripes that are present in the in the storage system at any given time is determined in accordance with the equation of:
S m =(N drives ×N stripes )/N_raid_drives

wherein S m is the maximum number of stripes, N drives is a count of SSDs in the RAID array, and N_raid_drives is a count of drives in an uber of RAID array.

2. The method of claim 1 , further comprising performing, by the storage processor, garbage collection on the storage array, wherein performing the garbage collection includes: identifying any given one of the plurality of stripes that includes PLBs marked for deletion, copying any clean PLBs that are part of the given stripe to a different stripe, and deleting the given stripe.

3. The method of claim 1 further comprising relinquishing each of the one more PLBs, wherein:
the write request is received by a client that is executed by the storage processor;
each of the plurality of stripes is associated with a counter that is used to detect when the stripe has become full,
reserving the one or more PLBs includes transmitting from the client to an allocation layer a request to reserve the one or more PLBs and receiving, at the client, a response from the allocation layer indicating that the one or more PLBs have been reserved, and
relinquishing each of the one or more PLBs includes providing, by the client to the allocation layer, an incremented value of the counter for the stripe of which the PLB is part.

4. The method of claim 1 , wherein writing the user data and the metadata to different data streams that are opened on the plurality of SSDs includes:
identifying one or more first data streams and writing the user data to the one or more first data streams; and
identifying one or more second data streams, and writing the metadata to the one or more second data streams.

5. The method of claim 1 , wherein deleting any given stripe includes deleting the given stripe by issuing, to the SSD, a physical layer unmap command, the unmap command including one of a Small Computer System Interface (SCSI) command, an Advanced Technology Attachment (ATA) TRIM command, or a Non-Volatile Memory Express (NVMe) DEALLOCATE command.

6. A storage processor, comprising:
a memory; and
at least one processor that is operatively coupled to the memory, the at least one processor being configured to perform the operations of:
receiving a write request, the write request including a request to store user data in a storage array that includes a plurality of solid-state drives (SSDs), the storage array and the storage processor being part of a storage system;
reserving, by an allocation layer, one or more physical logical blocks (PLBs), each of the reserved PLBs being part of a respective one of a plurality of stripes; and
executing the write request by: identifying metadata that is associated with the write request, and writing the user data and the metadata to different data streams that are opened on the plurality of SSDs, wherein writing the user data and the metadata to different data streams causes: (i) the user data to be stored in one or more first erase units that are part of any of the plurality of SSDs, and (ii) the metadata to be stored in one or more second erase units that are part of any of the plurality of SSDs, such that no part of the metadata is stored on any of the first erase units, and no part of the user data is stored on any of the second erase units, the first and second erase units being part of a plurality of erase units that are mapped to the reserved PLBs, and
wherein the storage array is a Redundant Array of Independent Disks (RAID) array,
wherein the allocation layer is configured to allocate only PLBs that are part of active stripes, and
wherein a maximum number of stripes that are present in the in the storage system at any given time is determined in accordance with the equation of:
S m =(N drives ×N stripes )/N_raid_drives

where S m is the maximum number of stripes, N drives is a count of SSDs in the RAID array, and N_raid_drives is a count of drives in an uber of the RAID array.

7. The storage processor of claim 6 , wherein the at least one processor is further configured to perform garbage collection on the storage array, the performing of garbage collection including: identifying any given one of the plurality of stripes that includes PLBs marked for deletion, copying any clean PLBs that are part of the given stripe to a different stripe, and deleting the given stripe.

8. The storage processor of claim 6 , wherein:
the at least one processor is further configured to perform the operation of relinquishing each of the one more PLBs,
the write request is received by a client,
each of the plurality of stripes being associated with a counter that is used to detect when the stripe has become full,
reserving the one or more PLBs includes transmitting from the client to an allocation layer a request to reserve the one or PLBs and receiving, at the client, a response from the allocation layer indicating that the one or more PLBs have been reserved, and
relinquishing each of the one or more PLBs includes providing, by the client to the allocation layer, an incremented value of the counter for the stripe of which the PLB is part.

9. The storage processor of claim 6 , wherein writing the user data and the metadata to different data streams that are opened on the plurality of SSDs includes:
identifying one or more first data streams and writing the user data to the one or more first data streams; and
identifying one or more second data streams, and writing the metadata to the one or more second data streams.

10. The storage processor of claim 6 , wherein deleting any given stripe includes deleting the given stripe by issuing, to the SSD, a physical layer unmap command, the unmap command including one of a Small Computer System Interface (SCSI) command, an Advanced Technology Attachment (ATA) TRIM command, or a Non-Volatile Memory Express (NVMe) DEALLOCATE command.

11. A non-transitory computer-readable medium storing one or more processor-executable instructions, which, when executed by one or

Assignees

• Emc Ip Holding Co Llc

Inventors

• Alkalay Amitai
• Kamran Lior
• Morley Steven

Classifications

• G06F3/064 Primary

  Management of blocks · CPC title

• G06F3/0604

  Improving or facilitating administration, e.g. storage management · CPC title

• G06F3/0652

  Erasing, e.g. deleting, data cleaning, moving of data to a wastebasket · CPC title

• G06F3/0689

  Disk arrays, e.g. RAID, JBOD · CPC title

• G06F12/0253

  Garbage collection, i.e. reclamation of unreferenced memory · CPC title