# US 2019/0377640 A1 · Raid systems and methods for improved data recovery performance

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0377640 A1
- Document ID: US20190377640A1
- Status: Published application
- Kind: A1
- Application number: 16/546,807
- Filed: 2019-08-21
- Published: 2019-12-12
- Earliest claimed: 2017-01-19
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Daniel S. Critchley; Gordon D. Hutchison; Gareth P. Jones; Jonathan W. L. Short
- Web page: https://patentgrep.com/patent/US20190377640A1

## CPC

- G06F 11/1088: {Reconstruction on already foreseen single or plurality of spare disks}
- G06F 3/061: {Improving I/O performance}
- G06F 3/0619: {in relation to data integrity, e.g. data losses, bit errors}
- G06F 3/064: {Management of blocks}
- G06F 3/0689: {Disk arrays, e.g. RAID, JBOD}
- G06F 11/1076: {Parity data used in redundant arrays of independent storages, e.g. in RAID systems}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A RAID system, RAID controller, method, and computer program product for reducing the number of reads of XOR data in a multi-storage-enclosure RAID array includes a RAID array controller that implements a selected distributed RAID scheme. The RAID array controller determines a set of drives and logical block addresses corresponding to a parity group and divides the set of drives into subsets of drives that are located within each individual storage enclosure of the multiple storage enclosures. The controller issues a single EnclosureXOR Read to each storage enclosure corresponding to the subsets of drives to read enclosure-level intermediate XOR data calculated by each storage enclosure for each subset of drives and in response to receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, and calculates an array level XOR result by XORing the enclosure-level intermediate XOR data results from the storage enclosures.

## Claims

### Claim 1

An apparatus comprising:
a RAID array controller that distributes a stride of parity data for a parity group across a RAID array consistent with a selected distributed RAID scheme, the RAID array comprising a plurality of storage enclosures each having a plurality of storage drives within, wherein the RAID array controller:
determines a set of drives and logical block addresses corresponding to the parity group;
divides the set of drives into subsets of drives that are located within each individual storage enclosure of the plurality of storage enclosures;
issues a single enclosure-level XOR read (“EnclosureXOR Read”) request to each storage enclosure corresponding to the subsets of drives to read enclosure-level intermediate XOR data calculated by individual storage enclosures for each subset of drives;
sends a read request to read a single enclosure-level intermediate XOR data result from the individual storage enclosures, the enclosure-level intermediate XOR data results corresponding to XOR data for the subset of drives within the individual storage enclosures; and
in response to receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, calculates an array level XOR result by performing an XOR operation on the enclosure-level intermediate XOR data results from the individual storage enclosures.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein in response to not receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, the RAID array controller performs at least one retry attempt of the EnclosureXOR Read.

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein the RAID array controller sends the read request to each of the individual storage enclosures concurrently.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein the RAID array controller further writes the array level XOR result to a spare space on one of the plurality of storage drives consistent with the selected distributed RAID scheme.

### Claim 5 (depends on 1)

The apparatus of claim 1, wherein the distributed RAID scheme is selected from RAID5 and RAID6.

### Claim 6 (depends on 5)

The apparatus of claim 5, wherein the RAID array controller regenerates a data chunk corresponding to a lost data chunk using the array level XOR result and raw data read from the plurality of storage drives.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein the RAID array controller sends a write request to store the regenerated data chunk to a predetermined storage drive of the RAID array consistent with the selected distributed RAID scheme.

### Claim 8 (depends on 7)

The apparatus of claim 7, wherein the RAID array controller reads enclosure-level intermediate XOR data from multiple parity groups corresponding to multiple strides and reconstructs multiple stride redundancy groups concurrently.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the EnclosureXOR Read function is implemented as an extension to a SCSI attached enclosure services protocol.

### Claim 10

A method comprising:
distributing a stride of parity data for a parity group across a RAID array consistent with a selected distributed RAID scheme, the RAID array comprising a plurality of storage enclosures each having a plurality of storage drives within;
determining a set of drives and logical block addresses corresponding to the parity group;
dividing the set of drives into subsets of drives that are located within each individual storage enclosure of the plurality of storage enclosures;
issuing a single enclosure-level XOR read (“EnclosureXOR Read”) request to each individual storage enclosure corresponding to the subsets of drives to read enclosure-level intermediate XOR data calculated by each storage enclosure for each subset of drives;
sending a read request to read a single enclosure-level intermediate XOR data result from the individual storage enclosures, the enclosure-level intermediate XOR data results corresponding to XOR data for the subset of drives within the individual storage enclosures; and
in response to receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, calculating an array level XOR result by performing an XOR operation on the enclosure-level intermediate XOR data results from the individual storage enclosures.

### Claim 11 (depends on 10)

The method of claim 10, wherein in response to not receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, performing at least one retry attempt of the EnclosureXOR Read.

### Claim 12 (depends on 10)

The method of claim 10, further comprising sending the read request to each of the individual storage enclosures concurrently.

### Claim 13 (depends on 10)

The method of claim 10, further comprising writing the array level XOR result to a spare space on one of the plurality of storage drives consistent with the selected distributed RAID scheme.

### Claim 14 (depends on 10)

The method of claim 10, wherein the distributed RAID scheme is selected from RAID5 and RAID6.

### Claim 15 (depends on 14)

The method of claim 14, further comprising regenerating data corresponding to a lost data chunk using the array level XOR result and raw data read from the plurality of storage drives.

### Claim 16 (depends on 15)

The method of claim 15, further comprising a write request to store the regenerated data chunk to a predetermined storage drive of the RAID array consistent with the selected distributed RAID scheme.

### Claim 17 (depends on 16)

The method of claim 16, further comprising reading enclosure-level intermediate XOR data from multiple parity groups corresponding to multiple strides and reconstructs multiple stride redundancy groups concurrently.

### Claim 18 (depends on 10)

The method of claim 10, wherein the EnclosureXOR Read function is implemented as an extension to a SCSI attached enclosure services protocol.

### Claim 19

A computer program product for reducing a number of reads of XOR data in a multi storage enclosure RAID array, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor of a RAID storage controller to cause the RAID storage controller to:
distribute a stride of parity data for a parity group across a RAID array consistent with a selected distributed RAID scheme, the RAID array comprising a plurality of storage enclosures each having a plurality of storage drives within;
determine a set of drives and logical block addresses corresponding to the parity group;
divide the set of drives into subsets of drives that are located within each individual storage enclosure of the plurality of storage enclosures;
issue a single enclosure-level XOR read (“EnclosureXOR Read”) request to each individual storage enclosure corresponding to the subsets of drives to read enclosure-level intermediate XOR data calculated by each storage enclosure for each subset of drives;
send a read request to read a single enclosure-level intermediate XOR data result from the individual storage enclosures, the enclosure-level intermediate XOR data results corresponding to XOR data for the subset of drives within the individual storage enclosures; and
in response to receiving the enclosure-level intermediate XOR data results from all storage drives in the parity group, calculate an array level XOR result by performing an XOR operation on the enclosure-level intermediate XOR data results from the individual storage enclosures.

### Claim 20 (depends on 19)

The computer program product of claim 19, wherein the program instructions are further executable by the processor of the RAID storage controller to cause the RAID storage controller to:
read enclosure-level intermediate XOR data from multiple parity groups corresponding to multiple strides and reconstruct multiple stride redundancy groups concurrently.
