System and method for ensuring write data integrity in a redundant array data storage system

• US 5,379,417 A
• Filed: 11/25/1991
• Issued: 01/03/1995
• Est. Priority Date: 11/25/1991
• Status: Expired due to Fees

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Tandem Computers Incorporated (HP Inc.)
• Inventors: Lui, Albert, Walker, Mark S.
• Primary Examiner(s): Baker, Stephen M.
• Application Number:
  US07/797,470
• Time in Patent Office: 1,135 Days
• Field of Search: 395/575, 371/7, 371/8.1, 371/8.2, 371/9.1, 371/10.1, 371/11.1, 371/11.3
• US Class Current: 713/300
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/1402 Saving, restoring, recovering or retrying
• CPC Class Codes: G06F 11/2015 Redundant power supplies power supply failure G06F1/30
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 11/2089 Redundant storage control functionality
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 2211/104 Metadata, i.e. metadata associated with RAID systems with parity
• CPC Class Codes: G11C 29/74 using duplex memories, i.e. using dual copies
• CPC Class Codes: Y10S 706/916 Electronic or computer, internal or network, circuit

Abstract

A system and method for ensuring the completion and integrity of data modification operations to a redundant array data storage system and for ensuring the integrity of redundancy values in such a system. The system combines four elements: reliable host processing provided by primary and secondary CPU'"'"'s with check pointing of operations, so that the secondary CPU can complete initiation of a data modification operation if the primary CPU fails; block-buffering of Write data in a host CPU adapter before transmission of the data to a RAID array controller, so that failure of the host CPU after buffering completes does not interfere with completion of the data modification operation; reliable array processing provided by primary and secondary array controllers with check pointing of operations, so that the secondary array controller can complete a data modification operation if the primary array controller fails; and battery-backup for the power supply to the array controller and the storage units comprising the storage array, or, in the alternative, non-volatile memory in the array controller sufficient to store all data modification operation command information for data records that have not been successfully written to the storage units, for use in subsequent retries.

Claims

1. A system for ensuring the completion and integrity of data modification operations to a redundant data storage array subsystem, comprising:

a. a first host processor configured to initiate data modification operations for data records and to generate host check point information about said data modification operations;

b. a second host processor, coupled to the first host processor, to receive the host check point information and assume control of initiated data modification operations if the first host processor fails;

c. first and second buffers, coupled to the first and second host processors, each buffer having sufficient capacity for receiving and storing entire generated data modification operations from the first or second host processor, whereby subsequent failure of either host processor does not affect the received data modification operations;

d. a redundant array of data storage devices coupled to both the first and second buffers for storing data records as indicated by said data modification operations originating from either host processor;

e. a first array processor, coupled to the first buffer and to the redundant array of data storage devices to receive data modification operations transmitted to the controller from the first buffer, the array processor being configured to control the completion of the data modification operations to the redundant array of data storage devices, and to generate array check point information about said data modification operations; and

f. a second array processor, coupled to the second buffer, to the redundant array of data storage devices, and to the first array processor, the second array processor being configured to receive data modification operations from the second buffer, the array check point information generated by the first array processor, and to assume control of the completion of the data modification operations to the redundant array of data storage devices if the first array processor fails.

2. The system of claim 1, further including a battery backup, coupled to the first and second array processors, and to the redundant array of data storage devices, for supplying electrical power to the first and second array processors and the redundant array of data storage devices if the primary power supply thereto fails.

3. The system of claim 1, wherein the redundant array of data storage devices is arranged in a RAID 3 configuration.

4. The system of claim 1, wherein the redundant array of data storage devices is arranged in a RAID 4 configuration.

5. The system of claim 1, wherein the redundant array of data storage devices is arranged in a RAID 5 configuration.

6. The system of claim 1, wherein each array processor further includes non-volatile storage for storing selected information about each data modification operation until such data modification operation is completed.

7. The system of claim 6, wherein the stored information is used to validate redundancy records, in the redundant array of data storage devices, for uncompleted data modification operations after a power loss to the redundant array of data storage devices.

8. The system of claim 7, wherein the stored information is communicated to one of the host processors, and wherein the receiving host processor includes means for re-initiating uncompleted data modification operations.

9. A method for ensuring the completion and integrity of data modification operations to a redundant data storage array subsystem, comprising the steps of:

a. providing a first host processor for initiating data modification operations for data records and for generating host check point information about said data modification operations;

b. providing a second host processor, coupled to the first host processor to receive the host check point information, and for assuming control of initiated data modification operations if the first host processor fails;

c. providing first and second buffers, coupled to the first and second host processors, to receive and store entire generated data modification operations from the first or second host processor, whereby subsequent failure of either host processor does not affect the received data modification operations;

d. providing a redundant array of data storage devices for storing data records as indicated by said data modification operations;

e. providing a first array processor, coupled to the first buffer and to the redundant array of data storage devices, for receiving data modification operations from the first buffer, for controlling the completion of the data modification operations to the redundant array of data storage devices, and for generating array check point information about said data modification operations; and

f. providing a second array processor, coupled to the second buffer, to the redundant array of data storage devices, and to the first array processor, for receiving data modification operations from the second buffer, for receiving the array check point information from the first array processor, and for assuming control of the completion of the data modification operations to the array of data storage devices if the first array processor fails.

10. The method of claim 9, further including the step of providing a battery backup, coupled to the first and second array processors, and to the redundant array of storage devices, for supplying electrical power to the first and second array processors and the redundant array of data storage devices if the primary power supply thereto fails.

11. The method of claim 9, further including the step of storing selected information about each data modification operation in a non-volatile storage component until such data modification operation is completed.

12. The method of claim 11, further including the step of using the stored information to validate redundancy records, in the redundant array of data storage devices, for uncompleted data modification operations after a power loss to the redundant array of data storage devices.

13. The method of claim 12, further including the step of communicating the stored information to one of the host processors, and using the stored information in such host processor to re-initiate uncompleted data modification operations.

14. A method for ensuring the completion and integrity of data modification operations to a redundant storage array subsystem, comprising the steps of:

a. initiating data modification operations for data records in a first host processor, and generating host check point information about said data modification operations;

b. receiving the host check point information in a second host processor, and assuming control by the second host processor of initiated data modification operations if the first host processor fails;

c. receiving and storing the entire data modification operations in first and second buffers, whereby subsequent failure of either host processor does not affect the received data modification operations;

d. receiving data modification operations from the first buffer in a first array processor, the first array processor controlling completion of the data modification operations to a redundant array of data storage devices and generating array check point information about said data modification operations; and

e. receiving the data modification operations from the second buffer and the array check point information from the first array processor in a second array processor, the second array processor assuming control of the completion of the data modification operations to the redundant array of data storage devices if the first array processor fails.

15. The method of claim 14, further including the step of supplying electrical power to the first and second array processors and the redundant array of data storage devices from a battery backup system if the primary power supply thereto fails.

16. The method of claim 14, wherein the first array processor controls completion of the data modification operations to a RAID 3 configuration of redundant data storage devices.

17. The method of claim 14, wherein the first array processor controls completion of the data modification operations to a RAID 4 configuration of redundant data storage devices.

18. The method of claim 14, wherein the first array processor controls completion of the data modification operations to a RAID 5 configuration of redundant data storage devices.

19. The method of claim 14, further including the step of storing selected information about each data modification operation in a non-volatile storage component until such data modification operation is completed.

20. The method of claim 19, further including the step of using the stored information to validate redundancy records, in the redundant array of data storage devices, for uncompleted data modification operations after a power loss to the redundant array of data storage devices.

21. The method of claim 20, further including the step of re-initiating uncompleted data modification operations indicated by the stored information.