Redundancy recovery within a distributed data-storage system

• US 20080126842A1
• Filed: 09/27/2006
• Published: 05/29/2008
• Est. Priority Date: 09/27/2006
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Jacobson, Michael B., Jackson, James P.
• Granted Patent: US 7,743,276 B2
• Time in Patent Office: Days
• US Class Current: 714/6
• CPC Class Codes: G06F 11/1662 the resynchronized component or unit being a persistent storage device re-synchronization of failed mirror storage G06F11/2082; rebuild or reconstruction of parity RAID storage G06F11/1008
• CPC Class Codes: G06F 11/18 using passive fault-masking of the redundant circuits error detection by comparing the output of redundant processing systems with continued operation after detection of the error G06F11/165
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence

Abstract

Embodiments of the present invention are directed to methods, and distributed data-storage systems employing the methods, for recovering redundancy within a distributed data-storage system upon failure of one or more mass-storage devices within a component data-storage system of the distributed data-storage system. In certain embodiments, failure of a mass-storage device within a component data-storage system elicits a redundancy-recovery operation in which segments affected by the mass-storage-device failure or failures are moved, by a process referred to as “migration,” to other component data-storage systems of the distributed data-storage system, and are recovered as a by-product of migration. Certain embodiments of the present invention more efficiently address redundancy recovery by moving only as many segments from the component data-storage system as needed to provide sufficient free space within the component data-storage system to recover the remaining segments affected by the mass-storage-device failure or failures within the component data-storage system.

Claims

1. A method for restoring redundancy, lost due to a mass-storage-device failure or failures within a component data-storage system of a distributed data-storage system comprising a number of component data-storage systems, each component data-storage system containing a number of mass-storage devices, the method comprising:

when sufficient free data-storage space is available within the component data-storage system in which the mass-storage-device failure or failures occurred, restoring data-storage units of segments and/or segment portions that were stored on the failed mass-storage device or devices;

when sufficient free data-storage space can be made available within the component data-storage system in which the mass-storage-device failure or failures occurred by moving one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems of the distributed data-storage system, moving the one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems and then restoring data-storage units of remaining segments and/or segment portions that were stored on the failed mass-storage device or devices; and

otherwise, moving segments and/or segment portions that were stored on the failed mass-storage device or devices to other component data-storage systems of the distributed data-storage system.

2. The method of claim 1 wherein each component data-storage system maintains a hierarchical-system-configuration-describing data structure that maps segments of virtual disks to particular component data-storage systems according to a redundancy scheme.

3. The method of claim 2 wherein each component data-storage system maintains a segment-to-logical-data-storage-unit map that maps segments and/or segment portions stored within the component data-storage system to logical data-storage units.

4. The method of claim 3 wherein each component data-storage system maintains a logical-to-physical map that maps logical data-storage units to physical data-storage units within mass-storage devices of the component data-storage system.

5. The method of claim 4 wherein moving the one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems includes updating the hierarchical-system-configuration-describing data structure to indicate the new locations of the segments and/or segment portions and updating the segment-to-logical-data-storage-unit map to remove mappings for moved segments and/or segment portions.

6. The method of claim 4 wherein restoring data-storage units of remaining segments and/or segment portions that were stored on the failed mass-storage device or devices includes importing redundantly stored data-storage units from other component data-storage systems.

7. The method of claim 1 wherein the one or small number of segments and/or segment portions are selected for migration in order to optimize redundancy recovery with respect to a set of goals and constraints.

8. The method of claim 1 wherein goals may include:

minimizing data transfer between component data-storage systems;

minimizing time required to restore redundancy; and

maximizing robustness of the distributed data-storage system following redundancy restoration.

9. Computer instruction stored in a computer-readable memory that implements the method of claim 1.

10. A distributed data-storage system comprising:

a number of component data-storage systems, each component data-storage system containing a number of mass-storage devices;

one or more communications media that interconnect the number of component data-storage systems; and

an implementation, within each component data-storage system, of a method for restoring redundancy, lost due to a mass-storage-device failure or failures within the component data-storage system by;

when sufficient free data-storage space is available within the component data-storage system, restoring data-storage units of segments and/or segment portions that were stored on the failed mass-storage device or devices;

when sufficient free data-storage space can be made available within the component data-storage system by moving one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems of the distributed data-storage system, moving the one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems and then restoring data-storage units of remaining segments and/or segment portions that were stored on the failed mass-storage device or devices; and

otherwise, moving segments and/or segment portions that were stored on the failed mass-storage device or devices to other component data-storage systems of the distributed data-storage system.

11. The distributed data-storage system of claim 10 wherein each component data-storage system maintains a hierarchical-system-configuration-describing data structure that maps segments of virtual disks to particular component data-storage systems according to a redundancy scheme.

12. The distributed data-storage system of claim 11 wherein each component data-storage system maintains a segment-to-logical-data-storage-unit map that maps segments and/or segment portions stored within the component data-storage system to logical data-storage units.

13. The distributed data-storage system of claim 12 wherein each component data-storage system maintains a logical-to-physical map that maps logical data-storage units to physical data-storage units within mass-storage devices of the component data-storage system.

14. The distributed data-storage system of claim 13 wherein moving the one or small number of segments and/or segment portions from the component data-storage system to other component data-storage systems includes updating the hierarchical-system-configuration-describing data structure to indicate the new locations of the segments and/or segment portions and updating the segment-to-logical-data-storage-unit map to remove mappings for moved segments and/or segment portions.

15. The distributed data-storage system of claim 13 wherein restoring data-storage units of remaining segments and/or segment portions that were stored on the failed mass-storage device or devices includes importing redundantly stored data-storage units from other component data-storage systems.

16. The distributed data-storage system of claim 10 wherein the one or small number of segments and/or segment portions are selected for migration in order to optimize redundancy recovery with respect to a set of goals and constraints.

17. The distributed data-storage system of claim 10 wherein goals may include:

minimizing data transfer between component data-storage systems;

minimizing time required to restore redundancy; and

maximizing robustness of the distributed data-storage system following redundancy restoration.