Device array system providing redundancy of disks from active system disks during a disk failure

• US 6,079,029 A
• Filed: 11/26/1997
• Issued: 06/20/2000
• Est. Priority Date: 03/17/1997
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Iwatani, Sawao, Taketa, Suijin
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Elmore, Stephen C.
• Application Number:
  US08/979,951
• Time in Patent Office: 937 Days
• Field of Search: 714/6, 714/7, 714/710
• US Class Current: 714/6.12
• CPC Class Codes: G06F 11/1084 Degraded mode, e.g. caused by single or multiple storage removals or disk failures
• CPC Class Codes: G06F 11/1092 Rebuilding, e.g. when physically replacing a failing disk
• 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 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 2211/1004 Adaptive RAID, i.e. RAID system adapts to changing circumstances, e.g. RAID1 becomes RAID5 as disks fill up
• CPC Class Codes: G06F 2211/103 Hybrid, i.e. RAID systems with parity comprising a mix of RAID types

Abstract

In a device array system, a failure alternate device is fetched from another logic group of a higher redundancy without needing a hot spare patrol, accordingly recovering redundancy of data. In a case corresponding to a mirrored device of RAID-1, as a logic group of a device array unit, a construction control unit constructs a first logic group in which the same data is stored into (N=2) devices and duplicated and a second logic group in which the same data is stored into (N+1=3) devices and triplicated. when one device has failed in the first logic group having a lower redundancy and the redundancy is lost, a failure processing unit fetches one device from the second logic group of a higher redundancy into the first logic group and reconstructs data by using the fetched device as a failure alternate device, accordingly recovering the redundancy of the first logic group.

Claims

1. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports, said plurality of devices being organized into a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided in said device array unit into a plurality of logic groups having a combination of RAID levels and accessing the plurality of devices on said logic group unit and port number basis in response to an access request from the host computer, said controller comprising;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which same data is stored into (N) devices and multiplexed and a second logic group in which the same data is stored into at least (N+1) devices and multiplexed, and

a failure processing unit fetching one of the at least (N+1) devices in said second logic group into said first logic group if one of the (N) devices in said first logic group has failed and redundancy is lost, and reconstructing data stored in the failed device by using said fetched device as a failure alternate device, thereby recovering the redundancy.

2. A device array system according to claim 1, wherein:

as said logic groups, said construction control unit constructs a first logic group in which same data is stored into two devices and is duplexed and a second logic group in which same data is stored into three devices and triplicated; and

when one of the devices in said first logic group has failed and the redundancy is lost, said failure processing unit fetches one of the devices in said second logic group into said first logic group and reconstructs data stored in the failed device by using said fetched device as a failure alternate device, thereby recovering the redundancy.

3. A device array system according to claim 1, wherein when the failed device in said first logic group is exchanged for a normal device, said failure processing unit fetches the exchanged device into said first logic group, reconstructs the same data as that of said failure alternate device, and returns said failure alternate device into said second logic group, thereby reconstructing the original data.

4. A device array system according to claim 1, wherein when the failed device in said first logic group is exchanged for a normal device, said failure processing unit fetches the exchanged device into said second logic group, reconstructs the data, and leaves said failure alternate device as it is in said first logic group.

5. A device array system according to claim 1, wherein in the case where one device in said first logic group has failed and a redundancy is lost and the device in said second logic group is fetched into the first logic group and the data is reconstructed by using said fetched device as a failure alternate device and, subsequently, one device in said first logic group has failed, said failure processing unit fetches another device in the second logic group into the first logic group and reconstructs the data by using said fetched device as a second failure alternate device, thereby recovering the redundancy.

6. A device array system according to claim 1, wherein when the failure alternate device in said first logic group is fetched from said second logic group, said failure processing unit preferentially selects as the failure alternate device a device in the second logic group connected to the same port as that of said failed device.

7. A device array system according to claim 1, wherein in the case where the device in said second logic group has failed, said failure processing unit continues to execute the access request in a state in which said failed device is disconnected, and when said failed device is exchanged for a normal device, said failure processing unit returns said device to the second logic group and reconstructs the data therein.

8. A device array system according to claim 1, wherein said failure processing unit measures an error ratio of said plurality of devices provided for said device array unit to the accesses and, when said error ratio is equal to or larger than a predetermined value, said failure processing unit determines that the device has failed, and executes a corresponding process.

9. A device array system according to claim 1, wherein:

as said second logic group, said construction control unit constructs a plurality of second logic groups having different redundancies comprising (N+1) multiplexing and (N+2) multiplexing; and

when the device in said first logic group has failed and the redundancy is lost, said failure processing unit fetches the device in the group having a highest redundancy among said plurality of second logic groups into said first logic group and reconstructs the data by using said fetched device as a failure alternate device, thereby recovering the redundancy.

10. A device array system according to claim 1, wherein:

said device array unit comprises a magnetic disk drive and a semiconductor memory;

said construction control unit arranges at least one semiconductor memory into said plurality of devices in said second logic group; and

when the device in said first logic group has failed and the redundancy is lost, said failure processing unit assembles the semiconductor memory in said second logic group into said first logic group and reconstructs the data by using said assembled semiconductor memory as an alternate device of said failed device, thereby recovering the redundancy.

11. A device array system according to claim 1, wherein said device array unit comprises input/output devices comprising one of magnetic disk drives, magnetooptic disk drives, semiconductor memories, and magnetic tape apparatuses.

12. A device array system according to claim 1, wherein said device array unit comprises input/output devices comprising magnetic disk drive, magnetooptic disk drive, semiconductor memory, and magnetic tape apparatus.

13. A device array system comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing said plurality of devices on said logic group unit and port number basis in response to an access request from an upper apparatus, said controller comprising;

a construction control unit constructing during activation of the device array system the plurality of logic groups, each logic group comprising a predetermined redundancy, a data device storing data, and a Hamming code device storing a Hamming code corresponding to the data stored in said data device; and

a failure processing unit detecting when the device in a certain logic group has failed and a redundancy is below said predetermined redundancy, assembling a device for redundancy in another logic group into said logic group in which the device failure occurred under a condition such that the redundancy is not below said predetermined redundancy as a prerequisite, reconstructing the data by using said assembled device as a failure alternate device, thereby recovering the redundancy.

14. A device array system according to claim 13, wherein when said failed device is exchanged for a normal device, said failure processing unit reconstructs the same data as that of said failure alternate device into the exchanged device, returns said failure alternate device into the original logic group, and reconstructs the original Hamming code.

15. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing the plurality of devices on said logic group unit and port number basis in response to an access request from the host computer, said controller comprising;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which data is divided and stored into each of a plurality of data devices and parities obtained from said storage data are stored into a plurality of parity devices and a second logic group in which data is divided and stored into a plurality of data devices and parities obtained from said storage data are stored in one parity device; and

a failure processing unit detecting when the device in said second logic group has failed and a redundancy is lost, fetching one parity device in said first logic group into said second logic group, and reconstructing the data reconstructed by using said fetched parity device as a failure alternate device, thereby recovering the redundancy.

16. A device array system according to claim 15, wherein said construction control unit constructs a first logic group in which a parity device of RAID-3 is multiplexed and a second logic group of RAID-3.

17. A device array system according to claim 15, wherein said construction control unit constructs a first logic group in which a parity device of RAID-4 is multiplexed and a second logic group of RAID-4.

18. A device array system according to claim 15, wherein when the failed device in said second logic group is exchanged for a normal device, said failure processing unit reconstructs the same parity as that in said second logic group into said exchanged device, duplicates the parity device, and leaves the failure alternate device fetched from said first logic group into said second logic group as it is.

19. A device array system according to claim 15, wherein when the failed device in said second logic group is exchanged for a normal device, said failure processing unit reconstructs the same data as that in said failure alternate device into said exchanged device, returns said failure alternate device into the first logic group, and reconstructs the original parity.

20. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and a port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing the plurality of devices on said logic group unit and port number basis in response to an access request from said host computer, said controller comprising;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group having a plurality of data devices in which data is divided and stored, one parity device to store the same parity obtained from said storage data, and at least one data multiplexing device in which the same data as that of a part of said plurality of data devices is stored and multiplexed and a second logic group having a plurality of data devices in which data is divided and stored and one parity device for storing parities obtained from said storage data, and

a failure processing unit detecting when the device in said second logic group has failed and a redundancy is lost, and fetching one data multiplexing device in said first logic group into said second logic group and reconstructs the data by using said fetched data multiplexing device as a failure alternate device, thereby recovering the redundancy.

21. A device array system according to claim 20, wherein said construction control unit constructs a first logic group in which a part of the data devices of RAID-3 is multiplexed and a second logic group of RAID-4.

22. A device array system according to claim 20, wherein said construction control unit constructs a first logic group in which a part of the data devices of RAID-4 is multiplexed and a second logic group of RAID-4.

23. A device array system according to claim 20, wherein when the failed device in said second logic group is exchanged for a normal device, said failure processing unit reconstructs the same data as that in a specific data device in said second logic group into said exchanged device, duplicates a part of the data devices, and leaves the failure alternate device fetched from said first logic group into said second logic group as it is.

24. A device array system according to claim 20, wherein when the failed device in said second logic group is exchanged for a normal device, said failure processing unit reconstructs the same data as that in said failure alternate device into said exchanged device, returns said failure alternate device into the first logic group, and reconstructs the original data.

25. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing the pluralitv of devices on said logic group unit basis in response to an access request from an upper apparatus, said controller comprising;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group having a plurality of devices for storing data and a parity corresponding to RAID-5 and a multiplexing device to store the same data and parity as those of a part of said plurality of devices and a second logic group having a plurality of devices to store the data and a parity corresponding to RAID-5, and

a failure processing unit determining when the device in said second logic group has failed and a redundancy is lost, and fetching the multiplexing device in said first logic group into said second logic group, and reconstructing the data by using said fetched multiplexing device as a failure alternate device, thereby recovering the redundancy.

26. A device array system according to claim 25, wherein when the failed device in said second logic group is exchanged for a normal device, said failure processing unit reconstructs the same data as that in said failure alternate device into said exchanged device, returns said failure alternate device into the first logic group, and reconstructs the original data.

27. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and for accessing said plurality of devices on said logic group unit basis in response to an access request from the host computer, wherein said controller comprises;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which data and a Hamming code corresponding to RAID-2 are stored in a plurality of devices and a second logic group in which data and a parity corresponding to RAID-1, 3, 4, or 5, and

a failure processing unit determining when the device in said second logic group has failed and a redundancy is lost, and fetching the Hamming code device in said first logic group into said second logic group and, reconstructing the data by using said fetched Hamming code device as a failure alternate device, thereby recovering the redundancy.

28. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number,

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing said plurality of devices on said logic group unit and port number basis in response to an access request from the host computer, wherein said controller comprises;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which data and a Hamming code corresponding to RAID-2 are stored into a plurality of devices and a second logic group in which data and a parity corresponding one of RAID-1, RAID-3, RAID-4, and RAID-5 are stored in a plurality of devices, and

a failure processing unit determining when the device in said second logic group has failed and a redundancy is lost, changing said first logic group to a logic group of RAID-3, fetching one of a plurality of Hamming code devices released from the first logic group into said second logic group, and reconstructing the data by using said fetched Hamming code device as a failure alternate device, thereby recovering the redundancy.

29. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices can be accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing said plurality of devices on said logic group unit and port number basis in response to an access request from said host computer, wherein said controller comprises;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which same data is stored into a plurality of devices and a second logic group in which data and a parity are stored into a plurality of devices, and

a failure processing unit determining the device in said first logic group has failed and a redundancy is lost, fetching a parity device in said second logic group into said first logic group, and reconstructing the data by using said fetched parity device as a failure alternate device, thereby recovering the redundancy.

30. A device array system according to claim 29, wherein said construction control unit constructs, as said first logic group, a duplicating logic group in which same data is stored into two devices and constructs, as said second logic group, a logic group of RAID-3 or 4 in which a parity device is fixed.

31. A device array system coupled to a host computer and comprising:

a device array unit comprising;

a plurality of ports, each of said ports being assigned a port number, and

a plurality of devices to input and output connected to each of the plurality of ports over a plurality of ranks and said plurality of devices being accessed in parallel on a rank unit and port number basis; and

a controller dividing said plurality of devices provided for said device array unit into a plurality of logic groups having a combination of RAID levels and accessing said plurality of devices on said logic group unit basis in response to an access request from the host computer, wherein said controller comprises;

a construction control unit constructing during activation of said device array system, as said logic groups, a first logic group in which same data is stored into a plurality of devices and a second logic group in which data and a parity are stored into a plurality of devices, and

a failure processing unit determining when the device in said second logic group has failed and a redundancy is lost, fetching one device in said first logic group into said second logic group, and reconstructing the data by using said fetched device as a failure alternate device, thereby recovering the redundancy.

32. A device array system according to claim 31, wherein said construction control unit constructs, as said first logic group, a multiplexing logic group in which same data is stored into at least three devices and constructs, as said second logic group, a logic group of one of RAID-3, RAID-4, and RAID-5.

33. A disk array system, comprising:

a disk array controller; and

disk drives, coupled to the disk array controller, and storing data, said disk drives corresponding to each other in groups and organized into ranks and combinations of RAID levels upon activation of the disk array system, each of said disk drives being grouped into one of the groups, each one of the groups comprising redundant disk drives comprising at least two of the disk drives, wherein, if one of the disk drives in one of the groups having a lower level of redundancy fails, one of the disk drives in one of the group s having a higher redundancy replaces the failed disk drive, said disk drives being accessed by said disk array controller based upon the ranks and port numbers of the disk drives.

34. A method for providing an alternate device in place of a failed device in a device array apparatus in which the alternate device and the failed device are two of disk drives organized into ranks and a combination of RAID levels upon activation of the device array apparatus, said method comprising the steps of:

determining which of the devices has failed; and

reconstructing data stored on the failed device of one of the RAID levels having a lower level of redundancy into another of the devices of another of the RAID levels having a higher level of redundancy an d which is storing redundant information, wherein the disk array apparatus accesses the disk drives based upon the ranks and upon respective port numbers of the disk drives.

35. The method according to claim 34, wherein the devices comprise disk drives.

36. The method according to claim 34, wherein the devices comprise semiconductor memory.

37. The method according to claim 34, wherein the devices comprise optical drives.

38. The method according to claim 34, wherein the devices comprise magnetic tape drives.

39. The method according to claim 34, wherein the devices comprise disk drives, semiconductor memory, optical drives, and magnetic tape.

41. The apparatus according to claim 35, wherein the devices comprise disk drives.

42. The apparatus according to claim 35, wherein the devices comprise semiconductor memory.

43. The apparatus according to claim 35, wherein the devices comprise optical drives.

44. The apparatus according to claim 35, wherein the devices comprise magnetic tape drives.

45. The apparatus according to claim 35, wherein the devices comprise disk drives, semiconductor memory, optical drives, and magnetic tape.

46. The apparatus according to claim 35, wherein the devices correspond to RAID-1.

47. The apparatus according to claim 35, wherein the devices correspond to RAID-2.

48. The apparatus according to claim 35, wherein the devices correspond to RAID-3.

49. The apparatus according to claim 35, wherein the devices correspond to RAID-4.

50. The apparatus according to claim 35, wherein the devices correspond to RAID-5.

51. The apparatus according to claim 35, wherein the devices correspond to a combination of RAID-1, RAID-2, RAID-3, RAID-4, and RAID-5.

40. An apparatus coupled to a host computer transmitting an instruction and comprising:

a device array unit comprising devices executing input/output based upon the instruction, one of said devices being redundant of another of said devices, said devices being organized into ranks and combinations of RAID levels upon activation of the device array unit; and

a controller, coupled to the device array unit, determining which of the devices has failed, and reconstructing data stored on the failed device of one of the RAID levels having a lower level of redundancy into the one of said devices of another of the RAID levels having a higher level of redundancy and which is storing redundant data, said devices being accessed based upon the ranks and respective port numbers of the devices.

52. A method for providing an alternate device in place of a failed device in a device array apparatus in which the alternate device and the failed device are two of disk drives organized into ranks and a combination of RAID levels upon activation of the device array apparatus, said method comprising the steps of:

determining which of the devices has failed;

determining which of the devices are grouped together in a logic group having a relatively higher redundancy of data;

reconstructing data stored on the failed device into another of the devices which is storing redundant data in the logic group having the relatively higher redundancy, wherein the devices are accessed based upon the ranks and port numbers of the devices.

53. The method according to claim 52, wherein the devices correspond to a RAID.

55. The apparatus according to claim 53, wherein the devices correspond to a RAID.

54. An apparatus coupled to a host computer transmitting an instruction and comprising:

a device array unit comprising devices organized into ranks and combinations of RAID levels upon activation of the device array unit and executing input/output based upon the instruction, one of said devices being redundant of another of said devices, said devices being grouped into logic groups and accessed based upon the ranks and respective port numbers of the devices; and

a controller, coupled to the device array unit, determining which of the devices has failed, and reconstructing data stored on the failed device into the one of said devices which is redundant and which is grouped into one of the logic groups having a relatively higher redundancy of data.

56. An apparatus coupled to a host computer transmitting an instruction and comprising:

means for storing data comprising devices executing input/output based upon the instruction, one of said devices being redundant of another of said devices, said devices being grouped into logic groups and organized into ranks and combinations of RAID levels upon activation of the apparatus; and

means for determining which of the devices has failed and for reconstructing data stored on the failed device into the one of said devices which is redundant and which is grouped into one of the logic groups having a relatively higher redundancy of data, said devices being accessed based upon the ranks and upon respective port numbers of the devices.