Disk array apparatus which predicts errors using mirror disks that can be accessed in parallel

• US 5,611,069 A
• Filed: 07/12/1994
• Issued: 03/11/1997
• Est. Priority Date: 11/05/1993
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Matoba, Tatsuo
• Primary Examiner(s): Harvey, Jack B.
• Assistant Examiner(s): Etienne, Ario
• Application Number:
  US08/273,887
• Time in Patent Office: 973 Days
• Field of Search: 395/441, 395/442, 395/488, 395/489, 395/497.04, 395/414, 395/415, 395/421.11, 395/182.02, 395/182.03, 395/182.04, 395/182.05, 395/181, 395/575 , 395/495, 395/650, 371/10.2, 371/10.3
• US Class Current: 711/114
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/2087 with a common controller
• CPC Class Codes: G06F 2211/1045 Nested RAID, i.e. implementing a RAID scheme in another RAID scheme

Abstract

Even if a plurality of disk units in a disk array which operate in parallel simultaneously causes faults, a system can instantaneously cope with a processing request without losing the data. For this purpose, a mirror disk unit in which two disk units are provided as one set is used as a component element of the disk array. The two disk units of the mirror disk unit are allocated to the disk unit for the present use and the disk unit for spare. Data is written into both the presently used disk unit and the spare disk unit. Data is read out from the present use disk unit. The occurrence of a fault of the disk unit is judged and the allocation is switched from the present use disk unit to the spare disk unit. In an idle state, a simulation to check the disk array is executed and fault information is collected. RAID3 or RAID5 in which a parity is stored in the mirror disk construction can be also used. Further, the present use disk unit is not constructed as a mirror disk and when the fault is judged in the present use disk unit, the data is copied to the spare disk unit, thereby dynamically realizing a mirror disk construction.

Claims

1. A disk array apparatus comprising:

disk array means in which a plurality of mirror disk means, each having a set of two disk units for storing same data, are arranged so that they can be accessed in parallel,

control means for striping write data from a host computer, and thereafter writing said stripe data into said disk array means in parallel and for synthesizing data read out in parallel from said disk array means and transferring said synthesized data to a host computer;

a memory table for managing states of said plurality of disk units;

check circuit means for executing an access to said disk array means;

access judging means for judging whether an abnormality has occurred in said check circuit means or not;

abnormality number counting means for counting a number of abnormalities judged by said access judging means; and

fault judging means for deciding a device fault when the number of abnormalities counted by said abnormality number counting means is equal to or larger than a predetermined threshold value and for registering fault information in a device management table means;

wherein said control means executes processes to said disk array means with reference to said memory table when there is an access request from said host computer.

2. An apparatus according to claim 1, wherein said memory table registers information indicating that one of said disk units provided for said mirror disk means is set into a disk unit for present use and the other is set into a disk unit for spare use, and

said control means reads out data from said disk unit on the present use side of said mirror disk means with reference to said memory table at the time of a read access operation and writes data into both said disk unit for present use and said disk unit for spare use at the time of a write access operation.

3. An apparatus according to claim 2, wherein said disk array means has change-over control means for switching, such that in response to an instruction from said reading operation from said control means, said disk unit is switched to said disk unit on the present use side of said mirror disk means and the data is read out and, in response to an instruction from said write operation, said change over control means switches said disk unit so that data is written into both said disk unit for present use and said disk unit for spare use.

4. An apparatus according to claim 2, wherein said disk array means is constructed in a manner such that each of said two disk units constituting said mirror disk means are connected in parallel to said control means through an individual port, and said control means reads out data by accessing said disk unit on the present use side by one of the ports of the mirror disk means at the time of said read operation and writes the data by accessing both said disk unit for present use and said disk unit for spare use by both of the ports at said write operation.

5. An apparatus according to claim 4, wherein when there is a write request from said upper apparatus, said control means stripes the write data into the data of the number as many as the number of disk units which can be accessed in parallel in said mirror disk means and, after that, said control means forms the same data as each of said stripe data and transfers the pair of same stripe data to the two ports of the mirror disk means.

6. An apparatus according to claim 1, wherein said disk array means is constructed by:

a disk unit in which at least two disk modules are detachably built; and

an apparatus casing which has a plurality of enclosing portions in which said plurality of disk units are installed in accordance with a predetermined array construction.

7. An apparatus according to claim 6, wherein said disk unit has a power source section and cooling means as common portions for said detachable disk modules.

8. An apparatus according to claim 6, wherein said disk module has at least a disk medium, a disk rotating mechanism, and a head mechanism.

9. An apparatus according to claim 6, wherein said disk unit has therein a circuit unit of change-over control means for switching in a manner such that the disk unit is switched to the disk unit on the present use side of said mirror disk means and the data is read out in response to an instruction of the reading operation from said control means and the data is written to both of the present use disk unit and the spare disk unit in response to an instruction of the writing operation.

10. An apparatus according to claim 6, wherein said disk array means is constructed in a manner such that the disk unit in which at least two disk modules are detachably built is installed into a disk array casing in which another disk unit having therein a single disk module of a different medium size will be enclosed and said disk unit is used as a component element of the mirror disk means.

11. An apparatus according to claim 1, further having preventive maintenance information storing means for holding preventive maintenance information formed in correspondence to the abnormality judgment of said access judging means into a state in which it can be recognized by said upper apparatus.

12. An apparatus according to claim 1, further having device check means for executing a false accessing operation to said disk array means in an idle state of said control means and for checking said state of each said disk unit.

13. An apparatus according to claim 12, wherein said device check means executes the writing operation using dummy data to a specific data region of said disk array means.

14. An apparatus according to claim 12, wherein each time said access judging means judges an abnormality by the false accessing operation to said disk array means, said device check means allows said abnormality number counting means to count the number of abnormality times, and in the case where said number of abnormality times is equal to or larger than a predetermined threshold value and said fault judging means decides the occurrence of the fault, fault information is registered into said device management table means.

15. An apparatus according to claim 1, further having preventive maintenance information storing means for holding preventive maintenance information corresponding to abnormality information judged by said access judging means by the false accessing operation by said device check means in a state in which said preventive maintenance information can be recognized by said upper apparatus.

16. An apparatus according to claim 1 or 14, wherein when the reading operation is executed, in the case where the fault disk unit is recognized with reference to the fault information of said device management table means, so long as said fault disk unit is the disk unit on the present use side, said control means switches the operation to the reading operation from the spare disk unit, and so long as the fault disk unit is the spare disk unit, said control means maintains the reading operation from the present use disk unit.

17. An apparatus according to claim 1 or 14, wherein when the writing operation is executed, in the case where the fault disk unit is recognized with reference to the fault information of said device management table means, said control means disconnects the fault disk unit from a target of the writing operation.

18. A disk array apparatus comprising:

disk array means in which a plurality of disk units are arranged so that they can be accessed in parallel;

control means for allocating at least one of the disk units which belong to a same rank of said disk array means to the disk unit for spare use, for allocating the other remaining plurality of disk units for data storage, for striping write data from a host computer, for writing said stripe data in parallel to said disk units for data storage, for synthesizing said data read out in parallel from the disk units for data storage, and for transferring said synthesized data to said host computer;

check circuit means for executing an access to said disk array means;

access judging means for judging whether an abnormality has occurred in said check circuit means or not;

abnormality number counting means for counting the number of abnormalities judged by said access judging means;

fault judging means for predictively judging that there is a possibility of the occurrence of a fault in near future when the number of abnormalities counted by said abnormality number counting means is equal to or larger than a predetermined threshold value that is smaller than a threshold value which is used in an ordinary fault judgement; and

construction control means for selecting said spare disk unit at the time of the predictive judgement of said fault by said fault judging means, for copying the storage data of a faulty disk unit, for constructing mirror disk means by said faulty disk unit and said spare disk unit after completion of the copying operation, and for setting said operation to simultaneously write the data to both of said disk units and to read out the data from either one of the disk units into said control means.

19. An apparatus according to claim 18, wherein said construction control means sets said spare disk unit into a disk unit for present use and allows the data to be read out by said control means.

20. An apparatus according to claim 18, wherein said fault judging means has means for outputting and displaying a fact that the predictive judgement of the fault has been executed.

21. A disk array apparatus comprising:

a plurality of disk units each of which is constructed by two disk modules for storing same data, in which one mirror disk is constructed by said two disk modules;

an array controller for striping write data from a host computer and, thereafter writing said stripe data into said plurality of disk units in parallel, and for synthesizing data read out in parallel from said plurality of disk units and transferring said synthesized data to said host computer;

a memory table for managing states of said plurality of disk units;

a check circuit section for executing an access to said plurality of disk units;

an access judging circuit section for judging whether an abnormality has occurred in said check circuit section or not;

a counter for counting the number of abnormalities judged by said access judging circuit section; and

a fault judging circuit section for deciding a device fault when the number of abnormalities counted by said counter is equal to or larger than a predetermined threshold value and for registering fault information into said memory table;

wherein said array controller executes processes to said plurality of disk units with reference to said memory table when there is an access request from said host computer.

22. An apparatus according to claim 21, wherein said memory table registers information indicating that one of the disk modules provided for the disk units is set into a disk module for present use and the other is set into a disk module for spare use, and

said array controller reads out said data from the disk module on the present use side of said disk unit with reference to said memory table at the time of a read access operation and writes the data into both of the present use disk module and the spare use disk module at the time of a write access operation.

23. An apparatus according to claim 22, wherein each of said plurality of disk units has a change-over control circuit section for switching in a manner such that in response to the instruction of the reading operation from said array controller, the disk module is switched to the disk module on the present use side and the data is read out and, in response to the instruction of the writing operation, the data is written into both of the present use disk module and the spare disk module.

24. An apparatus according to claim 22, wherein each of said disk units is constructed in a manner such that each of the two built-in disk modules is connected in parallel to said array controller through an individual port, and said array controller reads out the data by accessing the disk module on the present use side by one of the ports at the time of the reading operation and writes the data by accessing both of the present use disk module and the spare disk module by both of the ports at the time of the writing operation.

25. An apparatus according to claim 24, wherein when there is a write request from said upper apparatus, said array controller stripes the write data into the data of the number as many as the number of disk units which can be accessed in parallel and, after that, said array controller forms the same data as each of said stripe data and transfers the pair of same stripe data to the two ports of each of said disk units.

26. An apparatus according to claim 21, further having a register circuit for holding preventive maintenance information formed in correspondence to the abnormality judgment of said access judging circuit section into a state in which it can be recognized by said upper apparatus.

27. A disk apparatus according to claim 21, wherein said array controller writes said stripe data into said plurality of disk units in parallel together with parity data.

28. A disk array apparatus comprising:

a plurality of disk modules arranged so that they can be accessed in parallel;

an array controller for allocating at least one of said disk modules which belong to a same rank in said plurality of disk modules for spare, for allocating said other plurality of disk modules for data storage, for striping write data from a host computer, for writing said stripe data in parallel to said disk modules for data storage, for synthesizing data read out in parallel from said disk modules for data storage, and for transferring said synthesized data to said host computer;

a check circuit section for executing an access to said plurality of disk modules;

an access judging circuit section for judging whether an abnormality has occurred in said check circuit section or not;

a counter for counting the number of abnormalities judged by said access judging circuit section;

a fault judging circuit section for predictively judging a possibility of an occurrence of a fault in near future when the number of abnormalities counted by said counter is equal to or larger than a predetermined threshold value that is smaller than a threshold value which is used in an ordinary fault judgement; and

a construction control circuit section for selecting said spare disk module at the time of the predictive judgement of said fault by said fault judging circuit section, for copying storage data of a faulty disk module, for constructing a mirror disk unit by said faulty disk module and said spare disk module after completion of the copying operation, and for setting the operation to simultaneously write the data to both of said disk modules and read out the data from either one of said disk modules into said array controller.