Disk array apparatus which predicts errors using mirror disks that can be accessed in parallel
First Claim
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.
1 Assignment
0 Petitions
Accused Products
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.
116 Citations
28 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. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
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. - View Dependent Claims (19, 20)
-
-
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. - View Dependent Claims (22, 23, 24, 25, 26, 27)
-
-
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.
-
Specification