JP2006252336A - Inter-device data transfer apparatus, inter-device data transfer method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform a degenerate operation without stopping a system even when there is a fault in a storage device or a data processing device. <P>SOLUTION: An inter-device data transfer apparatus is provided with a table storing data transfer propriety between a plurality of magnetic disk devices and a plurality of I/O processors and fault flags indicating the presence of the fault of the respective devices and a bus connecting device. In the fault of the I/O processor, the magnetic disk device or the bus connecting device, by determining the ON/OFF of the fault flags for the respective devices and determining the transfer possibility of the magnetic disk device and the I/O processor constituting a logical volume, access to a fault device is limited without updating the entire management table. Thus, the system can be shifted to the degenerate operation without stopping the system. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a volume management module in data transfer between a plurality of devices connected to a bus.

  FIG. 15 shows an example of the system configuration of the inter-device data transfer apparatus.

  A central processing unit (CPU) 10 is in charge of controlling each device of the entire system. Reference numeral 9 denotes a main storage device, and software of the entire system is controlled by an operating system 90 stored in the main storage device 9. Reference numeral 1 denotes a bus (primary bus) that connects the central processing unit 10, the main storage device 9, and each device, and a bus (secondary bus) 2a by a bus coupling device 3a, 3b for connecting a plurality of buses. 2b. The bus coupling devices 3a and 3b are, for example, bridges. The magnetic disks 60a and 61a are connected to the bus 2a by a disk controller (adapter) 4a to constitute one storage device 7a. Similarly, the magnetic disks 60b and 61b are connected to the bus 2b by the disk controller 4b and constitute one storage device 7b. The magnetic disks 60a, 61a, 60b, 61b are hereinafter also referred to as physical disks. The device 50a is connected to the bus 2a, and the device 50b is connected to 2b. For example, the device 50a is an I / O processor including a storage device therein.

  Usually, a plurality of devices 7a and 7b and devices 50a and 50b are connected in one system, and the devices 7a and 7b and the devices 50a and 50b directly transfer data to each other. The data stored in the devices 7a and 7b is normally transferred to the main storage device 9 and processed by the central processing unit 10, but in order to reduce the load on the central processing unit and the data transfer on the bus, In some cases, direct data transfer (peer-to-peer data transfer) between devices without using a storage device is performed. That is, the data stored in the devices 7a and 7b is transferred to the storage device inside the devices 50a and 50b without going through the main storage device and processed by the I / O processor inside the devices 50a and 50b. There is.

  FIG. 16 is an internal configuration diagram of a volume management table for holding system configuration information disclosed in Patent Document 1. FIG. 17 is an internal configuration diagram of a configuration information management table for holding redundant configuration information of logical volumes disclosed in Patent Document 2.

In the conventional inter-device data transfer method, the system configuration information is held in a static table, and the transfer possibility between devices is determined by referring to the bus transfer possibility by following the hierarchy of the connected bus. (Patent Document 1). Also, a method has been proposed in which an inter-device data transfer apparatus has a redundant configuration by holding the same data in units of logical volumes and is substituted when a logical volume fails (Patent Document 2).
Japanese Patent No. 2978882 “Inter-device data transfer apparatus and inter-device data transfer method” JP 2003-308174 A "Inter-device data transfer apparatus, inter-device data transfer method and program"

  In the conventional inter-device data transfer method, the system configuration information is held in a static state as a volume management table or a redundancy management table, so that each device such as an I / O processor or a magnetic disk device or a bus coupling device has failed. In this case, it is necessary to regenerate the volume management table and the redundancy management table in order to enable the degenerate operation without using the failed device.

  Therefore, after the operating system recognizes the failure of the device, collect the system information again with the failed device disconnected from the system, or from the volume management table or redundancy management table, the failed device information or logical volume It was necessary to restrict access to the failed device by deleting the information. At that time, there is a problem that the system must be stopped for a while during the restart of the operating system and the generation of the volume management table and the redundancy management table.

  Further, there is a problem in that waiting for an error or timeout may occur when attempting to collect system information from the operating system in a state where the failed device is not disconnected from the system.

  In addition, since the information is completely deleted from the volume management table and redundancy management table in the event of a failure, when the failed device is restored, the system is There was a problem that I had to stop.

  In the conventional device-to-device data transfer method, the transferability level is determined based on the transferability of the bus information in the volume management table and the priority of the logical volume information in the redundancy management table. Therefore, the I / O processor or magnetic disk When a device such as a device or a bus coupling device does not necessarily become completely unusable but intermittently becomes a failure state, the transfer possibility is determined so as not to cause an error as much as possible. There was a problem that could not.

  In addition, the determination of transferability when a logical volume fails is because the priority is determined only between logical volumes that have the same data and have a redundant configuration. There was a problem that it was not possible to judge the transfer possibility.

  In addition, since transferability information was held only for the bus, when each device was used depending on the failure status of the individual I / O processors and logical disks that make up the logical volume connected to the same bus There was a problem that it was not possible to judge the transfer possibility.

  In addition, since the conventional inter-device data transfer method uses only the number of levels of each bus coupling device that connects the buses to determine the transfer possibility, the failure status of each bus coupling device and the bus hierarchy are overlapped. There was a problem that the transferability could not be determined by identifying the change in transferability due to.

  In addition, the conventional inter-device data transfer method is based on the possibility of transfer of the highest level bus traced from the magnetic disk device constituting the I / O processor or logical volume, or transfer of the bus directly connected to the I / O processor. Since only the possibility was determined, there was a problem that it was not possible to determine the transfer possibility on a bus having a connection relationship other than that.

  In the conventional inter-device data transfer method, it is impossible to prevent an error due to a failure.

  The present invention has been made to solve the above-described problems, and it is an object of the present invention to minimize a system stop period when an I / O processor, a magnetic disk device, or a bus coupling device fails. . It is another object of the present invention to determine transferability of inter-device data transfer according to the failure state of each device when an I / O processor, magnetic disk device, or bus coupling device fails. Another object of the present invention is to prevent errors in advance by making a failure prediction.

An inter-device data transfer apparatus according to the present invention is as follows.
A plurality of storage devices and a plurality of data processing devices are joined via a plurality of transfer paths and at least one transfer path coupling device, and between the storage device and the data processing device via a transfer path and a transfer path coupling device. In a device-to-device data transfer device that transfers data,
A table storage unit that stores information about a failure of the storage device, information about a failure of the data processing device, and a table showing information about the failure of the transfer path coupling device;
Whether data transfer between the storage device and the data processing device is possible based on the storage device failure information, the data processing device failure information, and the transfer path coupling device failure information shown in the table. And a data transfer control unit for identifying the device.

  According to the present invention, between the storage device and the data processing device based on the information on the failure of the storage device stored in the table, the information on the failure of the data processing device, and the information on the failure of the transfer path coupling device. Therefore, even if there is a failure in any of the devices, the table need not be regenerated and the system need not be stopped.

Embodiment 1 FIG.
FIG. 1 is an example of a system configuration of an inter-device data transfer apparatus for realizing the inter-device data transfer method of the present embodiment.

  A central processing unit (CPU) 10 is in charge of controlling each device of the entire system. Reference numeral 9 denotes a main storage device, and software of the entire system is controlled by an operating system 90 stored in the main storage device 9. Reference numeral 1 denotes a bus that connects the central processing unit 10, the main storage device 9, and each device, and is connected to the buses 2a and 2b by bus coupling devices 3a and 3b for connecting a plurality of buses. The bus coupling devices 3a and 3b are bridges, for example. The bus coupling devices 3a and 3b are examples of transfer path coupling devices. The magnetic disk devices 60a and 61a are connected to the bus 2a by the disk controller 4a to constitute one storage device 7a. Similarly, the magnetic disk devices 60b and 61b are connected to the bus 2b by the disk controller 4b to constitute one storage device 7b. Similarly, the magnetic disk devices 60c and 61c are connected to the bus 2b by the disk controller 4c to constitute one storage device 7c. The devices 50a and 51b are connected to the bus 2a, and the device 50b is connected to 2b. For example, the devices 50a and 51b are I / O processors including a storage device inside.

  Usually, a plurality of devices 7a, 7b, 7c and devices 50a, 51a, 50b are connected in one system, and the inter-device data transfer apparatus according to the present embodiment includes devices 7a, 7b. 7c and devices 50a, 51a, 50b directly transfer data to each other.

  The data stored in the devices 7a, 7b, and 7c is normally transferred to the main memory 9 and processed by the central processing unit 10, but in order to reduce the load on the central processing unit and the data transfer on the bus In some cases, direct data transfer between devices (peer-to-peer data transfer) is performed without using the main storage device. That is, in the inter-device data transfer apparatus according to the present embodiment, data stored in the devices 7a, 7b, and 7c is transferred to the storage devices inside the devices 50a, 51a, and 50b without passing through the main storage device. It is processed by an I / O processor inside 50a, 51a, 50b.

  In the present embodiment, it is assumed that the storage devices 7a, 7b, and 7c are set as logical volumes.

  Here, the logical volume is a combination of a plurality of partitions so that it can be handled as one logical storage device. For example, all levels (0 to 5) and “spanned volume” of so-called software RAID (Redundant Arrays of Inexpensive Disk) are also included. The partition may be the entire physical disk or a part thereof.

  In the present embodiment, a procedure for determining transferability of inter-device data transfer when the I / O processor 50a, the bus coupling device 3b, or the magnetic disk device 61b has failed will be described.

  In the present embodiment, the volume management method and the data transfer control method according to the present invention are software stored in the main storage device 9 as software stored in a volume management module 92 (hereinafter also referred to as volume management) and a data transfer control module 93 (hereinafter referred to as data). (Also called transfer control). The volume management 92 and the data transfer control 93 are implemented as software separated from the operating system 90, and can query the operating system 90 for system information.

  Note that the volume management module 92 and the data transfer control module 93 may be realized as firmware, hardware, or a combination of software, firmware, and hardware.

  The volume management module 92 and the data transfer control module 93 are execution circuits that execute a program by the CPU when the function is realized by the program. That is, the volume management module 92 is an example of a volume management unit that causes a CPU to execute a volume management program. The data transfer control module 93 is an example of a data transfer control unit that causes a CPU to execute a data transfer control program.

  The operating system is an execution circuit that executes an operating system program by the CPU, and the operating system is an operating system execution unit that executes the operating system program by the CPU.

  FIG. 2 is a configuration diagram showing the relationship between the operating system, volume management, and data transfer control in this embodiment.

  The arrow in FIG. 2 indicates the side that receives from the side that performs the inquiry, reference, processing request, and generation.

  The operating system 90 issues a request for collecting information to the device driver 96 at the initial time or when the user uses the tool, and generates the volume configuration table 91 and the bus information 95. The volume configuration table 91 stores volume configuration information managed by the operating system 90.

  The volume management 92 directly refers to the volume configuration table 91 after initialization of the operating system, or sends volume configuration information, bus information, physical disk / partition information, bus connection information, I / O processor information, etc. to the operating system 90. An information collection request is issued, and a volume management table 94 is generated. For example, the volume management table 94 is stored in the main storage device 9. In this case, the main storage device 9 is an example of a table storage unit.

  The data transfer control 93 issues a request for collecting information to the volume management 92 before data transfer to acquire transferability or transfer efficiency of the transfer path. At this time, the volume management 92 refers to the volume management table 94.

  The data transfer control 93 issues a data transfer request to the operating system 90 based on the transfer path information. The operating system 90 passes the data transfer request to the device driver 96 in accordance with the data transfer control request.

  In particular, the data transfer method of the data transfer control 93 and the volume management table 94 are different in this embodiment compared to Japanese Patent No. 2978882 “Inter-device data transfer apparatus and inter-device data transfer method”.

  FIG. 3 shows the internal structure of the volume management table 94 in FIG. The volume management table mainly includes bus transfer information 100, bus coupling device information 110, physical disk information 120, logical volume information 130, and I / O processor information 140.

  The bus transfer information 100 mainly stores a bus number 101 and a bus transfer possibility 102. The transfer possibility is transfer efficiency including the impossible of peer-to-peer data transfer. These sets are stored for the number of buses.

  The bus coupling device information 110 mainly stores an upper bus number 112 and a lower bus number 113 of a bus connected to the device number 111 of the bus coupling device, and a failure flag 114 indicating a failure of the bus coupling device. These number sets are stored as many as the number of bus coupling devices.

  The physical disk information 120 mainly stores a bus number 122 connected to the physical disk number 121, partition information 123 in the physical disk, and a failure flag 124 indicating a physical disk failure. A set of these numbers and information is stored as many as the number of physical disks.

  The partition information stores a partition number or partition offset identified by the operating system.

  The logical volume information 130 mainly stores the logical volume name 131 and the partition information 132 of the physical disk constituting the logical volume. A set of these numbers and information is stored for the number of logical volumes.

  The I / O processor information 140 mainly stores an I / O processor number 141, a connection bus number 142, and a failure flag 143 indicating an I / O processor failure. A set of these numbers is stored as many as the number of I / O processors.

  One of the features of the present embodiment is that the volume management table has failure flags 114, 124, and 143 indicating the failure state of each device. The value of the failure flag can be changed by a user or an external program.

  4 and 5 show a procedure for identifying transferability of a transfer path at the time of data transfer when the data transfer control according to the present embodiment fails in an I / O processor, a bus coupling device, or a magnetic disk device. It is a flowchart.

  First, in FIG. 4, the transfer control 93 refers to the I / O processor information 140 through the volume management 92, and obtains a connection bus number 142 corresponding to the transfer destination I / O processor number (step S11).

  It is checked whether or not the target bus can be transferred with reference to the transfer possibility 102 of the bus transfer information 100 (step S12).

  If the bus is not capable of peer-to-peer transfer, the process proceeds to step S15. The upper bus number 112 is obtained from the lower bus number 113 of the bus coupler information 110 (step S13).

  If the upper bus number exists, the process returns to step S12 with the upper bus number as the target bus number (step S13).

  After the loop from step S12 to step S14 is completed, the last bus number is stored as the highest number of the target I / O processor (step S15).

  Next, referring to the I / O processor information in the volume management table 94 in FIG. 5, if the failure flag is not OFF, the target logical volume and the I / O processor cannot be transferred (step S21).

  If the failure flag is OFF, the partition information 132 constituting the logical volume is listed with reference to the logical volume information 130 (step S22).

  If the failure flag of the physical disk information 120 corresponding to the listed partition information is referred to and the flag is not OFF, the target logical volume and the I / O processor cannot be transferred (step S23).

  If the failure flag is OFF, the connection bus number 122 connected to the partition information 123 of the physical disk information 120 corresponding to the listed partition information is obtained (step S24).

  It is checked whether or not the target bus can be transferred with reference to the transfer possibility 102 of the bus transfer information 100 (step S25).

  If the bus is not capable of peer-to-peer transfer, the process proceeds to step S26. If the lower bus number of the bus coupling device information 110 refers to the failure flag of the bus coupling device information corresponding to the bus number and is not OFF, the process goes through a loop following the bus hierarchy and proceeds to S29. If the failure flag is OFF, the upper bus number 112 is obtained from the lower bus number 113 of the bus coupler information 110 (step S27).

  If the upper bus number exists, the process returns to step S25 with the upper bus number as the target bus number. After the loop from step S25 to step S29 is completed, the last bus number is set as the highest bus number of the target partition and compared with the highest bus number of the I / O processor saved in step S15 in FIG. 4 (step S29). .

  If the highest bus numbers do not match, the target logical volume and the I / O processor cannot be transferred. If the highest bus numbers match, the process returns to step S23 if the next partition exists (step S30).

  When the highest bus number for all partition information matches the highest bus number of the I / O processor, the target logical volume and the I / O processor can perform peer-to-peer transfer.

  Data transfer control is performed in accordance with each of the above procedures (FIGS. 4 and 5) in a system in which an I / O processor, bus coupling device, or magnetic disk device has failed, from a logical volume storing data to peer-to- An I / O processor capable of peer transfer can be found.

  If there is no I / O processor that can transfer data, a warning may be issued to the user, or it may be transferred to the main memory and processed by a central processing unit (CPU).

  In this embodiment, a “magnetic disk” is used as an example of a storage device constituting a logical volume, but any storage medium such as a magnetic tape or a semiconductor disk may be used. Further, it may be a remote drive via a network.

  As described above, the inter-device data transfer apparatus and the inter-device data transfer method according to the present embodiment have a table for storing data transfer availability between a plurality of storage devices and a plurality of data processing devices, and an I / O processor. When a failure occurs in the magnetic disk device and the bus coupling device, the failure flag is set ON / OFF for each device, and the transfer possibility between the magnetic disk device and the I / O processor constituting the logical volume is determined. Thus, access to the failed device is limited without updating the entire management table.

  As described above, since the presence / absence of a failure in each of the I / O device, the magnetic disk device, and the bus coupling device is determined, it is possible to identify the failure state of each device and determine the degenerate operation.

  In other words, if there is no failure flag in the volume management table as in the prior art, the presence or absence of a failure can be determined after confirming the transferability of each device. Although the table had to be newly recreated, in this embodiment, the presence / absence of a failure in each of the I / O device, magnetic disk device, and bus coupling device is determined based on the failure flag, and the failure is determined. The determined device determines that peer-to-peer transfer is not possible and does not consider it as a transfer target. Therefore, the device is in the same state as when the failed device is disconnected, and there is no need to recreate the volume management table. For this reason, the degenerate operation can be started without stopping the system.

  In addition, in the bus hierarchy search loop for determining transferability, if there is a faulty device, it is interrupted, so it is possible to determine whether transfer is impossible without checking all connected devices. . In addition, since only some values in the table that holds the system configuration are changed, the configuration information can be changed promptly according to changes in the system status, and the system downtime can be minimized. Can do.

  In addition, since it is determined that the device cannot be used while retaining information such as the connection relationship of the failed device, the system configuration can be changed quickly when the failed device is restored, minimizing the system outage period. can do.

  As a result, it is possible to notify the user and other programs of the importance / necessity of recovery according to the state before the failure.

  Further, the data on the magnetic disk device can be sent to an I / O processor that has not failed. Further, if all logical volumes can be transferred by at least one or more I / O processors, it can be determined that the system can perform degenerate operation.

  Further, with respect to the I / O processor that processes the data of the logical volume, the number of I / O processors that can be transferred per logical volume is obtained, and a warning is issued when the specific number is not satisfied. It is possible to notify the user and other programs to keep the data processing performance above a certain level in the event of a failure.

  In addition, by issuing a warning when the number of I / O processors that can transfer logical volume data does not satisfy a specific number, users and others can ensure a certain level of reliability in the event of an I / O processor failure. It is possible to inform the program.

Embodiment 2. FIG.
In the first embodiment described above, the failure of the I / O processor, the bus coupling device, and the magnetic disk device is identified. Next, the I / O processor, the bus coupling device, and the magnetic disk device are not necessarily complete. An embodiment for identifying a case in which a failure occurs intermittently such that an error occurs with a certain probability, although it is not usable.

  FIG. 6 is an example of a system configuration for explaining the operation in such a case.

  A central processing unit (CPU) 10 is in charge of controlling each device of the entire system. Reference numeral 9 denotes a main storage device, and software of the entire system is controlled by an operating system 90 stored in the main storage device 9. Reference numeral 1 denotes a bus connecting the central processing unit 10, the main storage device 9, and each device, and is connected to the bus 2d by a bus coupling device 3d for connecting a plurality of buses. Further, the bus 2d is connected to the bus 2e by a bus coupling device 3e. The bus coupling devices 3d and 3e are, for example, bridges. The magnetic disk devices 60e and 61e are connected to the bus 2e by the disk controller 40e and are one storage device and constitute a logical volume 80e. Similarly, the magnetic disk devices 62e and 63e are connected to the bus 2e by the disk controller 41e and are one storage device and constitute a logical volume 81e. The device 5f is connected to the bus 1, the device 5d is connected to the bus 2d, and the device 5e is connected to 2e. For example, the device 5f is an I / O processor including a storage device.

  Normally, a plurality of logical volumes 80e and 81e and devices 5d, 5e, and 5f are connected in one system, and the inter-device data transfer apparatus according to the present embodiment is logical volumes 80e and 81e and devices 5d, 5e, and 5f. Data is directly transferred between each other.

  Data stored in the logical volumes 80e and 81e is normally transferred to the main storage device 9 and processed by the central processing unit 10, but in the inter-device data transfer apparatus according to the present embodiment, the data is stored in the logical volumes 80e and 81e. The stored data is transferred to the storage device inside the devices 5d, 5e, 5f without going through the main storage device, and is processed by the I / O processor inside the 5d, 5e, 5f.

  In the present embodiment, an example is shown in which logical volumes are set in units of physical devices, but it is also possible to configure logical volumes that span a plurality of storage devices.

  In the present embodiment, when the I / O processors 5d and 5e, the bus coupling device 3d, and the magnetic disk device 63e are intermittently in a state of failure such that an error occurs with a probability that they are not completely unusable. Shows a procedure for identifying the transferability of data transfer between devices.

  For example, when the failure level, which is a staged failure state, is the occurrence rate of failure, the failure level of the I / O processor 5d is 20%, the failure level of the I / O processor 5e is 50%, and the bus coupling device The failure level of 3d can be expressed as 10%, and the failure level of the magnetic disk device 63e can be expressed as 30%. This failure level is an example of a failure occurrence frequency value.

  The relationship between the operating system, volume management, and data transfer control in this embodiment is shown in FIG. 2 as in the first embodiment.

  In the present embodiment, the data transfer method of the data transfer control 93 and the volume management table 94 are different from those in the first embodiment.

  FIG. 7 shows the internal structure of the volume management table 94.

  The volume management table mainly includes bus transfer information 100, bus coupling device information 110, physical disk information 120, logical volume information 130, and I / O processor information 140.

  The bus transfer information 100 mainly stores a bus number 101 and a bus transfer possibility 102. The transfer possibility is transfer efficiency including the impossible of peer-to-peer data transfer. These sets are stored for the number of buses.

  The bus coupling device information 110 mainly stores an upper bus number 112 and a lower bus number 113 of a bus connected to the device number 111 of the bus coupling device, and a failure level 115 indicating the degree of failure of the bus coupling device. These number sets are stored as many as the number of bus coupling devices.

  The physical disk information 120 mainly stores a bus number 122 connected to the physical disk number 121, partition information 123 in the physical disk, and a failure level 125 indicating the degree of failure of the physical disk. A set of these numbers and information is stored as many as the number of physical disks.

  The partition information stores a partition number or partition offset identified by the operating system.

  The logical volume information 130 mainly stores the logical volume name 131 and the partition information 132 of the physical disk constituting the logical volume. A set of these numbers and information is stored for the number of logical volumes.

  The I / O processor information 140 mainly stores an I / O processor number 141, a connection bus number 142, and a failure level 144 indicating the degree of failure of the I / O processor. A set of these numbers is stored as many as the number of I / O processors.

  The volume management table has failure levels 115, 125, and 144 indicating the degree of failure of each device. The failure level can be changed by a user or an external program.

  FIGS. 8 and 10 show the transfer possibility of the transfer path during data transfer when the data transfer control according to this embodiment is intermittently in the state of failure of the I / O processor, the bus coupling device, or the magnetic disk device. It is the flowchart which showed the procedure which judges.

  First, in FIG. 8, the transfer control 93 refers to the I / O processor information 140 through the volume management 92 and stores the value of the failure level 144 (step S31).

  Subsequently, the connection bus number 142 corresponding to the transfer destination I / O processor number is obtained (step S32).

  It is checked whether or not the target bus can be transferred with reference to the transfer possibility 102 of the bus transfer information 100 (step S33).

If the bus is not capable of peer-to-peer transfer, the process proceeds to step S37. In the case of YES in S33, the failure level 115 relating to the bus bus device information 110 having the same lower bus number 113 is cumulatively calculated in the stored failure level and added to the connection bus list of the I / O processor. (Step S34)
FIG. 9 is a diagram showing the contents of the connection bus list of the I / O processor.

  The cumulative value 152 of the failure level is stored for each bus number 151. The cumulative calculation method may be, for example, a MAX operation in which the larger one of the already stored failure level value and the target failure level is stored in memory, simple addition may be performed, When the level is a failure occurrence rate, it may be (1- (1-recording failure level / 100) * (1-target failure level / 100)) * 100.

  Subsequently, the upper bus number 112 of the bus coupler information 110 is obtained (step S35).

  If the upper bus number exists, the process returns to step S33 with the upper bus number as the target bus number (step S36).

  After the loop from step S33 to step S36 is completed, the connection bus list of the target I / O processor is saved (step S37).

  FIG. 10 is a flowchart showing a procedure for obtaining transferability of data transfer between devices between the I / O processor and the logical volume.

  With reference to the logical volume information 130 of the volume management table 94, the partition information 132 constituting the logical volume is listed (step S41).

  A connection bus number 122 corresponding to the partition information 123 of the physical disk information 120 is obtained (step S42).

  The failure level corresponding to the partition information 123 of the physical disk information 120 is stored (step S43).

  It is checked whether or not the target bus can be transferred with reference to the transfer possibility 102 of the bus transfer information 100 (step S44).

  If the bus is not capable of peer-to-peer transfer, the target logical volume and the I / O processor cannot be transferred. The failure level 115 corresponding to the lower bus number 113 of the bus coupling device information 110 is acquired and cumulatively calculated to the stored failure level (step S45).

  The cumulative calculation method may be, for example, a MAX operation in which the larger one of the already stored failure level value and the target failure level is stored in memory, simple addition may be performed, When the level is a failure occurrence rate, it may be (1- (1-recording failure level / 100) * (1-target failure level / 100)) * 100. The bus number is compared with the bus number 151 on the connection bus list 150 of the I / O processor, and if it does not exist, the process proceeds to S49 (step S46).

  If not, the higher bus number 112 is obtained from the lower bus number 113 of the bus coupler information 110 (step S47). If the upper bus number exists, the process returns to step S44 with the upper bus number as the target bus number (step S48).

  If there is no higher bus number, the failure level 152 of the corresponding bus number 151 on the connection bus list 150 of the I / O processor is cumulatively calculated for the stored failure level (step S49).

  The cumulative calculation method is the same as in S45. If the stored failure level exceeds the predetermined value, transfer is impossible (step S50). If the next partition exists, the process returns to step S44 (step S51).

  For all partition information, if the bus number that matches the I / O processor connection bus number exists in a part of the bus hierarchy and the accumulated failure level is within the prescribed failure level, the target logical volume and I The / O processor can perform peer-to-peer transfer.

FIG. 11 is an example in which transferability is obtained according to the procedure of FIGS. 8 and 10 in the configuration diagram of FIG.
Peer-to-peer transfer is possible between the magnetic disk device 60e and the I / O processor 5d because the failure level is less than the predetermined failure level 40. Similarly, peer-to-peer transfer is possible between the magnetic disk device 63e and the I / O processor 5f because it is less than the predetermined failure level 40 (in the case of MAX operation).

  Data transfer control is performed in accordance with each of the above procedures (FIGS. 8 and 10) in a system in which the degree of failure of the I / O processor, bus coupling device, and magnetic disk device is set, from the logical volume storing the data to the peer- An I / O processor capable of to-peer transfer can be found. If there is no I / O processor that can transfer data, a warning may be issued to the user, or it may be transferred to the main memory and processed by a central processing unit (CPU).

  In this embodiment, a “magnetic disk” is used as an example of a storage device constituting a logical volume, but any storage medium such as a magnetic tape or a semiconductor disk may be used. Further, it may be a remote drive via a network.

  As described above, the inter-device data transfer apparatus and the inter-device data transfer method according to the present embodiment have a table for storing data transfer availability between a plurality of storage devices and a plurality of data processing devices, and an I / O processor. When a failure occurs in the magnetic disk device and the bus coupling device, a failure level (error occurrence frequency) is set according to the severity of the failure of each device, and an error occurs in the I / O processor, magnetic disk device, and bus coupling device. By accumulating the occurrence frequency and determining the transferability between the magnetic disk device constituting the logical volume and the I / O processor, the I / O having a high error occurrence frequency is performed without updating the entire management table. Access to a processor, a magnetic disk device, and a bus coupling device is restricted.

  As described above, the failure level of each device of the I / O device, the magnetic disk device, and the bus coupling device is determined, so that the degree of failure of each device can be identified and the degenerate operation can be determined.

  In addition, since only some values in the table that holds the system configuration are changed, the configuration information can be changed promptly according to changes in the system status, and the system downtime can be minimized. Can do.

  In addition, because the degree of transfer is determined while leaving information such as the connection relationship of the device under failure, the system configuration can be changed quickly when the degree of failure (fault occurrence frequency) is changed. The system downtime can be minimized.

  Further, since the failure level of the bus coupling device is cumulatively calculated, degeneration corresponding to the failure level can be determined even when the failure device is involved in a plurality of data transfers by overlapping the bus hierarchy.

  In addition, by setting the failure level based on the frequency of occurrence of failures in each device, it is possible to disable the devices in order from the devices that frequently generate errors.

  As a result, it is possible to notify the user and other programs of the importance / necessity of recovery according to the state before the failure.

  Further, data on the magnetic disk device can be sent to an I / O processor with a low failure level. Further, if all logical volumes can be transferred by at least one or more I / O processors, it can be determined that the system can perform degenerate operation.

  Further, with respect to the I / O processor that processes the data of the logical volume, the number of I / O processors that can be transferred per logical volume is obtained, and a warning is issued when the specific number is not satisfied. It is possible to notify the user and other programs to keep the data processing performance above a certain level in the event of a failure.

  In addition, by issuing a warning when the number of I / O processors that can transfer logical volume data does not satisfy a specific number, it is possible to ensure a certain level of reliability in the event of an I / O processor failure. It is possible to inform the program.

Embodiment 3 FIG.
In the first and second embodiments described above, the failure and the degree of failure of the I / O processor, the bus coupling device, and the magnetic disk device are identified to determine the transfer possibility. It is a user, an external program, or an operating system that detects this and determines the value of the failure level to be set.

  In the present embodiment, the failure level of the bus coupling device is changed according to the failure level value that has already been set, so that the possibility of failure can be predicted and used for determining transferability.

  FIG. 12 shows an example of a system configuration for explaining the operation in such a case.

  A central processing unit (CPU) 10 is in charge of controlling each device of the entire system. Reference numeral 9 denotes a main storage device, and software of the entire system is controlled by an operating system 90 stored in the main storage device 9. Reference numeral 1 denotes a bus that connects the central processing unit 10, the main storage device 9, and each device, and is connected to the buses 20g and 21g by bus coupling devices 30g and 31g for connecting a plurality of buses, respectively. Furthermore, the buses 20g and 21g are connected to the bus 2h by bus coupling devices 30h and 31h. The magnetic disk devices 60h and 61h are connected to the bus 2h by the disk controller 4h and are one storage device and constitute a logical volume 8h. Similarly, the magnetic disk devices 60i and 61i are connected to the bus 21g by the disk controller 4i and are one storage device and constitute a logical volume 8i. The device 5j is connected to the bus 1, the device 50g is connected to the bus 20g, and the device 51g is connected to 21g.

  Normally, a plurality of logical volumes 8h, 8i and devices 50g, 51g, 5j are connected in one system, and the inter-device data transfer apparatus according to the present embodiment is logical volumes 8h, 8i and devices 50g, 51g, 5j. Data is directly transferred between each other.

  The data stored in the logical volumes 8h, 8i is normally transferred to the main storage device 9 and processed by the central processing unit 10, but in the inter-device data transfer apparatus according to the present embodiment, the logical volumes 8h, 8i The data stored in is transferred to a storage device inside the devices 50g, 51g, 5j without going through the main storage device, and is processed by an I / O processor inside the 50g, 51g, 5j.

  In this embodiment, an example in which a logical volume is set in units of physical storage devices is shown, but it is also possible to configure a logical volume that spans a plurality of physical storage devices.

    In the present embodiment, the I / O processor 51g, the bus coupling device 30g, and the magnetic disk devices 61h and 61i are in an intermittent failure state that causes an error with a probability that they are not completely unusable. For example, when the failure level indicating the state of failure intermittently is a failure occurrence rate, the failure level of the I / O processor 51g is 30%, the failure level of the bus coupling device 30g is 30%, and the magnetic disk device 61h. The failure level is 20%, and the failure level of the magnetic disk device 61i is 30%. This failure level is an example of a failure occurrence frequency value.

  13 and 14 are flowcharts showing a failure prediction method according to this embodiment.

  FIG. 13 is substantially the same as FIG. 8 in the second embodiment. FIG. 14 is almost the same as FIG. 10 in the second embodiment. However, in FIG. 13, the maximum value operation MAX is used for the cumulative calculation method in S34 of FIG. 8, and the operation result is stored in each failure level 115 of the bus coupler information 110 referred to as the connected bus list. (S38). That is, the value of the failure level accumulation 152 of the I / O processor connection bus list 150 and the value of the failure level 115 of the bus coupler information 110 are subjected to MAX operation, and a larger value is set as a new failure level value. The value of the failure level is stored in both the failure level accumulation 152 of the I / O processor connection bus list 150 and the failure level 115 of the bus coupler information 110. Therefore, if the value of the failure level accumulation 152 in the I / O processor connection bus list 150 is larger than the value of the failure level 115 in the bus coupler information 110, the value of the fault level 115 in the bus coupler information 110 is The value of the failure level accumulation 152 in the I / O processor connection bus list 150 is rewritten, and as a result, the value of the failure level 115 in the bus coupler information 110 is increased.

  Similarly, in FIG. 14, the maximum value operation MAX is used in the cumulative calculation method of S35 of FIG. 10, and the operation result is stored in the failure level 115 of the bus coupler information 110 referring to the operation result (S52).

  In the case of the configuration of FIG. 12 and the value of the failure level, in the method of FIGS. 8 and 10 of the second embodiment, transfer is possible between the logical volume 8h and the I / O processor 5j because it is less than the prescribed failure level. Although determined, the buses related to the bus coupling devices 30h and 31h are set by newly setting the failure level based on the value of the failure level already set as shown in FIGS. 13 and 14 of the present embodiment. The value 20 is set to the failure level 115 of the coupling device information 110 (by S52 in FIG. 14), and the value 30 is set to the failure level 115 of the bus coupling device information 110 related to the bus coupling device 31g. As a result, it is determined that transfer between the logical volume 8h and the I / O processor 5j is impossible because the specified failure level is exceeded.

  As described above, the inter-device data transfer apparatus and the inter-device data transfer method according to the present embodiment, in the volume management table having a failure level, use the upper bus according to the failure level value of the lower bus device and the bus coupling device. By setting the value of the failure level of the coupling device in advance, access to the entire block having a high degree of failure is restricted, and an error at the time of failure is avoided in advance.

  As described above, the failure level of the bus coupling device is newly set based on the failure level of each of the I / O device, the magnetic disk device, and the bus coupling device. An error level can be avoided in advance by setting a failure level for the entire apparatus.

FIG. 3 is a diagram illustrating an example of a system configuration according to the first embodiment. FIG. 3 is a configuration diagram showing the relationship between the operating system, volume management, and data transfer control according to the first embodiment. FIG. 4 is a diagram showing an example of a volume management table according to the first embodiment. FIG. 3 is a flowchart showing a procedure for identifying transferability of a transfer path according to the first embodiment. FIG. 3 is a flowchart showing a procedure for identifying transferability of a transfer path according to the first embodiment. FIG. 4 is a diagram illustrating an example of a system configuration according to a second embodiment. FIG. 10 is a diagram showing an example of a volume management table according to the second embodiment. The flowchart figure which showed the procedure which identifies the transfer possibility of the transfer path | route which concerns on Embodiment 2. FIG. FIG. 10 is a diagram showing an example of an I / O processor connection bus list according to the second embodiment. The flowchart figure which showed the procedure which identifies the transfer possibility of the transfer path | route which concerns on Embodiment 2. FIG. The figure which shows the transfer possibility in the system configuration example concerning Embodiment 2. FIG. FIG. 10 is a diagram showing a system configuration example according to the third embodiment. FIG. 10 is a flowchart showing a procedure for identifying transferability of a transfer path according to the third embodiment. FIG. 10 is a flowchart showing a procedure for identifying transferability of a transfer path according to the third embodiment. The figure which shows the system structural example which concerns on a prior art. The figure which shows the volume management table which concerns on a prior art. The figure which shows the redundancy management table which concerns on a prior art.

Explanation of symbols

  1 bus, 2 bus, 3 bus coupling device, 4 disk controller, 5 I / O processor, 6 physical disk, 7 physical storage device, 9 main storage device, 10 central processing unit, 50 I / O processor, 51 I / O Processor, 90 operating system, 91 volume configuration table, 92 volume management, 93 data transfer control, 94 volume management table, 95 bus information, 96 device driver.

Claims (8)

A plurality of storage devices and a plurality of data processing devices are joined via a plurality of transfer paths and at least one transfer path coupling device, and between the storage device and the data processing device via a transfer path and a transfer path coupling device. In a device-to-device data transfer device that transfers data,
A table storage unit that stores information about a failure of the storage device, information about a failure of the data processing device, and a table showing information about the failure of the transfer path coupling device;
Whether data transfer between the storage device and the data processing device is possible based on the storage device failure information, the data processing device failure information, and the transfer path coupling device failure information shown in the table. An inter-device data transfer apparatus comprising: a data transfer control unit for identifying the device. The table further includes
The physical location of the storage device, the physical location of the data processing device and the data transfer efficiency of the transfer path;
The data transfer control unit
Information on storage device failure, information on data processing device failure, information on transfer path coupling device failure, physical location of storage device and data processing device stored in table 2. The inter-device data transfer apparatus according to claim 1, wherein whether or not data transfer is possible between the storage device and the data processing device is identified based on the position and the data transfer efficiency of the transfer path. The table further includes
The physical location of the storage device, the physical location of the data processing device and the physical location of the transfer path combiner;
The data transfer control unit
Based on the physical location of the storage device, the physical location of the data processing device, and the physical location of the transfer path coupling device shown in the table, the storage device and the data processing device to be identified for data transfer availability Extracting a transfer path coupling device used for data transfer between, information relating to the failure of the storage device that is subject to identification of whether data transfer is possible, information relating to the failure of the data processing device that is subject to identification of whether data transfer is possible, The inter-device data transfer apparatus according to claim 1, wherein whether or not data transfer is possible between the storage device and the data processing device is identified based on the extracted information relating to the failure of the transfer path coupling apparatus. The table above
Information relating to the failure of the storage device indicates a failure flag indicating the presence or absence of a failure of the storage device, information relating to the failure of the data processing device indicates a failure flag indicating the presence or absence of a failure of the data processing device, Indicates a failure flag indicating whether or not there is a failure in the transfer path coupling device,
The data transfer control unit
At least one of the failure flag of the storage device subject to identification of data transfer availability, the failure flag of the data processing device subject to identification of data transfer availability, and the failure flag of the extracted transfer path coupling device The inter-device data transfer apparatus according to claim 3, wherein it is determined that data transfer between the storage device and the data processing device is impossible. The table above
A failure occurrence frequency value indicating the occurrence frequency of the failure of the storage device is indicated as information relating to the failure of the storage device, a failure occurrence frequency value indicating the occurrence frequency of the failure of the data processing device is indicated as information relating to the failure of the data processing device, and the transfer path As the information about the failure of the coupling device, a failure frequency value indicating the frequency of occurrence of the failure of the transfer path coupling device is indicated,
The data transfer control unit
Failure occurrence frequency value of the storage device subject to identification of data transfer availability, failure occurrence frequency value of the data processing device subject to identification of data transfer availability, failure occurrence frequency value of the extracted transfer path coupling device 4. The inter-device data transfer apparatus according to claim 3, wherein when the accumulated calculated value of the data exceeds a predetermined threshold, it is determined that data transfer between the storage device and the data processing device is impossible. . The table above
A failure occurrence frequency value indicating the occurrence frequency of the failure of the storage device is indicated as information relating to the failure of the storage device, a failure occurrence frequency value indicating the occurrence frequency of the failure of the data processing device is indicated as information relating to the failure of the data processing device, and the transfer path As the information about the failure of the coupling device, a failure frequency value indicating the frequency of occurrence of the failure of the transfer path coupling device is indicated,
The data transfer control unit
Based on the physical location of the storage device, the physical location of the data processing device, and the physical location of the transfer path coupling device shown in the table, the storage device and the data processing device to be identified for data transfer availability A plurality of transfer path coupling devices used for data transfer between them are sequentially extracted, and for each of the extracted transfer path coupling devices, a failure frequency value of a storage device that is a target of data transfer availability identification and data transfer availability identification A new failure occurrence frequency value is set using at least one of the failure occurrence frequency value of the data processing device subject to the transfer and the failure occurrence frequency value of the transfer path coupling device previously extracted, and the object of data transfer availability identification The failure frequency value of the storage device to be extracted, the failure frequency value of the data processing device that is the object of identification of data transfer availability, and the extracted If the cumulative calculated value of each of the plurality of transfer path coupling devices and the new failure frequency value exceeds a predetermined threshold value, it is determined that data transfer between the storage device and the data processing device is impossible. The inter-device data transfer apparatus according to claim 3. About a plurality of storage devices and a plurality of data processing devices joined via a plurality of transfer paths and at least one transfer path coupling device, between the storage devices and the data processing devices via the transfer path and the transfer path coupling device In a device-to-device data transfer method for transferring data,
Information relating to storage device failure, information relating to data processing device failure, information relating to failure of transfer path coupling device, information relating to storage device failure, information relating to failure of data processing device, and transfer A failure information acquisition step for acquiring information on a failure of the path coupling device;
Data transfer between the storage device and the data processing device based on the information on the failure of the storage device acquired by the failure information acquisition step, the information on the failure of the data processing device, and the information on the failure of the transfer path coupling device A data transfer method comprising: a data transfer control step for identifying whether or not data transfer is possible. A plurality of storage devices and a plurality of data processing devices joined via a plurality of transfer paths and at least one transfer path coupling device, between the storage device and the data processing device via the transfer path and the transfer path coupling device In a program that causes a computer to execute data transfer,
Information relating to storage device failure, information relating to data processing device failure, information relating to failure of transfer path coupling device, information relating to storage device failure, information relating to failure of data processing device, and transfer Failure information acquisition processing for acquiring information related to the failure of the path coupling device;
Data transfer between the storage device and the data processing device based on the information on the failure of the storage device acquired by the failure information acquisition processing, the information on the failure of the data processing device, and the information on the failure of the transfer path coupling device A program for causing a computer to execute a data transfer control process for identifying whether or not the data can be transferred.

Images