JP5874933B2 - Path control device, path control method, and path control program - Google Patents

Description

  The present invention relates to a path control device, a path control method, and a path control program, and more particularly to a path control device, a path control method, and a path control program in a multipath storage device.

  Examples of storage devices that switch controllers when a failure occurs in a controller in a storage device that secures redundancy by providing a plurality of controllers connected to the host device are, for example, Patent Document 1 and Patent Document 2 It is described in.

  The storage device system of Patent Document 1 includes an interface device group that can be failed over to a group of storage devices. When a failure occurs in a certain interface device, the interface device included in the same interface device group as that interface device takes over the processing. For all interface devices included in the interface device group, another one interface device included in the interface device group is set as an interface device that takes over processing. The interface device that takes over the processing is an interface device that is set as an interface device that takes over the processing of the interface device in which the failure has occurred. Alternatively, the interface device that takes over the processing is a normal interface device having the lowest operating rate in the interface device group including the interface device in which the failure has occurred.

  Patent Document 2 describes a storage device connected to a host by two paths. The controller of the storage apparatus counts the number of occurrences of minor faults such as timeout, abort, and reset. Then, the controller of the storage apparatus returns a hardware error as a response to the access from the host when a minor failure of the number of times greater than or equal to a predetermined threshold occurs in the path being used for communication. The host that has received the hardware error switches the communication path to the other path.

JP 2003-208362 A JP 2006-107151 A

  Among a plurality of controllers that relay access from a host device to a storage unit and a storage device that includes a plurality of storage units, there are storage devices that have different access performance to the storage unit depending on the controller.

  However, the devices of Patent Literature 1 and Patent Literature 2 switch controllers regardless of the difference in access performance to the storage unit of each controller. Therefore, the devices of Patent Literature 1 and Patent Literature 2 can reduce a decrease in access performance when a controller that is accessed at the time of access to a storage device having different access performances of a plurality of controllers for a storage unit is changed due to a failure. Can not.

  It is an object of the present invention to provide a path control device that reduces a decrease in access performance when a storage device having different access performance of a plurality of controllers with respect to a storage unit is changed due to a failure. .

  The path control device of the present invention is connected to a storage unit and a host device, and each of a plurality of controllers that perform data transfer between the storage unit and the host device and a failure in communication between the host device Monitoring means for detecting the presence or absence, priority storage means for storing the priority of each of the controllers, selection means for selecting the controller having the highest priority among the controllers in which no communication failure is detected, Notification means for notifying an identifier of the selected controller to the host device that performs access to the storage unit via the controller notified of the identifier.

  According to the path control method of the present invention, a failure in communication between each of a plurality of controllers connected to a storage unit and a host device and performing data transfer between the storage unit and the host device, and the host device The priority of each of the controllers is stored in the priority storage means, and the controller having the highest priority is selected from the controllers in which no communication failure is detected, and the storage unit is accessed. Then, the identifier of the selected controller is notified to the host device that is transmitted via the controller that has been notified of the identifier.

  The path control program according to the present invention includes a computer connected between a storage unit and a host device, and each of a plurality of controllers that perform data transfer between the storage unit and the host device, and the host device. A selection unit that selects the controller having the highest priority among the monitoring unit that detects the presence or absence of a communication failure, the priority storage unit that stores the priority of each of the controllers, and the controller that does not detect a communication failure. And the host device that performs access to the storage unit via the controller that has been notified of the identifier is operated as a notification unit that notifies the identifier of the selected controller.

  The present invention has an effect that it is possible to reduce a decrease in access performance when a controller that is accessed at the time of access to a storage apparatus with different access performance of a plurality of controllers with respect to a storage unit is changed due to a failure.

FIG. 1 is a diagram illustrating a configuration of an information processing system 100 according to the first embodiment. FIG. 2 is a diagram illustrating an example of the configuration of the controller 20. FIG. 3 is a flowchart showing an example of the operation at the time of priority determination of the path control device 1 of the present embodiment of the first embodiment. FIG. 4 is a diagram illustrating an example of the priority order of each controller 20 stored in the priority order storage unit 11. FIG. 5 is a flowchart illustrating an example of an operation when a failure is detected in the path control device 1 according to the first embodiment. FIG. 6 is a diagram illustrating an example of a management table stored in the controller 20. FIG. 7 is a diagram illustrating an example of a management table stored in the controller 20. FIG. 8 is a diagram illustrating an example of a management table stored in the controller 20. FIG. 9 is a diagram illustrating an example of a management table stored in the controller 20. FIG. 10 is a diagram illustrating an example of a management table stored in the controller 20. FIG. 11 is a diagram illustrating the configuration of the path control device 1A according to the second embodiment.

  Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a diagram illustrating a configuration of an information processing system 100 according to the present embodiment.

  Referring to FIG. 1, the information processing system 100 includes a path control device 1, a storage system 2, and a host device 3. The information processing system 100 may include a maintenance terminal 5 connected to the storage system 2.

  The path control device 1 includes a monitoring unit 10, a priority order storage unit 11, a selection unit 12, a notification unit 13, and a priority order determination unit 14.

  The storage system 2 includes a plurality of controllers 20 and a storage device 22. The storage device 22 is composed of one or more hard disk drives and other physical storage devices. The storage device 22 includes a storage unit 21. The storage device 22 may include a plurality of storage units 21. The storage unit 21 may be the above-described physical storage device or a logical volume configured on one or more physical storage devices.

  Each controller 20 and each host device 3 are connected so that they can access each other. The controller 20 and the host device 3 may be directly connected. The controller 20 and the host device 3 may be connected via, for example, the switch device 4. A plurality of switch devices 4 may be connected between the controller 20 and the host device 3. The controller 20 and the host device 3 may be connected via a network.

  Each controller 20 and each storage unit 21 are connected so that the controller 20 can access each storage unit 21.

  The controller 20 relays data transfer between each host device 3 and each storage unit 21.

  In the present embodiment, the access performance for the storage unit 21 differs depending on the combination of the storage unit 21 and the controller 20 that is passed through when accessing the storage unit 21. The storage system 2 normally holds, for each storage unit 21, the identifier of the controller 20 having the highest access performance to the storage unit 21 as the optimum path to the storage unit 21. As will be described later, the path control device 1 sets an optimum route. Each host device 3 inquires of the storage system 2 about an optimum route for access to the storage unit 21. The storage system 2 transmits the identifier of the controller 20 set as the optimum path to the storage unit 21 to the inquiring host device 3. The host device 3 transmits the above inquiry to, for example, all the controllers 20 of the storage system 2. For example, the controller 20 set as the optimum route returns the identifier. The host device 3 accesses the storage unit 21 via the controller 20 that has received the identifier.

  The path control device 1 may be included in the storage system 2. The path control device 1 may be included in any or all of the controllers 20. When the path control device 1 is included in the plurality of controllers 20, the path control device 1 included in the predetermined controller 20 may perform the operation described later. In this case, for example, the other controller 20 monitors the presence or absence of a failure of the controller 20 including the operating path control device 1, and when a failure is detected, the path control device 1 included in the controller 20 that has detected the failure operates. Can be taken over. A method of selecting the controller 20 including the path control device 1 that takes over the operation may be arbitrary. For example, the order of taking over the operation of the path control device 1 may be set in each controller 20.

  The path control device 1 includes a monitoring unit 10, a priority order storage unit 11, a selection unit 12, a notification unit 13, and a priority order determination unit 14.

  The monitoring unit 10 detects, for each controller 20, the presence or absence of a communication failure between each of the plurality of controllers 20 that perform data transfer between the storage unit 21 and the host device 3 and the host device 3.

  The priority storage unit 11 stores the priority of each controller 20 for each storage unit 21.

  The selection unit 12 selects, for each storage unit 21, the controller 20 having the highest priority among the controllers 20 in which no communication failure is detected for each of the host devices 3.

  The notification unit 13 notifies the host device 3 to access the storage unit 21 via the selected controller 20.

  When a communication failure occurs between any of the host devices 3 and the controller 20, there is a risk of hindering data transfer between the other host devices 3 and the storage unit 21 via the controller 20. There is. Therefore, the selection unit 12 does not select the controller 20 in which a failure has been detected in communication with any of the host devices 3 as the controller 20 through which the storage unit 21 is accessed.

  The priority order determination unit 14 acquires the measured access performance from each of the controllers 20 that measure the access performance for each of the storage units 21. The priority order determination unit 14 determines the priority order of each controller 20 for each storage unit 21 so that the higher the acquired access performance, the higher the priority order. The priority order determination unit 14 associates the identifier of the storage unit 21 with the priority order determined for each controller 20 and stores them in the priority order storage unit 11. The priority order determination unit 14 may further transmit the priority order of each controller 20 for each storage unit 21 to each controller 20.

  FIG. 2 is a diagram illustrating an example of the configuration of the controller 20.

  Referring to FIG. 2, the controller 20 includes a first interface unit 201, a transfer unit 202, a second interface unit 203, a management table storage unit 204, and an access performance measurement unit 205.

  The first interface unit 201 transfers data between the transfer unit 202 and each host device 3.

  The transfer unit 202 relays data transfer between the first interface unit 201 and the second interface unit 203. Further, the transfer unit 202 detects whether there is a communication failure between the controller 20 and each host device 3. The transfer unit 202 may detect the presence or absence of a communication failure by detecting the presence or absence of a command transmitted from each host device 3 to the controller 20. For example, when a command from the host device 3 is detected within a predetermined time in the past, the transfer unit 202 determines that there is no communication failure between the host device 3 and the controller 20. For example, when a command from the host device 3 is not detected within a predetermined time in the past, the transfer unit 202 determines that there is a communication failure between the host device 3 and the controller 20. Then, the transfer unit 202 stores in the management table storage unit 204 the presence or absence of the detected communication failure for each host device 3. The command detected by the transfer unit 202 only needs to include commands such as an inquiry about an optimum path, data writing, and data reading.

  The second interface unit 203 transfers data between the transfer unit 202 and each storage unit 21.

  The management table storage unit 204 stores the presence / absence of a failure between the host device 3 and the controller 20 for each host device 3. Further, the management table storage unit 204 stores, for each host device 3, whether or not the storage system 2 is used by the host device 3. Further, the management table storage unit 204 stores, for each storage unit 21, whether the route passing through the controller 20 is an optimal access route or a non-optimal access route for the storage unit 21. When the controller 20 receives an inquiry about the optimum path from the host device 3 to the storage unit 21, for example, when the controller 20 itself is optimum, the controller 20 sends a response to the inquiring host device 3. The management table storage unit 204 may store the priority order of the controller 20 for each storage unit 21.

  The priority order stored in the management table storage unit 204 is input from the management terminal 5 by an administrator, for example. Alternatively, the priority order stored in the management table storage unit 204 is input by the notification unit 13 of the path control device 1, for example.

  The monitoring unit 10 described above detects a communication failure between the controller 20 and each host device 3 by, for example, reading the presence / absence of a failure corresponding to each host device 3 from the management table storage unit 204.

  The access performance measuring unit 205 measures the access performance between the controller 20 and the storage unit 21 for each storage unit 21. Then, the access performance measurement unit 205 transmits the measured access performance for each storage unit 21 to the priority order determination unit 14. The access performance may be an arbitrary index as long as it is an index representing the high access performance. The access performance is, for example, throughput. The access performance may be, for example, the number of input / output instruction processes per unit time. The access performance measurement unit 205 may store the measured access performance for each storage unit 21 in the management table storage unit 204.

  When the path control device 1 is included in the storage system 2 or the controller 20, the management table storage unit 204 may also serve as the priority order storage unit 11.

  Next, the operation of the path control device 1 at the time of priority determination will be described in detail with reference to the drawings.

  FIG. 3 is a flowchart showing an example of the operation of the path control device 1 at the time of priority determination.

  Referring to FIG. 3, first, the priority order determination unit 14 requests each controller 20 to transmit access performance to each storage unit 21 (step S101).

  Each controller 20 measures the access performance for each storage unit 21. Each controller 20 may read from the management table storage unit 204 the access performance for each storage unit 21 that has been measured in advance and stored in the management table storage unit 204. Each controller 20 transmits the access performance for each storage unit 21 to the priority order determination unit 14.

  The priority order determination unit 14 acquires the access performance for each storage unit 21 from each controller 20 (step S102).

  The priority order determination unit 14 determines the priority order of each controller 20 for each storage unit 21 (step S103). The priority order determination unit 14 determines the priority order so that the controller 20 having higher access performance to the storage unit 21 has a higher priority order.

  The priority order determination unit 14 stores the priority order of each controller 20 for each storage unit 21 in the priority order storage unit 11 (step S104).

  FIG. 4 is a diagram illustrating an example of the priority order of each controller 20 stored in the priority order storage unit 11.

  Next, the operation of the path control device 1 when a failure is detected will be described in detail with reference to the drawings.

  FIG. 5 is a flowchart illustrating an example of an operation when a failure is detected in the path control device 1 of the present embodiment.

  When the storage system 2 starts operation, the monitoring unit 10 detects a communication failure between each controller 20 and each host device 3 (step S201).

  As described above, each controller 20 may store the presence / absence of a command received from each host device 3 by the controller 20 within a predetermined time, for example, in the management table storage unit 204 as the presence / absence of a failure. The monitoring unit 10 may perform detection by obtaining from each controller 20 the presence or absence of a communication failure between the controller 20 and each host device 3. The monitoring unit 10 may read the presence / absence of a failure from the management table 204.

  FIG. 6 is a diagram illustrating an example of a management table stored in the management table storage unit 204 of the controller 20. FIG. 6 is a management table stored in the management table 204 of the controller 20 whose identifier is controller A, for example.

  The management table in FIG. 6 includes the presence / absence of a failure in communication with each host device 3. Further, the management table of FIG. 6 includes whether or not the storage system 2 is used by each host apparatus 3. In the example of FIG. 6, the failure is “present” in the communication between the controller 20 and the host apparatus 3 whose host identifier is the host m, which is an identifier for identifying the host apparatus 3. However, the path control device 1 and the controller 20 do not treat a failure detected for the host device 3 not using the storage system 2 as a communication failure. In the example of FIG. 6, the host m does not use the storage system 2. Therefore, in the example of FIG. 6, there is no host apparatus 3 that has a failure in communication with the controller 20.

  In the example of FIG. 6, the priority order of the controller A with respect to the storage unit a is 1. The controller A is set as an optimum path when the host device 3 accesses the storage unit a. The storage system 2 returns the identifier of the controller A in response to the inquiry about the optimum path for access to the storage unit a by the host device 3. The host device 3 that has received the reply accesses the storage unit a via the controller A.

  If there is no change in the controller 20 in which the failure has been detected (step S202, N), the process returns to step S201.

  7 to 10 are diagrams illustrating examples of management tables stored in the management table storage unit 204 of the controller 20. FIG. 7 is a management table stored in the management table 204 of the controller 20 whose identifier is controller A, for example. The priority of the controller A with respect to the storage unit a is 1. In the example of FIG. 7, unlike FIG. 6, there is a failure in communication between the host device 3 whose host identifier is the host 2 and the controller A. Controller A is selected as the optimum path for storage unit a.

  FIG. 8 is a management table stored in the management table 204 of the controller 20 whose identifier is controller B, for example. The priority order of the storage unit a of the controller B is 2. Controller B is not selected as the optimal path for storage unit a. The priority of the controller B with respect to the storage unit a is 1. Further, the controller B is selected as the optimum path for the storage unit a.

  In the present embodiment, the priority order of the storage units a of the other controllers 20 is lower than the priority order of the storage units a of the controllers A and B. The management table of the controller 20 of the controller A and the controller B is omitted.

  When the management tables stored in the controller A and the controller B are the examples of FIGS. 7 and 8, respectively, since there is a failure between the controller A and the host device 3 whose host identifier is the host 2, the monitoring unit 10 , Detect failure.

  For example, when the management table of the controller A changes from FIG. 6 to FIG. 7, a new failure is detected in the controller A. That is, the detection result of the communication failure between the host device 3 and the controller 20 changes (Y in step S202).

  In this case, the selection unit 12 selects, for each storage unit 21, the controller 20 having the highest priority among the controllers 20 in which no failure is detected with any host device 3 (step S203). The selection unit 12 may perform the operation of Step S203 when an instruction to change the access path to the storage unit 21 is transmitted from the host device 3 to the storage system 2. Such a change command is, for example, a SCSI command Set Target Port Groups. In this case, any controller 20 in the storage system 2 that has received the change command may notify the path control device 1 of the reception of the change command.

  In the example of FIGS. 7 and 8, the controller 20 that has the highest priority with respect to the storage unit a and has no failure detected with any host device 3 is the controller B. Accordingly, the selection unit 12 selects the controller B for the storage unit a. Similarly, the controller 20 having the highest priority with respect to the storage unit b and in which no failure is detected with any host apparatus 3 is the controller B. Accordingly, the selection unit 12 selects the controller B for the storage unit b.

  Next, the notification unit 13 transmits to each host device 3 an instruction to access the storage unit 21 through a path that passes through the controller 20 selected for each storage unit 21 (step S204).

  For example, the notification unit 13 may instruct each host device 3 of a route for each storage unit 21 by changing the management table of each controller 20. That is, for each storage unit 21, the notification unit 13 may set “optimal” in the management table of the selected controller 20 and set “non-optimal” in the management table of the controller 20 that is not selected. Each host device 3 inquires each controller 20 about an optimum route to the storage unit 21 periodically, for example. When the controller 20 for which “optimum” is set receives an inquiry about the optimum route to the storage unit 21, the controller 20 sends a reply that can specify the identifier of the own device, for example, to the host device 3 that sent the inquiry. For example, the host device 20 can identify the controller 20 for which “optimum” is set based on the address of the transmission source of the received response. For example, the response to be transmitted may include the identifier of the controller 20 for which “optimum” is set. The host device 20 accesses the storage unit 21 via the controller 20 specified by the response to the inquiry.

  For example, the notification unit 13 may directly transmit a set of the identifier of the storage unit 21 and the identifier of the controller 20 selected for the storage unit 21 to the host device 3 as the above-described instruction. When the host device 3 that has received the set of the identifier of the storage unit 21 and the identifier of the controller 20 accesses the storage unit 21 that includes the identifier in the received set, the host device 3 passes through the controller 20 that includes the identifier in the same set. Access should be done.

  In the example of FIGS. 7 and 8, a failure has occurred in the controller A that is selected as the optimum path for the storage unit a and has the highest priority for the storage unit a. The controller A is the controller 20 having the highest access performance to the storage unit a. On the other hand, the controller B having the priority order 2 for the storage unit a has no failure. The controller B is the controller 20 having the next highest access performance to the storage unit a after the controller A.

  The storage system 2 changes the management tables of the controller A and the controller B as shown in the examples of FIGS. 9 and 10 by the operations of step S203 and step S204, respectively.

  FIG. 9 is a management table stored in the management table 204 of the controller 20 whose identifier is controller A after rewriting by the notification unit 13. In FIG. 9, “non-optimal” is set for the storage unit a.

  FIG. 10 is a management table stored in the management table 204 of the controller 20 whose identifier is the controller B after rewriting by the notification unit 13. In FIG. 9, “optimum” is set for the storage unit a.

  In the examples of FIGS. 9 and 10, the controller B having the highest priority for the storage unit a among the controllers 20 having no failure for all the host devices 3 is notified as the optimum path for the storage unit a. . The controller B is the controller 20 having the highest access performance with respect to the storage unit a among the controllers 20 that do not have any trouble with respect to all the host devices 3. In other words, the controller with the highest priority for the storage unit a is the controller 20 having the optimal path to the storage unit a from the controller A having the failure to the host 2 to the controller 20 having no failure for all the host devices 3. B has been changed.

  The storage system 2 returns the identifier of the controller B in response to an inquiry from the host device 3 about the optimum path to the storage unit a. The host device 3 that has made the inquiry accesses the storage unit a via the controller B instead of the controller A.

  The optimum path for the storage unit b remains the controller B having the highest priority for the storage unit b and the highest access performance for the storage unit b.

  On the other hand, when the failure that has occurred in the communication between the controller A and the host 2 is resolved, the detection result by the monitoring unit 10 changes such that the failure is not detected in the communication between the controller A and the host 2 ( Step S202, Y).

  In this case, the selection unit 12 selects the controller A as the optimum path for the storage unit a in step S203. In step S204, the notification unit 13 notifies each host device 3 of the controller A as an optimum route to the storage unit a. That is, the notification unit 13 sets “optimum” for the storage unit a in the management table of the controller A. After the setting, when the storage system 2 receives an inquiry from the controller 20 via any of the host devices 3 when accessing the storage unit a, the controller A, for example, Notify to go through itself. The controller A performs this notification by transmitting the identifier of the controller A, for example. The host device 3 that has received the notification accesses the storage unit a via the controller A. The host device 3 that has accessed the storage unit a via the controller B before receiving the notification changes the controller 20 that is accessed via the access to the storage unit a to the controller A after receiving the notification.

  In the present embodiment described above, it is possible to reduce a decrease in access performance when the controller 20 through which access is made to a storage system 2 having different access performance to the storage units 21 of the plurality of controllers 20 is changed due to a failure. There is an effect that can be.

  The reason is that the selection unit 12 selects the controller 20 having the highest priority for the storage unit 21 among the controllers 20 that do not have any trouble in communication with all the host devices 3 for each storage unit 21. Because it does. The controller 20 selected for each storage unit 21 is the controller 20 having the highest access performance to the storage unit 21. Then, the notification unit 13 notifies each host device 3 of the controller 20 selected for each storage unit 21 as an optimum path for the storage unit 21.

(Configuration example)
Next, a configuration example of the first embodiment will be described in detail with reference to the drawings.

  In this configuration example, each host device 3 transmits a SCSI (Small Computer System Interface) command to the storage unit 21 (LUN, Logical Unit Number) via one of the controllers 20 to the storage unit 21. to access. In the information processing system 100 of this configuration example, as in the first embodiment, the performance of access to each storage unit 21 is determined according to the path that passes through the access, that is, the controller 20 that passes through at the time of access. Different for each. Such a configuration is referred to as an ALUA (Asymmetric Logical Unit Access) configuration. In a system that supports ALUA, path switching control between the host and the storage is performed by a SCSI command that controls ALUA. The path when the host device 3 accesses the LUN is specified by the controller 20 through which the path passes. In other words, the identifier of the controller 20 through which the LUN is accessed represents a path.

  When there are two controllers 20, a path that passes through the controller 20 having better access performance to the LUN is called a preferred path. In this case, the priority storage unit 11 stores which controller 20 passes the preferred path. In this case, a value indicating that the path is a preferred path and a value indicating that the path is not a preferred path correspond to the priority order. However, the priority order of the controller 20 whose path is the preferred path is higher than the priority order of the controller 20 whose path is not the preferred path. When there are three or more controllers 20, the priority storage unit 11 stores the priority of each controller 20 for each LUN. Even when there are two controllers 20, the priority order storage unit 11 may store the priority order of each controller 20 for each LUN. In this configuration example, the priority order stored in the priority order storage unit 11 is set by an administrator who uses the maintenance terminal 5.

  The path through which each host device 3 accesses the LUN is an Active / Optimized path. The Active / Optimized path is a path that passes through the controller 20 in which the above-described optimization is set. A path other than the Active / Optimized path is an Active / Non-Optimized path. The Active / Non-Optimized path is a path that passes through the controller 20 in which the above-mentioned non-optimal is set. When receiving Report Target Port Groups, which is an SCSI command for ALUA control, from the host device 3, the controller 20 transmits information on the Active / Optimized path to the host device 3. The information on the Active / Optimized path is, for example, an identifier of the controller 20 that passes through the Active / Optimized path. The response to the Report Target Port Groups command may be performed by the controller 20 via the Active / Optimized path, for example.

  Further, when a failure occurs in the path used for accessing the LUN and the connection is disconnected, the host device 3 transmits Set Target Port Groups, which are SCSI commands for ALUA control, to each controller 20. The controller 20 that has received Set Target Port Groups notifies the path control device 1 that it has received Set Target Port Groups. The path control apparatus 1 that has received this notification selects the controller 20 as in the first embodiment described above. The path control device 1 sets a path passing through the selected controller 20 as an Active / Optimized path. The path control device 1 sets a path passing through the controller 20 that has not been selected as an Active / Non-Optimized path.

  The transfer unit 202 of this configuration example detects whether or not a SCSI command has been received from the host device 3 within a predetermined time before storing the detection result in the management table storage unit 204, as will be described later. The transfer unit 202 stores, in the management table storage unit 204, whether or not a SCSI command is received from each host device 3 within a predetermined time, for example, every predetermined time. The management table storage unit 204 of this configuration example stores, in the management table, whether or not a SCSI command has been received within a predetermined time as the presence or absence of a communication failure between the host device 3 and the controller 20. When the transfer unit 202 receives the SCSI command within a predetermined time, the transfer unit 202 determines that there is no failure in communication between the host device 3 and the controller 20. If the transfer unit 202 does not receive the SCSI command within a predetermined time, the transfer unit 202 determines that there is a failure in communication between the host device 3 and the controller 20. Further, the host identifier stored in the management table by the management table storage unit 204 may be the WWn (World Wide name) of the host device 3. For example, the administrator of the information processing apparatus 100 uses the management terminal 5 to set whether to use the storage system 2 for each host device 3. The storage system 2 stores the set usage information in the management table of the management table storage unit 204. As in the first embodiment, when there is a host device 3 in which “not used” is set in “use / non-use” of the management table of the controller 20, the path control device 1 and the controller 20 and the host device 3 Communication is excluded from fault monitoring.

  Next, the operation of this configuration example will be described.

  A user or administrator of the information processing system 100 uses one of the host devices 3 or the maintenance terminal 5 to input whether or not the storage system 2 is used by each host device 3. For example, the storage system 2 stores, in association with the management table of the management table storage unit 204 of each controller 20, used or unused, which is the usage status input as WWn that is the identifier of the host device 3.

  Each controller 20 periodically detects whether or not each host device 3 has issued a SCSI command, and stores the detection result in the management table of the management table storage unit 204 in association with the WWn of the host device 3.

  The monitoring unit 10 of the path control device 1 monitors the management table of the management table storage unit 204 of each controller 20 and detects the host device 3 that has not issued the SCSI command within a predetermined time, thereby detecting the controller 20 and the host. A communication failure between the devices 3 is detected. However, the monitoring unit 10 excludes the host device 3 for which nonuse is set in the management table from the monitoring target. When the controller 20 receives a SCSI command from the monitored host device 3 within a predetermined time, the monitoring unit 10 indicates that the path between the controller 20 and the host device 3 is valid and there is no failure. judge.

  Each host device 3 issues a Report Target Port Groups command for each storage unit 21 to inquire about an optimal path for accessing the storage unit 21. When communication between each host device 3 and the storage system 2 is normal, the storage system 2 returns the identifier of the controller 20 with the highest priority as an Active / Optimized path in response to the Report Target Port Groups command. . That is, the storage system 2 returns the identifier of the controller 20 with the highest priority as the optimum path for accessing the storage unit 21. When the storage system 2 includes two controllers 20, the storage system 2 returns the identifier of the controller 20 set in the Preferred path. The host device 3 accesses the storage unit 21 via the controller 20 to which the identifier is returned.

  In the following description, the controller A is a preferred path of LUN 0 that is one of the storage units 21. Alternatively, the controller A is the controller 20 with the highest priority in access to LUN0. When a failure occurs in the controller A, the path between the controller A and each host device 3 is lost. In this case, any of the host apparatuses 3 that access the storage system 2 via the controller A detects the loss of the path that passes through the controller A. The host device 3 that has detected the loss of the path issues a Set Target Port Groups command to change the Active / Optimized path for access to LUN0. When the storage system 2 receives the Set Target Port Groups command, the path control device 1 has the highest priority for the active / optimized path for access to the LUN 0 among the controllers 20 having no failure as described above. Change to controller 20. When the Active / Optimized path for access to the LUN 0 is changed to the controller B, for example, the host device 3 continues to access the LUN 0 via the controller B. Thereafter, when the failure of the controller A recovers, each host device 3 can access the LUN 0 via either the controller A or the controller B.

  When the monitoring unit 10 detects that the communication failure between the controller A and each host device 3 has been resolved, the selection unit 12 determines, for each LUN, the controller 20 having the highest priority among the controllers 20 that have no failure. Select. Since the priority in the access to the LUN 0 of the controller A is the highest, the selection unit 12 selects the controller A for the LUN 0.

  When each controller 20 includes the path control device 1, for example, the selection unit 12 of the path control device 1 of each controller 20 may select the controller 20 described above. Then, the path control device 1 of the controller 20 that has selected itself may perform the following operation.

  The notification unit 13 rewrites the management table stored in the controller 20 selected by the selection unit 12 for a certain LUN, and sets the selected controller 20 to the Active / Optimized path for the LUN. Further, the notification unit 13 rewrites the management table stored in the controller 20 that has not been selected, and sets the controller 20 that has not been selected as an Active / Non-Optimized path for the LUN.

  As described above, when the failure of the controller A is recovered and the LUN 0 can be accessed through either the controller A or the controller B, the notification unit 13 sets the controller A to the active / optimized path of the LUN 0. To do. Then, the notification unit 13 sets the controller B to the Active / Non-Optimized path of LUN0.

  The host apparatus 3 that has issued the Report Target Port Groups command after these settings receives the identifier of the controller A as an Active / Optimized path for LUN0. The host device 3 that has received the identifier of the controller A accesses the LUN 0 via the controller A.

  Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 11 is a block diagram showing the configuration of the path control apparatus 1A of this embodiment.

  Referring to FIG. 11, the path control device 1 </ b> A of the present embodiment is connected to a storage unit 21 and a host device 3, and includes a plurality of controllers 20 that perform data transfer between the storage unit 21 and the host device 3. A monitoring unit 10 that detects whether there is a communication failure between the host device 3 and the host device 3, a priority storage unit 11 that stores the priority of each of the controllers 20, and the controller in which a communication failure is not detected 20, the selection unit 12 that selects the controller 20 having the highest priority and the host device 3 that performs access to the storage unit 21 via the controller 20 that is notified of the identifier are selected. And a notification unit 13 for notifying the identifier of the controller 20.

  The present embodiment described above has the same effect as the first embodiment.

  The path control device 1, the path control device 1A, and the storage system 2 are each realized by a computer and a program for controlling the computer, dedicated hardware, or a combination of a program for controlling the computer and the computer and dedicated hardware. Can do.

  The monitoring unit 10, the selection unit 12, the notification unit 13, the priority order determination unit 14, the first interface unit 201, the transfer unit 202, the second interface unit 203, and the access performance measurement unit 205, for example, a record that stores a program It can be realized by a dedicated program for realizing the function of each unit read from the medium into the memory and a processor that executes the program. The priority order storage unit 11 and the management table storage unit 204 can be realized by a memory or a hard disk device included in a computer. Alternatively, the monitoring unit 10, the priority order storage unit 11, the selection unit 12, the notification unit 13, the priority order determination unit 14, the first interface unit 201, the transfer unit 202, the second interface unit 203, the management table storage unit 204, Part or all of the access performance measurement unit 205 can be realized by a dedicated circuit that implements the function of each unit.

  The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1, 1A path control apparatus 2 Storage system 3 Host apparatus 4 Switch apparatus 5 Maintenance terminal 10 Monitoring part 11 Priority order memory | storage part 12 Selection part 13 Notification part 14 Priority order determination part 20 Controller 21 Storage unit 22 Storage apparatus 100 Information processing system 201 First interface unit 202 Transfer unit 203 Second interface unit 204 Management table storage unit 205 Access performance measurement unit

Claims (10)

A plurality of controllers connected to the storage unit and the host device for transferring data between the storage unit and the host device, and monitoring means for detecting the presence or absence of a communication failure between the host device and ,
Priority storage means for storing the priority set so as to be higher as the access performance of each of the controllers to the storage unit is higher ;
When the communication failure is detected, a selection unit that selects the controller having the highest priority among the plurality of controllers in which the communication failure is not detected;
A path control device comprising: a notification unit configured to notify the host device that performs access to the storage unit via the controller that has been notified of an identifier, the identifier of the selected controller. The plurality of controllers are connected to the plurality of host devices, respectively, and transfer data between the storage unit and the plurality of host devices,
The monitoring means detects the presence or absence of a communication failure between the controller and the plurality of host devices,
The path control device according to claim 1, wherein the selection unit selects the controller having the highest priority among the controllers in which no communication failure is detected for any of the plurality of host devices. Obtaining the measured access performance from each of the controllers that measure the access performance to the storage unit, and the higher the access performance obtained, the higher the priority. The path control device according to claim 1, further comprising: a priority order determining unit that stores the priority order determined for each controller in the priority order storage unit.   An information processing system including the storage unit, the host device, and the path control device according to claim 1. Each of a plurality of controllers connected to the storage unit and the host device for transferring data between the storage unit and the host device, and detecting the presence or absence of a communication failure between the host device,
Storing the priority order set so as to be higher as the access performance of each of the storage units of the controller is higher ;
If a failure of the communication is detected, among the plurality of the controllers failure of the communication is not detected, it selects the highest priority and the controller,
A path control method of notifying an identifier of the selected controller to the host device that performs access to the storage unit via the controller that is notified of the identifier. Each of the plurality of controllers connected to the plurality of host devices transfers data between the storage unit and the plurality of host devices,
Detecting the presence or absence of a communication failure between the controller and the plurality of host devices;
The path control method according to claim 5, wherein the controller having the highest priority is selected from the controllers in which no communication failure is detected for any of the plurality of host devices. Obtaining the measured access performance from each of the controllers that measure the access performance to the storage unit, and the higher the access performance obtained, the higher the priority. The path control method according to claim 5 or 6, wherein the priority order determined for each controller is stored in the priority order storage means. Computer
A plurality of controllers connected to the storage unit and the host device for transferring data between the storage unit and the host device, and monitoring means for detecting the presence or absence of a communication failure between the host device and ,
Priority storage means for storing the priority set so as to be higher as the access performance of each of the controllers to the storage unit is higher ;
When the communication failure is detected, a selection unit that selects the controller having the highest priority among the plurality of controllers in which the communication failure is not detected;
A path control program that causes the host device that performs access to the storage unit via the controller notified of the identifier to operate as a notification unit that notifies the identifier of the selected controller. Computer
Communication failure between each of the plurality of controllers and the plurality of host devices, each connected to the plurality of host devices and transferring data between the storage unit and the plurality of host devices. The monitoring means for detecting
The path control program according to claim 8, wherein the controller that detects no communication failure with any of the plurality of host devices is operated as the selection unit that selects the controller having the highest priority. Computer
Obtaining the measured access performance from each of the controllers that measure the access performance to the storage unit, and the higher the access performance obtained, the higher the priority. The path control program according to claim 8 or 9, wherein the path control program is operated as priority order determining means for storing the priority order determined for each controller in the priority order storage means.

Images