JP2005122763A - Storage device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To increase reliability and performance by distributing a load among a plurality of storage controllers in a redundant configuration, and realize live maintenance. <P>SOLUTION: A plurality of disk drive controllers 200 and 400 in a redundant configuration for controlling a disk device are connected to an upper device by the same SCSI ID, and a communication mechanism 300 and a common management table 310 are interposed for mutual monitoring of operating states and setting of load distribution information. In a normal state, the plurality of disk drive controllers 200 and 400 realize high performance by load distribution under concurrent operation. At a failure or maintenance, switching mechanisms 220 and 420 execute a switching operation for saving and recovery by isolating the failing side to realize uninterrupted operation and live maintenance. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an external storage device, and more particularly to a technique effective when applied to a redundant configuration external storage subsystem or the like having multiple input / output control devices that control information input / output requests from a host device.

  In the external storage device constituting the computer system, if the storage control device that is interposed between the storage device having the storage medium and the host device and controls the exchange of information between the two is not redundant, the storage control device When a failure occurs, the subsystem is forced to stop, and recovery work is performed during this time. When this recovery work is completed, the storage control device is restarted, or the subsystem that has been restarted is restarted.

Recently, in various information processing operations using a computer system, the operation mode of 24-hour operation is increasing, and the external storage subsystem is required to be continuously operated. For this reason, for example, as described in Patent Document 1, a redundant configuration is adopted for the storage control device in which one storage control device is in operation and the other storage control device is stopped and enters a standby state. There is known a technique for enabling continuous operation of a system by switching to a standby storage control device of another system when a storage control device fails.
Japanese Patent Laid-Open No. 3-206529

  However, in the above-described prior art, continuous operation at the time of failure is possible, but only one of them actually operates despite the fact that it has two storage control devices, and it is There was no difference from the time of one. In other words, the redundant other-system storage control device is only for hot standby, and is merely a substitute for the failed storage control device.

  In recent years, there are various requests to the system, and there are various connection modes in which access requests are issued from the host device to the same storage device or to separate storage devices through a plurality of paths. In the conventional redundant configuration of the storage control device, it is difficult to construct a system in accordance with various user requests.

  Conventionally, an inexpensive system has a configuration in which a storage controller and a data buffer are mounted on a single board. When performing maintenance management such as adding a data buffer in the storage controller, only the data buffer is used. Because it is impossible to disconnect, the data buffer is expanded while the system is temporarily stopped, and after the expansion work is completed, the storage controller and system are restarted, and the previously interrupted work is resumed. It is necessary to follow a procedure, and it has been impossible to perform maintenance management work such as expansion while processing input / output requests (I / O) from the host device.

  An object of the present invention is to provide an external storage device capable of improving reliability and performance by distributing a load to a plurality of redundant storage control devices.

  Another object of the present invention is to improve reliability by multiplexing storage control devices and realize various control operations of the storage control devices without making the host device side aware of the redundant configuration of the storage control devices. It is an object of the present invention to provide an external storage device that can be used.

  Still another object of the present invention is to provide an external storage device that can easily perform maintenance management work such as hardware and software in a plurality of storage control devices in a redundant configuration without stopping operation. It is in.

  Still another object of the present invention is to provide an external storage device capable of executing maintenance management work having a configuration in which a storage control device and a data buffer are mounted on a single board during operation.

  The external storage device of the present invention monitors the other storage control devices between the plurality of storage control devices and interface means for connecting the storage control devices to the higher order devices so that the plurality of storage control devices look the same when viewed from the higher order devices. Monitoring means, communication means capable of transmitting information between the storage control devices, switching means for switching the storage control device receiving a request from the host device, and the host device received by one storage control device And load distribution means for distributing the load of the processes associated therewith by a plurality of storage control devices.

  Also, a data buffer that is provided in each storage control device and temporarily stores write data from the host device by adopting the same redundant configuration as the storage control device, and a point in time when the write data from the host device is stored in the data buffer The data transfer means for reporting the end to the host device and writing from the data buffer to the storage device asynchronously with the request from the host device and controlling whether to write to all of the redundant data buffers or selectively It is set as the structure containing these.

  In addition, it is made commonly accessible from a plurality of storage control devices, and which of the first management information, write after processing, and write through processing for identifying whether each storage control device is healthy is executed. Second management information to be specified, third management information to specify which of the plurality of storage control devices accepts an input / output request from a higher-level device, and first to specify the load sharing in each of the plurality of storage control devices 4 includes management information storage means for storing at least one of the management information.

  In the external storage device of the present invention, for example, when it has a redundant configuration including a pair of first and second storage control devices, the first and second storage control devices are daisy chained by a host device and, for example, a SCSI interface. Connected and accessed with the same SCSIID. For example, when the first storage controller receives a fixed input / output request from the host device, the load accompanying the processing of the input / output request designates the load sharing in each of the plurality of storage controllers. Based on the fourth management information, it is distributed to other second storage control devices, improving the reliability by the redundant configuration and improving the processing capability of the input / output processing by the parallel operation of the first and second storage control devices. I can plan.

  Also, for example, when the first storage controller receives a fixed input / output request from the host device, when the first storage controller fails, the first management information and monitoring means By switching the storage control device that receives the request by the switching means to the second storage control device, the host device may issue an I / O request for the same SCSI ID even after a failure. There is no need to be conscious of anything. Thereafter, the faulty storage control device is disconnected and the degenerate operation is started. After maintenance work such as replacement of parts and microprograms is completed, the first storage controller is restored and restored to the original redundant configuration.

  Each storage controller has a data buffer. When the first storage controller receives write data from the host device and executes write-after processing, a failure occurs in the first storage controller. When the storage control device that receives the input / output request from the first storage control device to the second storage control device is switched by the switching means, the multiple data buffers are simultaneously multiplexed. The data writing process is switched to a process of selectively writing data to the data buffer provided in the operating storage control device.

  At this time, whether to execute the write after process or the write through process is selected. This selection can be made by the user setting the second management information in the management information storage means. That is, when the user's request for data reliability is high, the write-through mode is set. When the performance is required rather than the reliability, the write-after mode is set.

  After restoration of the second storage control device, switching to selective writing or multiple writing can be switched to multiple writing to the data buffer to restore the redundant configuration.

  The first storage controller receives an input / output request from the host device, and the write request from the host device is written twice in the data buffer of the first storage controller and the data buffer of the second storage controller. When write after processing is performed, switch from multiple writing processing to selective writing processing, disconnect the second storage controller, and perform maintenance such as adding data buffers or exchanging microprograms. After that, the original redundant configuration is restored. After that, the second storage control device notifies the first storage control device of the completion of maintenance work such as the addition of the data buffer by using communication means capable of transmitting information between the storage control devices, and switches after the notification. Using the means, the reception of the request from the host device is switched to the own device.

  On the other hand, the first storage control device that has received the notification degenerates itself and performs maintenance such as addition of a data buffer to restore it. After the restoration, the first storage control device notifies the second storage control device of the completion of maintenance such as the addition of the data buffer by using communication means capable of transmitting information. In response to this, switching from selective data writing to a single data buffer to multiple writing processing to a plurality of data buffers is performed. This makes it possible to perform maintenance management operations such as adding data buffers and exchanging microprograms for the pair of first storage control device and second storage control device while continuing input / output processing with the host device. Become.

  Further, according to the present invention, by referring to the third management information for designating which of the plurality of storage control devices set in the management information storage means accepts input / output requests from the host device. The first storage control device and the second storage control device can determine which one receives the request from the host device. Thereby, for example, not only one of the first storage control device and the second storage control device accepts a request from the host device, but both storage control devices accept and process input / output requests. It is also possible to do. Further, by arbitrarily setting the third management information by the user, the storage control device that receives a request from the host device can be arbitrarily designated by the user.

  According to the external storage device of the present invention, it is possible to improve the reliability and performance by distributing the load to a plurality of redundant storage control devices.

  In addition, it is possible to improve the reliability by multiplexing the storage control devices without realizing the redundant configuration of the storage control devices on the host device side, and to realize various control operations of the storage control devices. Is obtained.

  Further, it is possible to easily perform maintenance management work such as hardware and software in a plurality of redundantly configured storage control devices without stopping the operation.

  Further, there is an effect that maintenance management work having a configuration in which the storage control device and the data buffer are mounted on a single board can be executed during operation.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a conceptual diagram showing an example of a computer system including an external storage device according to an embodiment of the present invention. The computer system of the present embodiment includes a host device 100 that is a central processing unit, a disk drive control device 200, a disk drive control device 400, and a disk device 500. The disk drive control device 200 and the disk drive control device 400 are connected to the host device 100 through a daisy chain of SCSI interfaces, and the disk drive control devices 200 and 400 have the same SCSI ID and have a redundant configuration. In the case of the present embodiment, the disk drive control device 200 receives a request from the higher-level device 100, and executes a process associated with the request by the disk drive control device 200 and the redundant disk drive control device 400. The apparatus 500 is controlled.

  FIG. 2 is a block diagram showing an example of the internal configuration of the disk drive control devices 200 and 400. As shown in FIG. Since the internal configurations of the disk drive control device 200 and the disk drive control device 400 are the same, the disk drive control device 200 will be described as an example. The explanation is omitted with the same digit.

  The microprocessor unit 250 (hereinafter referred to as MPU) executes a random access memory (RAM, not shown) while sequentially decoding it, and controls the entire disk drive controller 200.

  The host I / F control unit 210 performs protocol control with the host device 100. The DRVI / F control unit 270 performs protocol control with each drive. The data buffer 240 is used when the host I / F control unit 210 and the DRVI / F control unit 270 transfer data. This memory may be a volatile memory or a non-volatile memory. In the present embodiment, a case where the data buffer 240 is constructed with a volatile memory will be described as an example.

  The switching mechanism 220 receives the I / O from the host device 100 to the host I / F control unit 210 and the host I / F control unit 410 of each disk drive control device 200 and the disk drive control device 400. This is for switching the F control unit. In this embodiment, it is assumed that the host I / F control unit 210 has received. The data transfer control unit 230 controls data transfer between the host device 100 and the data buffer 240. The data transfer control unit 230 has both functions of writing the write data from the host device 100 in two planes, the data buffer 240 and the data buffer 440, or writing only the data buffer 240 in a single manner. . Further, it is possible to switch between single writing and double writing according to an instruction from the MPU 250.

  The DRV transfer control unit 260 controls data transfer between the data buffer 240 and the disk device 500.

  The communication mechanism 300 is a mechanism for transmitting information between the MPU 250 and the MPU 450. This communication mechanism 300 enables bidirectional transmission between the MPU 250 and the MPU 450.

  The common management table 310 is a management table that can be referenced / updated from both the MPU 250 and the MPU 450.

  In the present embodiment, a drive storage method using an array configuration in which logical data from the host device 100 is distributed and stored in a plurality of disk devices 500 will be described as an example.

  The ECC generation circuit 280 has a function of generating redundant data for data sent from the host device 100, and this function can also be used for data restoration. The unit for adding redundant data may be one logical data unit sent from the host or a plurality of logical data units. In this embodiment, redundant data is added to four logical data, and description is made in a RAID 5 system in which the drive for storing the redundant data is not fixed.

  Next, an example of the configuration of the common management table 310 will be described with reference to FIG. The monitoring information 320 is used to check whether each disk drive control device 200/400 is operating normally. The monitoring information A321 sets information at regular intervals when it is determined that the MPU 250 of the disk drive control apparatus 200 can operate normally. Also, when the MPU 250 determines that it cannot normally operate, information indicating an abnormality is set. Note that the MPU 450 of the disk drive control device 400 also sets information in the monitoring information B 322 in the same manner as the MPU 250.

  The data transfer mode information 330 indicates an end report trigger for a write data write request from the higher-level device 100 when the system is in a degenerated state. That is, when this information is written to the data buffer 240 or the data buffer 440, completion is reported to the host device 100 (hereinafter referred to as a write-after mode) or when data is written from the data buffer 240 to the disk device 500. This is information for determining whether to report the end (hereinafter referred to as write-through mode).

  The host I / O reception information 340 indicates instruction information of a disk drive controller that receives I / O out of the two disk drive controllers 200/400. In the present embodiment, the disk drive control device 200 will be described as being set on the host I / O reception side.

  The load distribution information 350 is information for distributing the load associated with the I / O from the host device between the two disk drive control devices 200/400. The load distribution method may divide the disk device to be accessed into each disk drive control device, and the processing of the I / O request from the host device 100 and the I / O request from the host device 100 are asynchronous. It may be shared with the process of storing write data from the data buffer to the disk device 500. Alternatively, all the matters that must be processed may be written in the load distribution information, and the competition logic between the two MPUs may be used, and the processing may be executed when there is an empty MPU.

  In the present embodiment, the processing of the I / O request from the host device 100 and the processing of storing the write data from the data buffer 240/440 to the disk device 500 asynchronously with the I / O request from the host device 100 are shared. The method to do is demonstrated. Therefore, in the present embodiment, it is assumed that the load distribution information 350 contains information on the write data stored in the data buffer 240/440.

  Next, data write processing and data read processing from the host device 100 to the disk device 500 in the computer system according to the present embodiment will be described.

  The disk drive control device 200 normally receives write logical data from the host I / F control unit 210 at the time of a write request from the host device 100, and the data transfer control unit 230 stores two data buffers 240 and 440 in the data buffer 240. The storage information is stored in the load distribution information 350 of the common management table 310, and the end is reported to the host device 100 at this point. The MPU 450 sequentially refers to the load distribution information 350, and if there is stored information, the MPU 450 corresponds to the data already stored in the drive (hereinafter referred to as old data) having the same address as the write data and the write data. Parity data is read from the disk device 500 by the DRVI / F control unit 470 and the DRV transfer control unit 460, and the ECC data generation circuit 480 uses the write data, the old data, and the parity data as parity data corresponding to the write data (hereinafter referred to as the new parity). Data). The generated new parity data and write data are written into the disk device 500 by the DRVI / F control unit 470 and the DRV transfer control unit 460, thereby storing the write data in the disk device 500. This process is performed asynchronously with the I / O request from the host apparatus 100. The old data / old parity data read processing, new parity generation processing, and new parity data storage processing performed to store the write data are called write penalties in RAID5.

  As described above, the write data storage request from the host device 100 is a processing with a very high load when the plurality of disk devices 500 function as the disk array device. Executing this process by dividing the roles between the two disk drive control devices 200 and 400 is more efficient and improves the performance of the system than when executed by only one disk drive control device. In particular, as a recent market trend, it is very important to install an inexpensive processor and reduce the cost of the entire system together with high performance and high reliability. Therefore, in the write penalty processing, a large number of accesses to the drive may lead to performance degradation. However, since the running time of the microprogram of the processor that controls the drive before that is long, it becomes a processor neck as a system. There are also many. At this time, as in the present embodiment, by performing processing with the two disk drive control devices 200 and 400, nearly double performance can be achieved.

  Next, at the time of a read request from the host device 100, the MPU 250 starts reading data from the physical drive (disk device 500) by the DRVI / F control unit 270 and the DRV transfer control unit 260 and transfers the data to the host device 100. At this time, if the read request addresses from the host device 100 are continuous, the disk drive control device 400 determines that the sequential read processing is performed, and determines a certain amount of data following the read request address from the host device 100 as the host device. Processing to read data buffers 240 and 440 asynchronously with I / O from 100 may be performed. By doing so, the next time an I / O request is received from the host device, the target data is already stored in the data buffer 240/440, and the data does not cause time-consuming access to the disk device 500. Can improve the overall performance.

  As described above, the redundant portion (in this embodiment, the disk drive control device 400) is not simply put on standby for switching when a failure occurs, but a part of the processing is executed even though it is a redundant configuration. By doing so, not only reliability but also performance can be improved.

  Next, in the present embodiment, an operation in which the two disk drive control devices 200/400 execute automatic switching and recovery at the time of failure while executing processing will be described. First, a monitoring procedure for automatically detecting a failure will be described.

  The MPUs 250 and 450 control the disk drive control devices 200 and 400, and each time a certain time elapses, the MPU 250 is in the monitoring information 321 and the MPU 450 is normal in the monitoring information 322 (hereinafter referred to as normal information). ). However, in order to show that the information is set at regular intervals, information that changes sequentially is set in this information. For example, the information is added one by one. Further, if each MPU 250, 450 determines that the disk drive control device 200, 400 cannot normally operate, for example, when the data buffer becomes inaccessible from the MPU, the monitoring information is faulty. Information indicating this (hereinafter referred to as failure information) is set. Hereinafter, an example of the aforementioned monitoring procedure will be described with reference to the flowchart of FIG.

  Here, an operation in which the MPU 450 of the disk drive control device 400 monitors the other disk drive control device 200 will be described as an example.

  First, the MPU 250 determines in step 600 whether a certain time has elapsed. If the predetermined time has not elapsed, the process proceeds to step 608, and the disk drive control device 200 determines that it is normal.

  If the predetermined time has elapsed, the process proceeds to step 601 to set normal information indicating that the MPU 450 is normal. In step 602, the monitoring information 322 of the disk drive control device 200 is referred to. It is determined whether this information is normal. If it is determined in step 603 that the information is normal, the process proceeds to step 604. If it is determined that there is a failure, the process proceeds to step 607 and the disk drive control device 200 determines that there is a failure.

  If it is normal information, the process proceeds to step 605 and it is determined in step 605 whether or not this normal information has been changed. That is, the MPU 250 may be unable to set monitoring information due to a microprogram failure or the like. Such a failure is determined by checking this step 605. If there is a change, the process proceeds to step 608 and is determined to be normal. When there is no change, the process proceeds to step 606, and it is determined whether or not a margin time longer than a certain time has elapsed. As a result, if it has elapsed, the process proceeds to step 607, where it is determined as a failure, and if it has not elapsed, the process proceeds to step 608, where it is determined to be normal. According to the above monitoring procedure, both a hardware failure and a microprogram failure can be detected simultaneously.

  Next, an example of processing in which the disk drive control device 400 recognizes a failure of the other disk drive control device 200 and switches to it will be described with reference to the flowchart of FIG.

  First, the MPU 450 sequentially refers to the load distribution information 350 in step 700. As a result, if there is no write data from the host device 100 in the data buffers 240 and 440 in step 701, the process proceeds to step 704. If it exists, the process proceeds to step 702, and in order to generate parity corresponding to the write data in the data buffer 440, the old data and old parity data corresponding to the write data are read from the disk device 500, and the ECC generation circuit 480 generates a new one. Generate parity data. Thereafter, the process proceeds to step 703, where the write data and the new parity data are stored in the disk device 500 by the DRV transfer control unit 460 and the DRVI / F control unit 470. Next, at step 704, the failure of the disk drive control device 200 is checked by the monitoring procedure after step 600 in FIG. If the result is normal, the process proceeds to step 700 to continue the process. If it is determined that the switching is necessary, the process proceeds to step 710 and the reception of I / O from the higher-level device 100 is switched from the disk drive control device 200 to the disk drive control device 400 using the switching procedure. Then, in step 720, the disk drive control device 400 substitutes for the I / O processing from the host device 100 that was performed by the disk drive control device 200.

  Next, an example of the switching procedure will be described with reference to the flowchart of FIG.

  First, in step 711, when write data is received from the host device 100, the data transfer control unit 430 is instructed to write the data to the data buffer 440 in a single manner. That is, since a failure has occurred in the disk drive control device 200, the data buffer exists only in the disk drive control device 400 until the disk drive control device 200 is degenerated and cut off, the failed portion is replaced and restored. Therefore, the double writing as in the normal redundant configuration cannot be performed.

  In step 712, the switching mechanism 420 instructs the host I / F control unit 210 to switch the I / O request from the host device 100 to the host I / F control unit 410. As a result, the host I / F control unit 210 does not accept a request from the higher-level device 100, and the host I / F control unit 410 receives a request from the higher-level device 100, substantially controlling the disk drive. In this embodiment, since the SCSI ID is the same, the host device 100 may issue the same SCSI ID as before the I / O switching, and the receiving disk drive control device has been switched. There is no need to know.

  Next, an example of a procedure for performing I / O in the disk drive control device 400 after switching is described using the flowchart of FIG.

  When an I / O process is received from the host device 100 in step 721, the process proceeds to step 722, where it is determined whether the request is a read request or a write request. When the read request is made, the process proceeds to step 729, and the target data is read into the data buffer 440 from the disk device 500 corresponding to the read request. Proceeding to step 730, the data is transferred from the data buffer 440 to the host device 100, and the end is reported to the host device 100 at step 728.

  When a write request is made, the process proceeds to step 723 and the write data is stored in the data buffer 440. In step 724, the data transfer mode information 330 is referred to, and in step 725, it is determined whether or not the write-through mode is set. As a result, in the write after mode, that is, in the mode for reporting the end to the host device 100 when the data is stored in the data buffer 440, the process proceeds to step 728 to report the end, and then asynchronously to the data buffer 440. To the disk device 500. In the write-through mode, the process proceeds to step 726, where parity data for the write data is created, write data and new parity data are stored in the disk device 500 in step 727, and the end is reported in step 728. Further, thereafter, steps 700 to 703 in the flowchart of FIG. 5 are executed, and the process before the switching is also executed.

  As described above, according to the present embodiment, the disk drive control device 200/400 automatically switches between each other without instructing the host device 100 without an instruction from the host device 100. Operation and processing can continue.

  Next, an example of a method when the disk drive control apparatus 200 recovers and returns to the original redundant configuration will be described.

  First, an example of the recovery operation on the disk drive control device 200 side will be described with reference to the flowchart shown in FIG. In step 810, the communication mechanism 300 notifies the disk drive control device 400 that the restoration has been completed. Thereafter, the disk drive control device 200 becomes a redundant disk drive control device, and the position is changed from the previous one. In step 811, asynchronous destage processing (steps 700 to 705 in FIG. 5) previously performed by the disk drive control device 400 is performed.

  Further, an example of the operation of the disk drive control device 400 on the receiving side will be described with reference to the flowchart shown in FIG.

  When the communication mechanism 300 recognizes that the recovery of the disk drive control device 200 has been completed in step 820, the data transfer control unit 430 is instructed to perform double writing to the data buffers 240 and 440 in step 821, and in step 821, the host device. Only I / O processing from 100 is executed. In this way, it is possible to restore the original redundant configuration while receiving an I / O request from the host device 100, and further by distributing the processing load between the two disk drive control devices 200/400, The performance can also be improved.

  Next, an example of a data buffer expansion method during operation of the disk drive control apparatus 200/400 will be described with reference to the flowchart of FIG. The disk drive control device that receives I / O from the host device 100 is the disk drive control device 200.

  When there is a request for adding a data buffer, it is determined in step 911 whether the disk drive control device is on the I / O receiving side. First, processing contents of the disk drive control device 200 will be described. Since the disk drive control device 200 is on the I / O receiving side, the process proceeds to step 912, where it is recognized that the disk drive control device 400 is first degenerated and disconnected. Therefore, the data transfer control unit 230 is instructed to write the write data to the data buffer 240 as a single write. Thereafter, in step 913, I / O processing from the host device 100 is executed, and in step 914, steps 700 to 703 in FIG. 5 are executed. That is, the part executed by the disk drive control device 400 is substituted. While the steps 913 and 914 are repeated, the completion of the recovery of the disk drive control device 400 is awaited.

  Next, the disk drive controller 400 also determines in step 911 whether the disk drive controller is on the I / O reception side. As a result, since it is not the I / O receiving side, the process proceeds to step 915, the disk drive control device 400 is disconnected, and the data buffer 440 is expanded in step 916. After completion of the expansion, the disk drive control device 200 is notified via the communication mechanism 300 that it has been restored in step 917.

  Since the disk drive control device 200 needs to be expanded this time, the disk drive control device 400 uses the switching mechanism 420 in step 919 to replace the I / O reception. Switch the control unit to your system. Thereafter, in step 920, I / O processing from the host device 100 is executed, and in step 921, steps 700 to 703 in FIG. 5 are executed, and the completion of recovery of the disk drive control device 200 is awaited.

  The disk drive control device 200 that has recognized the recovery through the communication mechanism 300 in step 918 disconnects it in step 922, and adds the data buffer 240 in step 923. After completion of the expansion, the communication mechanism 300 notifies the disk drive control device 400 of the restoration in step 924. After the notification, since the disk drive control device 200 is not the host I / O receiving side, in step 925, the disk drive control device 200 goes to the side executing step 700 to step 705 in FIG.

  When the communication mechanism 300 recognizes the recovery of the other system in step 926, the disk drive control device 400 instructs the data transfer control unit 230 to write the write data to the data buffer 240/440 in step 927. To do. In step 928, I / O processing from the host device 100 is executed.

  In this way, it is possible to increase the number of data buffers 240/440 for each system while executing I / O from the host device 100. In other words, according to the present embodiment, the data buffer can be increased while online, whereas the data buffer can be increased only after the system is stopped. In particular, when the disk drive control device is constructed on one board realized at a low cost, replacement for each board is necessary, so that it is not possible to add it during operation. In the present embodiment, it is possible to increase the number of data buffers while degrading / recovering one disk at a time in the redundantly configured disk drive control apparatus 200/400.

  Further, according to the present embodiment, by replacing the processing of step 916 and step 923 of FIG. 10 with the microprogram replacement work, the operating microprogram can be replaced, and the demand for 24-hour operation is remarkable. This is particularly effective for maintenance management work in recent computer systems.

  Further, during degeneration at the time of a one-system failure, the user can instruct whether to write a write request from the host device 100 to the data buffer and report the end or to the disk device 500 to report the end. That is, the data transfer mode information 330 may be rewritten automatically by a user program. In other words, when the data buffer has a single-side configuration, if the end is reported at the time when the data buffer is stored, the responsiveness is excellent. However, if a failure occurs in the disk drive control device at this time, the data guarantee is guaranteed. become unable. On the other hand, when storing up to the disk device 500, write penalty processing occurs, and the responsiveness deteriorates considerably. However, a reliable response can be reported to the host device 100, and the reliability is high. high. In the case of the external storage device of the present embodiment, according to the request level of the reliability of the file handled by the user, it is possible to arbitrarily select whether to give priority to the reliability or the response speed according to the user instruction, It becomes possible to construct a flexible file system.

  Furthermore, the present invention can provide not only a redundant configuration of a plurality of disk drive control devices, but also a system that can be accessed simultaneously from a plurality of host devices or a plurality of buses. An example of this system configuration is shown in FIGS.

  11 has the same configuration as that of FIG. 1 of the embodiment described so far. However, when the I / F with the host device 100 is SCSI, the storage control device 0 and the storage control device 1 have the same SCSI ID in the configuration of FIG. 11 is different in that the storage control device 0 (400A) and the storage control device 1 (200A) are connected with different SCSI IDs. In the case of the configuration of FIG. 11, both receive and process an I / O request from the host device 100. FIG. 12 is a block diagram showing an example of a system configuration in which a plurality of storage control devices 0 (400B) and storage control devices 1 (200B) are connected to the same host device 100 by multipath. Also in the configuration of FIG. 12, both the storage control device 0 (400B) and the storage control device 1 (200B) can execute an I / O request from the host device 100. The designation of which I / O request is executed is realized by rewriting the host I / O reception information 340 in the common management table 310. That is, each storage control device first refers to the host I / O reception information 340 to determine whether or not the storage control device receives I / O from the host device. Thus, in the present invention, it is possible to cope with various user connection methods, and a flexible system can be constructed.

  As described above, according to this embodiment, the plurality of disk drive control devices 200/400 having a redundant configuration execute a request from the host device 100 while distributing the load, thereby improving not only the reliability. A file system capable of improving performance at the same time can be provided. Even if all disk drive control devices 200/400 execute I / O requests from the host device 100 while distributing the load, they automatically switch over without any instruction from the host device 100 when a failure occurs. It is possible to continue and continue to recover. This makes it possible to increase the data buffer and replace the microprogram while executing an I / O request from the host device 100, thereby realizing non-stop maintenance. Further, not only a redundant configuration but also a configuration in which all disk drive control devices simultaneously receive requests from the host device 100 can be flexibly adapted to various file systems requested by users. .

It is a conceptual diagram which shows an example of the computer system containing the external storage device which is one embodiment of this invention. 1 is a block diagram illustrating an example of an internal configuration of a disk drive control device that constitutes an external storage device according to an embodiment of the present invention. FIG. It is a conceptual diagram which shows an example of a structure of the common management table used in the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a flowchart which shows an example of an effect | action of the external storage device which is one embodiment of this invention. It is a conceptual diagram which shows the modification of the connection form with the high-order apparatus in the external storage device which is one embodiment of this invention. It is a conceptual diagram which shows the modification of the connection form with the high-order apparatus in the external storage device which is one embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 ... Host apparatus, 200 ... Disk drive control apparatus, 210 ... Host I / F control part, 220 ... Switching mechanism, 230 ... Data transfer control part, 240 ... Data buffer, 250 ... Microprocessor unit, 260 ... DRV transfer control part 270 ... DRVI / F control unit, 280 ... ECC generation circuit, 300 ... communication mechanism, 310 ... common management table, 320 ... monitoring information, 321 ... monitoring information, 322 ... monitoring information, 330 ... data transfer mode information, 340 ... Host I / O reception information, 350 ... load distribution information, 400 ... disk drive control device, 410 ... host I / F control unit, 420 ... switching mechanism, 430 ... data transfer control unit, 440 ... data buffer, 450 ... microprocessor Unit, 460... DRV transfer control unit, 470... DRVI / F control unit, 48. ... ECC generation circuit, 500 ... disk drive.

Claims (5)

A storage device that controls writing of data received from a host device to a disk device,
The storage device includes first and second disk control devices connected via a communication mechanism,
Each of the first and second disk control devices includes a host I / F control unit connected to the host device,
A data buffer for temporarily holding data received via the host I / F control unit;
A data transfer control unit that controls transfer of the data between the host I / F control unit and the data buffer;
A drive I / F control unit for controlling writing and reading of data held in the data buffer to the disk device;
A microprocessor for controlling the operation of each control unit,
The storage device is characterized in that the storage device holds a common management table to which both the microprocessors of the two disk control devices refer. The storage device according to claim 1.
The data transfer control unit of the first disk control device sends the data received by the host I / F control unit of the first disk control device to the data buffers of the first and second disk control devices. A storage device characterized by writing. The storage device according to claim 1 or 2,
Information for performing load distribution between the first and second disk control devices is held in the common management table, and the second disk control is performed based on an instruction from the microprocessor of the second disk control device. The drive I / F control unit of the device writes the write data written in the data buffer in the second disk control device by the data transfer control unit of the first disk control device to the disk control device. A storage device. A storage device that controls writing of data received from a host device to a disk device,
The storage device includes first and second disk control devices connected via a communication mechanism,
Each of the first and second disk control devices includes a host I / F control unit connected to the host device,
A data buffer for temporarily holding data received via the host I / F control unit;
A data transfer control unit that controls transfer of the data between the host I / F control unit and the data buffer;
A drive I / F control unit for controlling writing and reading of data held in the data buffer to the disk device;
A microprocessor for controlling the operation of each control unit,
The storage device holds a common management table to which the microprocessors of the two disk control devices refer;
The microprocessor transfers the data received by the host I / F control unit in the first disk control unit to the first and second disk by the data transfer control unit in the first disk control unit. A storage device for controlling whether data is written to the data buffer in the control device or only to the data buffer in the first disk control device. A storage device that controls writing of data received from a host device to a disk device,
The storage device includes first and second disk control devices connected via a communication mechanism,
The first disk control device includes a first host I / F control unit connected to the host device,
A first data buffer for temporarily holding data received via the first host I / F control unit;
A first data transfer control unit that controls transfer of the data between the first host I / F control unit and the first data buffer;
A first drive I / F control unit that controls writing and reading of data held in the first data buffer to and from the disk device;
A first microprocessor for controlling the operation of each control unit in the first disk control device,
The second disk control device includes a second host I / F control unit connected to the host device,
A second data buffer for temporarily holding data received via the second host I / F control unit;
A second data transfer control unit that controls transfer of the data between the second host I / F control unit and the second data buffer;
A second drive I / F control unit for controlling writing and reading of data held in the second data buffer to and from the disk device;
A second microprocessor for controlling the operation of each control unit in the second disk control device,
The first data transfer control unit writes the data received from the host device only in the first data buffer according to an instruction from the first microprocessor, or the first data buffer and the second data buffer A storage device in which whether the data is written to both data buffers is controlled.

Images