JP4904942B2 - Disk control system, disk control device, and disk control method - Google Patents

Description

  The present invention relates to a disk control technique for performing data check (hereinafter referred to as “verify”) of a storage device such as a magnetic disk.

  In recent years, RAID (Redundant Arrays of) is a technique for improving access speed by combining a plurality of HDDs (hard disk drives) to be used as a single HDD as a disk array or a technique for multiplexing data to improve fault tolerance. Independent Disks) technology is employed.

  For example, in the conventional verification method in such RAID technology, once the verification is started, the processing cannot be stopped until the verification is completed, and a constant load is continuously applied to the system until the verification is completed. For this reason, the CPU usage rate is increased, and the processing speed of other processes may be slowed down or temporarily hung up (see, for example, Patent Document 1).

As a method for solving this problem, a CPU load control processing technique (for example, refer to Patent Document 2) that stops the verify process when the CPU usage rate exceeds a certain level, or an area for performing the verify is subdivided. There has been an area division processing technique (see, for example, Patent Document 3) that shortens the processing time per one verify command.
JP-A-6-75703 JP 2000-284976 A JP-A-6-324893

  However, in the conventional CPU load control process, the verification process is temporarily stopped when the CPU usage rate exceeds a predetermined value. Therefore, when the process with a high CPU usage rate continues during operation, the time during which the verification process is stopped is stopped. In some cases, the verification process could not be performed until the end due to the length.

  This problem has become even more important with the recent increase in capacity of HDDs. In other words, as the capacity of the HDD increases, the system operation ends before the verification is completed, and at the next startup, the verification is performed again from the first address, and the verification cannot be completed indefinitely. There was a problem.

  Also, in another area division process, if a process with a high CPU usage rate continues during operation, it may compete with and affect other processes. Similarly, with the recent increase in capacity of HDDs, operation will not be completed before verification is completed, and verification will be performed again from the first address at the next startup, and verification may be completed indefinitely. The problem of being unable to do so occurred.

  The present invention employs the following configuration in order to solve the above-described problems. That is, according to the present invention, in a disk control system that performs verification processing of a storage device, a parameter storage unit that stores at least verified information is provided, and verification processing is performed based on information in the parameter storage unit.

  Alternatively, in another invention, a usage rate measuring unit that measures a usage rate of a CPU or the like, and a parameter storage unit that stores a determination threshold value such as a CPU usage rate, the measurement result by the usage rate measuring unit, and the determination threshold value are provided. Based on the above, the frequency of performing the verify process was changed.

  As described above, according to the disk control system of the present invention, in the disk control system that performs the verification process of the storage device, the parameter storage unit that stores at least the verified information is provided, and based on the information in the parameter storage unit Thus, the verify process is performed, so that the influence on other processes can be minimized, and the verify process can be reliably completed without the need for performing a duplicate verify process.

  Alternatively, a usage rate measuring unit that measures the usage rate of a CPU or the like and a parameter storage unit that stores a determination threshold value such as a CPU usage rate are provided, and verification is performed based on the measurement result by the usage rate measuring unit and the determination threshold value. Since the frequency of processing is changed, the influence on other processing can be further minimized, and the verification processing can be performed reliably.

  Embodiments of the present invention will be described below with reference to the drawings. In the following description, a disk array using RAID technology will be described as an example. However, any system that performs verification of a storage device such as a magnetic disk in the background may be used. It can also be applied to.

  The disk control system according to the first embodiment divides the storage area to be verified into the set number of divisions, performs verification, and stores the verified area number. In the next verification, the verification is performed from an unverified area. It is a thing.

(Constitution)
FIG. 1 is a diagram illustrating the configuration of the disk control system according to the first embodiment. As shown in the figure, in a file system using RAID technology, striping (RAID0) for distributed writing for speeding up, mirroring (RAID1) for improving reliability, and a combination of these ( RAID levels such as RAID 1 + 0) are supported, and are controlled by the RAID controller 1 and the disk controller 7.

  For example, in the case of RAID 1 + 0 control, the two HDDs 2 and 3 obtained by dividing the entire storage area by the number of divisions arbitrarily set as will be described in detail later pass through the disk controller 7 that controls them. The host device 10 is connected to the RAID controller 1, the higher-level operation system (hereinafter referred to as “OS6”), and the higher-level application (hereinafter referred to as “AP”).

  The RAID controller 1 is provided with a parameter storage unit 4 for storing disk control parameters. The division number 4a is used as a value for dividing the entire storage area, and the area number 4b that has been verified last. Can be stored. The configuration in which the parameter storage unit 4 is provided in the RAID controller 1 will be described below as an example. However, the configuration may be such that the AP (5) and the OS (6), the RAID controller 1 and the like are stored. The contents of 4 may be described in a registry file and stored in the HDDs 2 and 3.

(Operation)
With the above configuration, the disk control system of the first embodiment operates as follows. This operation will be described below with reference to the operation flowcharts of FIGS.

  First, as shown in FIG. 2, before the start of verification, the division number 4a and the executed area number 4b of the parameter storage unit 4 are read (step S1), and the total capacity to be verified is divided by the division number 4a (step S2). A start area is determined based on the executed area number 4b (step S3), and verification is executed (step S4). For example, as shown in FIG. 1, when the verification is completed up to # 002, the next verification is performed from # 003.

  Then, referring to FIG. 3, when the operation of this system is finished, it is determined whether or not the verification is completed (step S5). When the verification of the new area is completed, the previous verification is performed. The executed area number 4b which is a record at the time of the update is updated. In this case, in the next verify, the verify is executed from the area next to the newly completed verify.

  On the other hand, when the verification is not completed in step S5, the executed area number 4b is not updated. In this case, the next verification is performed from the same area as this time.

  In the above description, the last number of the executed area is stored, and in the next verification, it is described that the process is performed from the area next to the executed area. However, all the executed area numbers are stored. In the next verification, verification may be performed for an area arbitrarily selected from unexecuted areas.

(Effect of Example 1)
As described above, according to the disk control system of the first embodiment, since the verification is performed by dividing the storage area to be verified into the set number of divisions, the verified area information is stored. The time per one time can be shortened and the influence on other processing can be minimized, and in the next verification, verification can be performed from the next area that has been executed.

  The disk control system according to the second embodiment stores a verified address, and in the next verification, the verification is performed from the next of the stored verified address.

(Constitution)
FIG. 4 is a diagram illustrating the configuration of the disk control system according to the second embodiment. In the disk control system according to the second embodiment, the last-performed address 4c can be stored in the parameter storage unit 4 of the RAID controller 4. Other configurations are the same as the configuration of the disk control system according to the first embodiment, and the description thereof is omitted for the sake of brevity.

(Operation)
With the above configuration, the disk control system of the second embodiment operates as follows. This operation will be described below with reference to the operation flowcharts of FIGS.

  First, as shown in FIG. 5, before the verification is started, the executed address 4c of the parameter storage unit 4 which is the address at which the verification is finally performed is read (step S10), and the verification start address is determined from the executed address 4c. (Step S11), verify is executed (Step S12). For example, as shown in FIG. 4, when the executed address 4c is “1234AB (H)”, the verify start address is “1234AC (H)” which is the next address.

  Then, referring to FIG. 6, when the operation of this system is finished, it is determined whether or not the verification of all the storage areas to be verified is completed (step S13). The completed address 4c is updated to the head address, that is, 0 (H). In this case, in the next verification, verification is newly executed from the head address 0 (H).

  On the other hand, when the verification of all the storage areas to be verified is not completed, the address that has been verified last is overwritten and updated to the performed address 4c. In this case, the next verification is performed from the next of the completed address 4c.

(Effect of Example 2)
As described above, according to the disk control system of the second embodiment, since the last verified address is stored as the verification start address, the next verification is performed from the next address after verification. Even in a system that cannot take a long time for verification, the entire area to be verified can be verified.

  In the disk control system of the third embodiment, an interval for performing verification can be arbitrarily set.

(Constitution)
FIG. 7 is a diagram illustrating the configuration of the disk control system according to the third embodiment. In the disk control system according to the third embodiment, the verification execution interval 4d representing the verification execution interval (time) can be stored in the parameter storage unit 4 of the RAID controller 4. Other configurations are the same as the configuration of the disk control system according to the first embodiment, and the description thereof is omitted for the sake of brevity.

(Operation)
With the above configuration, the disk control system according to the third embodiment operates as follows. This operation will be described below using the operation flowchart of FIG.

  First, before the verification is started, the verification execution interval 4d stored in the parameter storage unit 4 is read (step S20), and the verification interval (time) is determined from the verification execution interval 4d (step S21), and the verification is executed. (Step S22).

  Then, the verification is repeated at the verification execution interval 4d (time), and the verification is completed together with the operation of the entire system.

(Effect of Example 3)
As described above, according to the disk control system of the third embodiment, the interval for performing the verification can be arbitrarily set in advance, so that the influence of the verification on other processes can be reduced as much as possible.

  In the disk control system of the fourth embodiment, the verification start time can be arbitrarily set.

(Constitution)
FIG. 9 is a diagram illustrating the configuration of the disk control system according to the fourth embodiment. In the disk control system according to the fourth embodiment, the verification start time 4e representing the verification start time (start time) can be stored in the parameter storage unit 4 of the RAID controller 4. Other configurations are the same as the configuration of the disk control system according to the first embodiment, and the description thereof is omitted for the sake of brevity.

(Operation)
With the above configuration, the disk control system according to the fourth embodiment operates as follows. This operation will be described below using the operation flowchart of FIG.

  First, before the start of verification, the verification start time 4e stored in the parameter storage unit 4 is read (step S30), the time to start verification is set from the verification start time 4e (step S31), and the verification start time 4e is reached. (Step S32) The verify is executed (Step S32). The verification start time may be monitored by a real time timer, and verification may be performed by interrupt processing when the set time is reached.

  Then, the verification is repeated, the verification is repeated, and the verification is completed together with the operation of the entire system.

(Effect of Example 4)
As described above, according to the disk control system of the fourth embodiment, the verification start time can be arbitrarily set. Therefore, the verification start time is set in consideration of other processes such as initial setting after the system is started. By setting, the influence on other processing can be minimized.

  In the disk control system according to the fifth embodiment, the frequency of verification is changed in accordance with the CPU usage rate.

(Constitution)
FIG. 11 is a diagram illustrating the configuration of the disk control system according to the fifth embodiment. In the disk control system of the fifth embodiment, the parameter storage unit 4 of the RAID controller 4 stores in the CPU-High (4f), which is the upper limit value of the CPU usage rate to stop verification, and the lower limit value of the CPU usage rate, which extends the verification execution interval. CPU-Low (4g) can be stored. Other configurations are the same as the configuration of the disk control system according to the first embodiment, and the description thereof is omitted for the sake of brevity.

  The CPU usage rate can usually be obtained by making an inquiry to the OS (6). For example, when the OS (6) is Windows (Windows is a registered trademark of Microsoft Corporation, USA), the CPU usage rate can be measured using the PerformanceCounter class.

(Operation)
With the above configuration, the disk control system of the fifth embodiment operates as follows. This operation will be described below using the operation flowchart of FIG.

  First, the parameter values CPU-High (4f) and CPU-Low (4g) are read before the verification is started (step S40), and after the verification is started (step S41), the CPU usage rate α is measured by the method described above (step S40). In step S42, the following processing is performed according to the CPU usage rate α (step S43).

  That is, when the CPU usage rate α is equal to or higher than the upper limit value CPU-High (4f), since the CPU is in the Busy state, the verification is temporarily stopped (step S45), and after a predetermined time has elapsed (step S46), again, The CPU usage rate α is measured (step S47), and when it is less than the upper limit value CPU-High (4f) by about Zch% and is not in the busy state (step S48), the verification is restarted (step S49). Note that Zch, which is the determination threshold for the busy state, is usually preferably about 10% to 20%.

  Then, after a predetermined time has passed, it is determined whether or not the verification has been completed (steps S55 and S56). If the verification has not been completed, the process returns to step S42, and the CPU usage rate α is measured. Repeat the process.

  On the other hand, when the CPU usage rate α is equal to or lower than the lower limit value CPU-Low (4 g) in step S43, the CPU is in an empty state, so the verification execution interval is shortened by a predetermined time and the verification is continued (step S50). -S51). Here, as a method of shortening the verification execution interval, the verification execution interval may be subtracted by a predetermined amount or may be divided at a constant rate.

  Then, the CPU usage rate α is periodically measured (step S52). When the CPU usage rate α exceeds the lower limit value CPU-Low (4g) by about Zcl% (step S53), the busy state is being reached. It returns to the initial value (step S54), and after a predetermined time has passed, it is determined whether or not the verification is completed (steps S55 and S56). If the verification is not completed, the process returns to step S42 and the CPU usage rate α is set. Measure and perform the subsequent processing in the same manner.

  Note that Zcl, which is the determination threshold value in step 53, is usually about 10% to 20%, similar to the determination threshold value in step S48 described above.

  On the other hand, if the CPU usage rate α is smaller than the upper limit value CPU-High (4f) and larger than the lower limit value CPU-Low (4g) in step S43, verification is continued because the CPU is in a normal state ( In step S44), it is determined whether or not the verification has been completed after a lapse of a predetermined time. If not, the process returns to step S42, the CPU usage rate α is measured, and the subsequent processing is similarly performed (step S55). S56).

  In the case where the CPU is in the Busy state, it has been described that the verification is temporarily stopped in step S45. However, the verification execution interval may be lengthened instead of temporarily stopping. Further, when the CPU is changing from the idle state to the Busy state, it has been described in step S54 that the verification execution interval is returned to the initial value and lengthened. However, the verification execution interval is increased by a predetermined amount or rate. Also good.

(Effect of Example 5)
As described above, according to the disk control system of the fifth embodiment, the state of the CPU is determined based on the CPU usage rate, and in the Busy state, the verification is interrupted to wait for the CPU usage rate to decrease, Since the verify interval is shortened in the empty state, the influence of the verify on other processing can be minimized.

  In the disk control system according to the sixth embodiment, the frequency of verification is changed according to the usage rate of the disk control device.

(Constitution)
FIG. 13 is a diagram illustrating the configuration of the disk control system according to the sixth embodiment. In the disk control system according to the sixth embodiment, the disk I / O-High (4h), which is the upper limit value of the disk controller that stops the verification, is stored in the parameter storage unit 4 of the RAID controller 4 and the verification execution interval is extended. Disk I / O-Low (4i), which is the lower limit value of the usage rate of the control device, can be stored. Other configurations are the same as the configuration of the disk control system according to the first embodiment, and the description thereof is omitted for the sake of brevity.

  Note that the usage rate of the disk control device can be obtained by making an inquiry to the OS (6), similarly to the usage rate of the CPU. For example, when the OS (6) is Windows (Windows is a registered trademark of Microsoft Corporation, USA), the usage rate of the disk control device can be measured using the PerformanceCounter class.

(Operation)
With the above configuration, the disk control system according to the sixth embodiment operates as follows. This operation will be described below using the operation flowchart of FIG.

  First, the parameter values DiskI / O-High (4h) and DiskI / O-Low (4i) are read before the verification is started (step S60). After the verification is started (step S61), the usage rate β of the disk controller 7 is measured. (Step S62), the following processing is performed according to the usage rate β of the disk controller 7 (Step S63).

  That is, when the usage rate β of the disk control device 7 is equal to or greater than the upper limit value DiskI / O-High (4h), the disk control device 7 is in the Busy state, so verification is temporarily stopped (step S65), and a predetermined time is set. After the elapse (step S66), the usage rate β of the disk controller 7 is measured again (step S67), and when it is about Zdh% below the upper limit value DiskI / O-High (4h) and the busy state is not reached (step S67). S68) The verify is resumed (step S69). Here, Zch, which is the determination threshold value for the busy state, is usually about 10% to 20%, as in the fifth embodiment.

  Then, after a predetermined time has passed, it is determined whether or not the verification has been completed (steps S75 and S76). If the verification has not been completed, the process returns to step S62, and the usage rate β of the disk controller 7 is measured. The same processing as described above is repeatedly performed.

  On the other hand, if the usage rate β of the disk control device 7 is equal to or lower than the lower limit value DiskI / O-Low (4i) in step S63, the disk control device 7 is in an empty state, so the verification execution interval is set to a predetermined amount. The verification is continued after shortening (steps S70 and S71). Here, as a method of shortening the verification execution interval, a predetermined amount may be subtracted, or it may be decreased by dividing at a constant rate.

  Then, the usage rate β of the disk controller 7 is periodically measured (step S72), and when it exceeds the lower limit value DiskI / O-Low (4i) by Zdl% or more (step S73), the busy state is being entered. Therefore, the verification execution interval is returned to the initial value (step S74), and after a predetermined time has passed, it is determined whether the verification has been completed (steps S75 and S76). If the verification has not been completed, the process returns to step S62. Then, the usage rate β of the disk controller 7 is measured, and the subsequent processing is similarly performed.

  Note that Zdl, which is the determination threshold value in step 73, is preferably about 10% to 20%, similar to the determination threshold value in step S68.

  On the other hand, if the usage rate β of the disk controller 7 is smaller than the upper limit value DiskI / O-High (4h) and larger than the lower limit value DiskI / O-Low (4i) in step S63, it is in a normal state. Therefore, the verification is continued (step S64). After a predetermined time elapses, it is determined whether the verification is completed. If the verification is not completed, the process returns to step S62 to measure the usage rate β of the disk controller 7 and thereafter. Processing is performed in the same manner (steps S75 and S76).

  Note that, in the case where the disk control device 7 is in the Busy state, it has been described that the verification is temporarily stopped in Step S65. However, instead of temporarily stopping, the verification execution interval may be lengthened. Further, in the case where the disk control device 7 is changing from the idle state to the Busy state, the verification execution interval is returned to the initial value in step S74. However, the verification execution interval is increased by a predetermined amount or rate. May be.

(Effect of Example 6)
As described above, according to the disk control system of the sixth embodiment, the state of the disk control device is determined based on the usage rate of the disk control device 7. When the status is Busy, the verification is interrupted and the usage rate of the disk control device decreases. On the other hand, since the verification interval is shortened in the empty state, the influence of the verification on other processing can be minimized.

<< Other modifications >>
In the above description of the examples, each embodiment has been described as being implemented independently, but may be implemented in combination with each other.

  For example, the embodiment 1 and the embodiment 5 are combined, the storage area of the storage device is divided into predetermined areas, the verification is executed, and the executed area number is stored. The verification is performed from the next area, and it is determined by the CPU usage rate whether the CPU is in the busy state. If the CPU is in the busy state, the frequency of verification is decreased, and if it is empty, the frequency of verification is increased. You may do it.

  As described above, the present invention can be widely used in disk control systems that perform verification processing of magnetic disks and the like.

1 is a configuration diagram of a disk control system of Embodiment 1. FIG. 3 is an operation flowchart of the disk control system according to the first embodiment. 3 is an operation flowchart of the disk control system according to the first embodiment. FIG. 6 is a configuration diagram of a disk control system according to a second embodiment. 6 is an operation flowchart of the disk control system according to the second embodiment. 6 is an operation flowchart of the disk control system according to the second embodiment. FIG. 6 is a configuration diagram of a disk control system according to a third embodiment. 10 is an operation flowchart of the disk control system according to the third embodiment. FIG. 10 is a configuration diagram of a disk control system according to a fourth embodiment. 10 is an operation flowchart of the disk control system according to the fourth embodiment. FIG. 10 is a configuration diagram of a disk control system according to a fifth embodiment. 10 is an operation flowchart of the disk control system according to the fifth embodiment. FIG. 10 is a configuration diagram of a disk control system according to a sixth embodiment. 10 is an operation flowchart of a disk control system according to a sixth embodiment.

Explanation of symbols

1 RAID controller 2, 3 HDD
4 Parameter storage 6 OS (operation system)
7 Disk controller

Claims (1)

A measurement unit for measuring the usage rate of the CPU;
A parameter storage unit for storing the determination threshold consisting upper and lower limits of the utilization provided,
The disk control device characterized in that the verification is stopped when the usage rate measured by the measurement unit is larger than the upper limit value, and the verification execution interval is shortened when the usage rate is smaller than the lower limit value. .

Images