JP2008257638A - Magnetic disk system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a magnetic disk system, capable of attaining reduced power consumption by stopping a RAID group without stopping on-line operation. <P>SOLUTION: The system comprises a RAID group composed of a large number of small-capacity drives, and a RAID group composed of a small number of large-capacity drives, in which the RAID group composed of a large number of small-capacity drives is operated in a time zone with high access frequency, and the RAID group composed of a small number of large-capacity drives is operated in a time zone with low access frequency. The operating RAID group of the RAID group composed of the small-capacity drives and the RAID group composed of the large-capacity drives can regularly hold latest data by copying the respective differential data at the time of changing the operation thereof. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a magnetic disk system composed of a plurality of RAIDs, and more particularly to a magnetic disk system that stops some RAIDs.

  2. Description of the Related Art In recent years, MAID (Massive Arrays of Inactive Disks) systems have become widespread as disk systems that can reduce costs while having high speed and large capacity. This consists of a number of RAID (Redundant Array of Independent Disks) groups consisting of a small number of magnetic disks, and the rotation of magnetic disks without data access is stopped by controlling the rotation of the magnetic disks in units of RAID groups. By trying to rotate the magnetic disk only when this occurs, it is intended to bring out cost merit.

  However, in order to access a magnetic disk that has once stopped rotating and is in a standby state in an accessible active state (normal state), it takes time in seconds, which caused a significant drop in access performance. .

A magnetic disk device (Patent Document 1) has been proposed in which access performance is improved by returning to normal from standby. In the magnetic disk device described in Patent Document 1, a slow rotation speed at the time of data reading and a high rotation speed at the time of data writing are set as the rotation speed of the disk, and this can improve the access performance. Yes.
JP 2004-362768 A

  However, since the magnetic disk device described in Patent Document 1 eventually stops the magnetic disk, there is a problem that it cannot be applied to a user who handles online work that cannot be stopped.

  Also, users who perform online work often place logical devices that hold frequently accessed data on physical devices in order to place importance on performance, and rotate the magnetic disk only for RAID groups that are accessed less frequently. It was also difficult to apply the MAID system for stopping the operation.

  The present invention has been made in view of such circumstances, and by stopping the RAID group without stopping online, it is possible to realize a long life of the magnetic disk with low power consumption and response speed. It is an object of the present invention to provide a magnetic disk system in which there is no decrease in the above.

  The problems of the present invention can be solved by the following inventions.

  That is, a magnetic disk system of the present invention includes a magnetic disk storage device including a plurality of magnetic disk groups constituting a RAID (Redundant Array of Independent Disks), and a control device that controls the magnetic disk storage device. In the system, the magnetic disk storage device is powered on in a set first time zone and is turned off in other time zones, and a set second magnetic disk group. And a second magnetic disk group that is powered off during the other time periods, and the number of the second magnetic disk groups is smaller than the number of the first magnetic disk groups. The control device is configured to store storage areas of the first magnetic disk group and the second magnetic disk group. A copy pointer unit, which is an address indicating the position of a track that is a unit, and into which an address of a track that is a data copy target is written between the first magnetic disk group and the second magnetic disk group; A copy difference management table section in which information indicating whether there is a difference in data between the first magnetic disk group and the second magnetic disk group is written for each track; An on / off timer management table section in which setting information of one time zone and setting information of the set second time zone are written; and a magnetic disk constituting the magnetic disk storage device includes the first magnetic disk A pair management table portion in which information indicating which group belongs to which of the first magnetic disk group, the first magnetic disk group, and the second magnetic disk group is written. The main feature is that each of the two magnetic disk groups includes a RAID group state management table portion in which information indicating whether or not each is in a standby state is written.

  Thereby, in a time zone with low access frequency, the power of the RAID group having a large number of magnetic disks can be turned off and the RAID group having a small number of magnetic disks can be operated, so that power consumption can be reduced. In addition, since the amount of heat generated from the magnetic disk is reduced, the air conditioning load is reduced and the power cost required for air conditioning can be reduced.

Furthermore, since any RAID group is always in operation during a high access frequency period or a low access time period, there is no need to stop online and the access speed does not decrease.
Furthermore, since the magnetic disk is stopped, the life of the magnetic disk can be increased, and since there is no frequent spin-up and spin-down, the mean time interval MTBF (Mean Time Between Failure) can be improved. .

  In the magnetic disk system of the present invention, the number of magnetic disks constituting the second magnetic disk group is less than half of the number of magnetic disks constituting the first magnetic disk group. It is said.

  Further, in the magnetic disk system of the present invention, the addresses of all tracks to be copied from the first magnetic disk group to the second magnetic disk group are sequentially written in the copy pointer section, and the copy pointer section When copying the track indicated by the address from the first magnetic disk group to the second magnetic disk group in order for each written address, in the track indicated by the address written in the copy pointer section, From the first magnetic disk group only when information indicating that there is a difference in data between the first magnetic disk group and the second magnetic disk group is written in the copy difference management table section. The main feature is that the track is copied to the second magnetic disk group.

  As a result, the difference data can be copied in a short time between a RAID group operating in a time zone with high access frequency and a RAID group operating in a time zone with a low access frequency.

  Furthermore, when there is an access for reading or writing data from the host while copying data from the first magnetic disk group to the second magnetic disk group,

  If the address of the accessed track has already been written in the copy pointer section, the host is made to access the second magnetic disk group, and if the access is a write, the copy difference management table If the information indicating that there is a data difference in the accessed track is written in the section, and the address of the accessed track is not yet written in the copy pointer section, the host is set to the first The main feature is that when the access is made to the magnetic disk group and the access is writing, information indicating that there is a data difference in the accessed track is written in the copy difference management table section. .

As a result, even if there is an access request from the host during the differential data copy between the RAID group that operates during a high access frequency period and the RAID group that operates during a low access frequency period, the access performance is not degraded. Enable access. In addition, since which of the two RAID groups is accessed is automatically switched, the host does not need to be aware of which one is accessed.

As described above, according to the magnetic disk system of the present invention, the operation time zones of the two RAID groups can be set separately.
Thereby, in a time zone with low access frequency, the power of the RAID group having a large number of magnetic disks can be turned off and the RAID group having a small number of magnetic disks can be operated, so that power consumption can be reduced.
In addition, since the amount of heat generated from the magnetic disk is reduced, the air conditioning load is reduced and the power cost required for air conditioning can be reduced.

In addition, since either RAID group is always in operation during a high access frequency period or a low access time period, there is no need to stop online and the access speed does not decrease.
Furthermore, since the magnetic disk is stopped, the life of the magnetic disk can be increased, and since there is no frequent spin-up and spin-down, the mean time interval MTBF (Mean Time Between Failure) can be improved. .

  In a data copy between a RAID group operating in a high access frequency period and a RAID group operating in a low access frequency period, only the data difference is copied, so the copy can be performed in a short time. It becomes possible.

In addition, even if there is an access request from the host during a data copy between a RAID group operating during a high access frequency period and a RAID group operating during a low access frequency period, the host is conscious of It is possible to switch the access to the two RAID groups without causing the access performance to deteriorate.

  The best mode for carrying out the magnetic disk system of the present invention will be described below in detail with reference to the accompanying drawings.

<Configuration>
FIG. 1 is a schematic view showing an embodiment of a magnetic disk system of the present invention. The magnetic disk system of the present invention includes a host 11, a PC 12, a control device 20, and a magnetic disk storage device 30.

The host 11 and the PC 12 are connected to the control device 20. The control device 20 is connected to the magnetic disk storage device 30. The host 11 writes data to the magnetic disk storage device 30 or reads data from the magnetic disk storage device 30 via the control device 20.
The PC 12 performs various settings of the control device 20. The control device 20 controls the magnetic disk storage device 30.

  The control device 20 includes a cache memory 21, a shared memory 23, and a control unit 23A. User data 22 is temporarily stored in the cache memory 21. That is, when the host 11 writes data to the magnetic disk storage device 30, the user data is temporarily stored in the cache memory 21 and written from the cache memory 21 to the magnetic disk storage device 30. When the host 11 reads data from the magnetic disk storage device 30, user data from the magnetic disk storage device 30 is temporarily stored in the cache memory 21, and is passed to the host 11 from there. At this time, the delivery of the user data 22 between the cache memory 21 and the magnetic disk storage device 30 is controlled by the control unit 23A.

  In the shared memory 23, a copy pointer 24, a copy difference management table 25, an on / off timer management table 26, a pair management table 27, and a RAID group state management table 28 are created.

  The copy pointer 24 is written with information indicating up to which track the difference data was copied when a difference volume is copied between a logical volume (primary) 32 and a logical volume (secondary) 34 described later. In the copy difference management table 25, information for each track indicating whether or not a difference has occurred between the logical volume (primary) 32 and the logical volume (secondary) 34 is written.

  The on / off timer management table 26 includes information indicating whether or not the logical volume (primary) 32 and the logical volume (secondary) 34 are powered on, and the RAID group (primary) 31 is the target of power on / off. Information indicating whether or not there is information, and information regarding the time when the power is turned on and the time when the power is turned off are written. The on / off timer management table 26 includes an after-spin-up access monitoring timer 26A. In the access monitoring timer 26A after re-spin-up, a set time until the power is turned off again after re-spin-up of a logical volume (primary) 32 described later is written.

In the pair management table 27, information on the arrangement and configuration of the logical volume (primary) 32 and the logical volume (secondary) 34 and information on the volume to be turned on / off are written. In the RAID group status management table 28, each RAID group constituting the logical volume (primary) 32 and the logical volume (secondary) 34 is in normal operation, during transition from normal to standby, and during transition from standby to normal. Information indicating that one of the devices is on standby is written.
The control unit 23A controls a physical volume 37, which will be described later, based on the contents of the cache memory 21 and the shared memory 23.

  The magnetic disk storage device 30 includes a RAID group (primary) 31 and a RAID group (secondary) 33. Each of the RAID group (primary) 31 and the RAID group (secondary) 33 is composed of physical volumes 37 that are physical hard disks. For example, the RAID group (primary) 31 includes two 72 GB hard disks. It can be configured with 14 sets ((2D + 2D) × 14 = 56 units) configured with RAID 0 + 1. The RAID group (secondary) 33 can be composed of eight 300 GB hard disks in RAID5. Of course, the present invention is not limited to this, and a RAID group configuration corresponding to the required capacity can be adopted.

  In this way, by using a hard disk with a large capacity for the RAID group (secondary) 33, the number of hard disks can be greatly reduced while keeping the overall capacity substantially the same as that of the RAID group (primary). Thereby, when the RAID group (secondary) 33 is used, the power cost can be suppressed, and the heat generated from the hard disk can be reduced, so that the power cost of indoor air conditioning can be suppressed. These power cost reduction effects can be obtained sufficiently if the number of hard disks in the RAID group (secondary) is less than half the number of hard disks in the RAID group (primary). If the number of hard disks in this group is 1/7 or less of the number of hard disks in the RAID group (primary), extremely great effects can be obtained.

The RAID group (primary) 31 includes a logical volume (primary) 32 and a non-copy target logical volume 35. The logical volume (primary) 32 is copied to the logical volume (secondary) 34 of the RAID group (secondary) 33 and mirrored. The non-copy target logical volume 35 is a logical volume that is not copied to the logical volume (secondary) 34.
The RAID group (secondary) is composed of logical volumes (secondary) 34.

<Operation>
The operation of one embodiment of the present invention will be described with reference to FIG.
The PC 12 creates information of the pair management table 27 and the on / off timer management table 26, accesses the shared memory 23, and writes it in the respective tables.

  When the writing to the table is completed, the PC 12 instructs the control unit 23A to create a pair, that is, to create a logical volume (secondary) 34 in which the contents of the logical volume (primary) 32 are copied. The control unit 23A controls the magnetic disk storage device 30 based on the contents of the pair management table 27, and creates a logical volume (secondary) 34 obtained by copying the logical volume (primary) 32 of the RAID group (primary) 31 as a RAID group. (Sub) 33.

  Next, the control unit 23A refers to the on / off timer management table 26, reads the information of the time when the RAID group (primary) 31 is turned off and the RAID group (secondary) 33 is turned on, and the time comes. When the RAID group (secondary) 33 is turned on, the hard disk spins up.

  When the spin-up is completed, the control unit 23A causes the logical volume (primary) 32 to be copied to the logical volume (secondary) 34. Copying is performed as follows. That is, the head address of the logical volume (primary) 32 is written into the copy pointer 24. Next, the information in the copy difference management table 25 corresponding to the address written in the copy pointer 24 indicates that there is a difference in data between the logical volume (primary) 32 and the logical volume (secondary) 34. If the difference bit is on (hereinafter referred to as the case where the difference bit is on), the track corresponding to the address is copied from the logical volume (primary) 32 to the logical volume (secondary) 34 and the copy difference management table 25 is logically copied. Information indicating that there is no difference in data between the volume (primary) 32 and the logical volume (secondary) 34 is written (hereinafter referred to as turning off the difference bit). After copying is completed, the value of the copy pointer 24 is rewritten to the value of the next address.

  When the information in the copy difference management table 25 indicates that there is no difference in data between the logical volume (primary) 32 and the logical volume (secondary) 34 (hereinafter referred to as a case where the difference bit is off). ) Is not copied, and the value of the copy pointer 24 is rewritten to the value of the next address.

  In this way, the first to last addresses of the logical volume (primary) 32 are written to the copy pointer 24, and the tracks corresponding to the written addresses are copied in order.

If a request for reading or writing user data is received from the host 11 during the copying operation, the following processing is performed.
When a read or write request is made for a track whose address has already been written to the copy pointer (hereinafter referred to as a track after the copy pointer has passed), the logical volume (secondary) 34 is used for reading. In the case of reading and writing data, information indicating that a difference has occurred is written in the copy difference management table 25 (hereinafter referred to as turning on the difference bit), and the data is written in the logical volume (secondary) 34.

  When a read or write request is made for a track whose address has not yet been written to the copy pointer (hereinafter referred to as a track before passing through the copy pointer), the logical volume (primary) 32 is used for reading. In the case of reading and writing data, the differential bit is turned on and the data is written to the logical volume (primary) 32.

  Thus, information indicating whether or not there is a difference in data between the track of the logical volume (primary) 32 and the corresponding track of the logical volume (secondary) 34 is written in the copy difference management table 25. Therefore, when copying data between the (primary) and (secondary) logical volumes, by referring to the copy difference management table 25, it is possible to copy only tracks with differences in data. Compared with the case of copying, copying can be performed in a short time.

  When the copy of the logical volume (primary) 32 to the logical volume (secondary) 34 is completed, the control unit 23A turns off the power of the logical volume (primary). As a result, the power supply of the (primary) 32 of the logical volume having a large number of hard disks can be turned off and operated only with the logical volume (secondary) 34 having a small number of hard disks, so that the life of the hard disk is reached. Can save electricity costs. Also, if the number of hard disks is small, the amount of heat generation is also reduced, so that the load of air conditioning in the room is reduced, and the electrical cost required for air conditioning can be reduced.

  Subsequently, the control unit 23A refers to the on / off timer management table 26, reads the information of the time when the power of the RAID group (secondary) 33 is turned off and the power of the RAID group (primary) 31 is turned on. Comes on, the RAID group (primary) 31 is turned on and the hard disk spins up.

  When the spin-up is completed, the control unit 23A executes a differential copy from the (secondary) 34 of the logical volume to the logical volume (primary) 32. This is because the logical volume (primary) 32 and the logical volume in the above-described copy are logically copied. The same operation as replacing the (secondary) 34 of the volume is performed.

  In this way, only the RAID group (primary) 31 is operated during the time zone when the RAID group is frequently accessed, and only the RAID group (secondary) 33 is activated during the low time zone, only the setting of the on / off timer management table. Can be realized easily.

<Create logical volume primary / secondary pair>
Next, creation of primary and secondary pairs of logical volumes will be described in detail with reference to FIG. FIG. 2 is a flowchart of logical volume pair creation.
The PC 12 creates information on the arrangement of the logical volume (primary) 32 and the logical volume (secondary) 34 and information on the logical volume to be turned on / off (S200).

  Next, the PC 12 determines whether or not the hard disk constituting the RAID group is powered on, information indicating whether or not the RAID group (primary) 31 is the target of power on / off, and the time when the power is turned on. And information on the setting of the time to turn off the power (S210).

  Thereafter, the PC 12 instructs the control unit 23A to start creating a logical volume (secondary) 34 that is a pair of the logical volume (primary) 32 and in which the contents of the logical volume (primary) 32 are copied (S220).

The PC 12 communicates with the control unit 23A of the control device 20, and the control unit 23A communicates with the magnetic disk storage device 30 in response to an instruction from the PC 12 to control the magnetic disk storage device 30 (S230).
The control unit 23A checks whether the RAID group (secondary) 33 is spinning up (S240). If the RAID group (secondary) 33 is not spinning up, the RAID group (secondary) 33 is powered on to spin up (S250) and spinning up. In this case, S260 is performed.

  The control unit 23A reads the information created by the PC 12 in S210 and S200 from the PC 12 (S260), writes the information created in S210 to the pair management table 27 (S270), and writes the information created in S200 to the on / off timer management table 26. Write (S280).

In addition, the control unit 23A causes the magnetic disk storage device 30 to perform a full differential copy of the logical volume (primary) 32 and the logical volume (secondary) 34 instructed to start pair creation from the PC 12 (S290).
The full difference copy is performed as follows. All the difference bits in the copy difference management table 25 are turned on. Next, the logical address (primary) 32 is sequentially written from the head address to the final address in the copy pointer 24, and the track at the written address is copied to the logical volume (secondary) 34 and corresponding to the copy difference management table 25. The difference bit is turned off. In this way, all tracks from the start address to the end address are copied.

  When there is an access from the host 11 during the entire differential copy, the following processing is performed. When a read access is made to the track before passing through the copy pointer, it is read from the logical volume (primary) 32.

When a write access is made to a track before passing through the copy pointer, it is written to the logical volume (primary) 32. When a read access is made to the track after passing the copy pointer, the track is read from the logical volume (primary) 32. When a write access is made to the track after passing the copy pointer, the difference bit of the copy difference management table 25 is turned on and written to the logical volume (primary) 32.
After all the difference copies are completed, the track for which the difference bit of the copy difference management table 25 is turned on is copied from the logical volume (primary) 32 to the logical volume (secondary) 34, and the difference bit is turned off.

Next, when pair creation is completed normally, the next S310 is performed, and when it is not completed normally, S290 is performed again (S300).
The controller 23A spins down the RAID group (secondary) 33 by turning off the power (S310).

  The PC 12 ends the process when pair creation is completed normally, and reports an error (S330) and ends the process when pair creation is not completed normally (S320).

<Periodic switching of logical volume (primary) and (secondary)>
Next, switching between the logical volume (primary) and the logical volume (secondary) will be described in detail with reference to FIG. FIG. 3 is a switching flowchart of logical volumes (primary) and (secondary).

  The control unit 23A refers to the on / off timer management table 26, and determines whether or not the logical volume (primary) is a target to be switched between power on and off, that is, whether or not the on / off function is set to invalid. If it is set as invalid, the process jumps to S36 and ends the process (S30).

When the power is turned on / off, that is, when the on / off function is set to be valid, the next step S31 is executed.
Next, the control unit 23A refers to the on / off timer management table 26 and determines whether it is time to turn off the power. If it is time to turn off the power, the flow of FIG. 4 described later is performed, and if it is not time to turn off the power, the next S32 is performed (S31).

  The control unit 23A refers to the RAID group state management table 28, and when the operation state of the RAID group (primary) is displayed as standby transient or standby, that is, when the RAID group (primary) is spinning down. If there is an access from the host 11, the flow of FIG. 5 to be described later is performed, and if not, S33 is performed (S32).

  When the flow shown in FIG. 5 is executed, that is, when the RAID group (primary) is re-spun up, the control unit 23A refers to the access monitoring timer (26A) after re-spin-up and monitors access after re-spin-up. It is determined whether or not the time set in the timer (26A) has elapsed. If it has elapsed, the flow of FIG. 6 described later is performed to spin down the RAID group (positive) (S33).

  If not applicable to S33, the control unit 23A refers to the on / off timer management table 26 to determine whether it is time to turn on the power. If the time is to turn on the power, the control unit 23A executes the flow of FIG. If it is not time to insert, the next S35 is performed (S34).

The control unit 23A has passed the time to turn on the RAID group (primary) indicated in the on / off timer management table 26, and the status of the RAID group (positive) indicated in the RAID group management table 28 is normal. If not, the flow of FIG. 7 to be described later is performed, and if not, the process ends in S36.
From S36, it jumps to S30 again and repeats similarly.

<Data update to the RAID group (secondary) side and spindown of the RAID group (primary)>
Next, update of data from the RAID group (primary) 31 to the RAID group (secondary) 33 and spin-down of the RAID group (primary) 31 will be described with reference to FIG. FIG. 4 is a flowchart of data update of the RAID group (secondary) 33 and spin-down of the RAID group (primary) 31.

  The controller 23A turns on the power of the RAID group (secondary) 33 to spin it up (S41). Next, the control unit 23A causes the logical volume (primary) 32 to copy the difference data from the logical volume (secondary) 34 (S42).

If the differential copy has been completed normally, that is, if the pair creation has been completed normally, the next step is performed. If not, the process returns to S42 (S43).
The control unit 23A changes the status display of the RAID group (primary) 31 in the RAID group status management table 28 to standby (S44), and powers down the RAID group (primary) 31 to spin it down (S45).
When S45 ends, the process is executed from S32 of the flow of FIG.

<Process when there is access during RAID group (primary) spindown>
Next, processing when the RAID group (primary) 31 is accessed from the host 11 during spin down will be described with reference to FIG. FIG. 5 is a flow diagram of operations during RAID group (primary) 31 spin down.

  The control unit 23A determines whether or not the logical volume requested to be accessed from the host 11 is a copy target to the logical volume (secondary) 34. If it is not a copy target, S55 is executed (S51).

  If it is a copy target, it is determined whether or not the access from the host 11 is a write. If the access is a write, S53 is executed, and if it is not a write, S54 is executed (S52).

If the access is a write, the control unit 23A performs writing from the host to the logical volume (secondary) 34, turns on the corresponding difference bit in the copy difference management table 26 (S53), and becomes End.
If the access is not a write in S52, that is, if it is a read, the control unit 23A performs a read process from the logical volume (secondary) 34 (S54) and becomes End.

  If it is determined in S51 that the logical volume (secondary) 34 is not to be copied, the control unit 23A determines whether or not the non-copy target logical volume 35 is being spun up, and is spinning up. In the case of (5), S58 is executed, and if it is not spinning up, S56 is executed (S55).

  If not, the RAID group (primary) 31 including the non-copy target logical volume 35 is spun up (S56). After the spin-up, the status of the corresponding RAID group (primary) 31 in the RAID group status management table 28 is changed to normal (S57).

After the change is completed, the control unit 23A accesses the non-copy target logical volume 35 (S58). After the access is completed, the control unit 23A sets the access monitoring timer 26A after the re-spin up and becomes End.
After performing up to End, execution is performed from S33 in the flowchart of FIG.

<Spin down of non-copy target logical volume 35>
Next, spin down of the non-copy target logical volume 35 will be described with reference to FIG. FIG. 6 is a flow chart of the spin-down of the non-copy target logical volume 35.

The control unit 23A spins down the RAID group (primary) 31 including the non-copy target logical volume 35 (S62).
Next, the control unit 23A changes the state of the corresponding RAID group (primary) 31 in the RAID group state management table 28 to standby (S63).
After the change, execution is performed from S34 in the flowchart of FIG.

<Data update to the RAID group (primary) side and spindown on the RAID group (secondary) side>
Next, update of data from the RAID group (secondary) 33 to the RAID group (primary) 31 and spin down of the RAID group (secondary) 33 will be described with reference to FIG. FIG. 7 is a flowchart of data update of the RAID group (primary) 31 and spin-down of the RAID group (secondary) 33.

  The control unit 23A refers to the RAID group status management table 28 to check whether the RAID group (primary) status is standby, and executes S72 if it is standby, and executes S77 if it is not standby. (S71).

In the case of standby, the control unit 23A turns on the RAID group (primary) 31 and spins it up (S72).
After the spin-up, the control unit 23A changes the contents of the RAID group state management table 28 from standby to standby transition (S73).

  Next, the control unit 23A checks whether or not the differential copy from the RAID group (secondary) 33 to the RAID group (primary) has been completed. If the copy has been completed, the control unit 23A executes S75 to complete the copying. If not, the difference bit of the copy difference management table 26 is checked, and only the track for which the difference bit is turned on is copied from the logical volume (secondary) 34 to the logical volume (primary) 32 (S78) End and Become.

  If the copying is complete, the contents of the RAID group status management table 28 are changed normally (S75), the RAID group (secondary) 33 is turned off and spun down (S76) to End.

In S71, if it is not standby, it is checked whether or not the contents of the RAID state management table are in a transition to standby. If transition is in progress, the process is executed from S74, and if it is not in transition, that is, normal.
In the flow chart of FIG. 3, when it is End, when it jumps from S34 to this flow, it executes from S35, and when it jumps from S35 to this flow, it jumps to S36.

<Initial copy from logical volume (primary) to logical volume (secondary)>
The initial copy from the logical volume (primary) 32 to the logical volume (secondary) 34 will be described in more detail with reference to FIG. FIG. 7A is a schematic diagram showing an overview of all differential copies from a logical volume (primary) to a logical volume (secondary).

  In the copy difference management table 25, information indicating whether or not there is a difference in data for each track corresponding to the logical volume (primary) 32 and the logical volume (secondary) 34 (hereinafter referred to as a difference track bit). Have. In the case of an initial copy from the logical volume (primary) 32 to the logical volume (secondary) 34, all the differential bits are turned on.

  The control unit 23A sequentially writes from the first track address to the last track address of the difference track bit in the copy pointer 24, refers to the difference track bit of the written address, and sets the track corresponding to the difference bit to the logical volume. Copy from the (primary) 32 to the logical volume (secondary) 34 and turn off this differential bit.

As a result, copying from the logical volume (primary) 32 to the logical volume (secondary) 34 can be performed.
If there is an access from the host 11 during copying, the following processing is performed.
When there is an access to the track after passing the copy pointer 24, the control unit 23A accesses the logical volume (primary) 32, and when the access is a write, turns on the corresponding differential bit. .

When there is an access to the track before passing the copy pointer 24, the control unit 23A accesses the logical volume (primary) 32.
After all of the differential data between the logical volume (primary) 32 and the logical volume (secondary) 34 and all the differential copies that occurred when writing was performed on the track after passing the copy pointer 24, the logical volume (primary) The difference data between the logical volume (primary) 32 and the logical volume (secondary) 34 generated when the data is written to 32 is the difference copy target from the logical volume (primary) 32 to the logical volume (secondary) 34, which will be described later Become.

<Differential copy from logical volume (primary) to logical volume (secondary)>
The difference copy from the logical volume (primary) 32 to the logical volume (secondary) 34 will be described in more detail with reference to FIG. FIG. 8 is a schematic diagram showing an outline of differential copy from the logical volume (primary) to the logical volume (secondary).

  In the copy difference management table 25, information indicating whether or not there is a difference in data for each track corresponding to the logical volume (primary) 32 and the logical volume (secondary) 34 (hereinafter referred to as a difference track bit). If there is a difference in the data, the difference bit is on, and if there is no difference, the difference bit is off.

  The control unit 23A sequentially writes from the first track address to the last track address of the difference track bit in the copy pointer 24, refers to the difference track bit of the written address, and only when the difference bit is on The track corresponding to the bit is copied from the logical volume (primary) 32 to the logical volume (secondary) 34, and the differential bit is turned off.

As a result, the copy from the logical volume (primary) 32 to the logical volume (secondary) 34 is not lost, and only a track with a difference can be copied. Therefore, it can be performed more quickly than a full copy.
If there is an access from the host 11 during copying, the following processing is performed.
When there is an access to the track after passing through the copy pointer 24, the control unit 23A accesses the logical volume (secondary) 34, and when the access is a write, turns on the corresponding differential bit. .

  When there is an access to the track before passing through the copy pointer 24, the control unit 23A makes the logical volume (primary) 32 access, the access is write, and the corresponding difference bit is off. If so, turn it on.

  As a result, the control unit 23A automatically switches access to the logical volumes (primary) and (secondary), so that the host is not aware of the logical volumes (primary) and (secondary) even during copying. Access to a magnetic disk storage device.

<Differential copy from logical volume (secondary) to logical volume (primary)>
The difference copy from the logical volume (secondary) 34 to the logical volume (primary) 32 will be described in more detail with reference to FIG. FIG. 9 is a schematic diagram showing an outline of differential copy from the logical volume (secondary) 34 to the logical volume (primary) 32.

  Similar to the differential copy from the logical volume (primary) to the logical volume (secondary), the control unit 23A sequentially writes the write from the address of the first track to the address of the last track in the differential track bit. The track corresponding to the difference bit is copied from the logical volume (secondary) 34 to the logical volume (primary) 32 only when the difference bit is turned on with reference to the difference track bit of the read address, and the difference bit is turned off. To do.

As a result, the copy from the logical volume (secondary) 34 to the logical volume (primary) 32 is not lost, and only a track with a difference can be copied. Therefore, it can be performed more quickly than a full copy.
If there is an access from the host 11 during copying, the following processing is performed.
When there is an access to the track after passing the copy pointer 24, the control unit 23A accesses the logical volume (primary) 32, and when the access is a write, turns on the corresponding differential bit. .

  When there is an access to the track before passing the copy pointer 24, the control unit 23A accesses the logical volume (secondary) 34, the access is write, and the corresponding difference bit is off. If so, turn it on.

  As a result, the control unit 23A automatically switches access to the logical volumes (primary) and (secondary), so that the host is not aware of the logical volumes (primary) and (secondary) even during copying. Access to a magnetic disk storage device.

<Differential copy using two copy difference management tables>
The differential copy between the logical volume (primary) 32 and the logical volume (secondary) 34 can also be performed by using two copy difference management tables. This will be described in detail with reference to FIG. FIG. 10 is a schematic diagram showing an outline of differential copy using two copy difference management tables.

  In this case, the copy difference management table 25A used for differential copy from the logical volume (primary) to the logical volume (secondary) and the copy difference used for differential copy from the logical volume (secondary) to the logical volume (primary) The management table 25B is used.

  When performing differential copy from the logical volume (primary) to the logical volume (secondary), the control unit 23A sequentially writes from the address of the first track of the differential track bit to the address of the last track in the copy pointer 24, and the written address. The copy difference management table 25A is referred to, and only when the difference bit is on, the track corresponding to the difference bit is copied from the logical volume (primary) 32 to the logical volume (secondary) 34, and the corresponding copy difference management after copying is performed. The difference bits of both the table 25A and the copy difference management table 25B are turned off.

If there is an access from the host 11 during copying, the following processing is performed.
The control unit 23A refers to the copy difference management table 25A. When the difference bit of the accessed track is off, the control unit 23A accesses the logical volume (secondary) 34. When the access is write, the control unit 23A writes The corresponding difference bit in the post-copy difference management table 25B is turned on.

  If the difference bit of the track that has been accessed is ON, the corresponding track is copied from the logical volume (primary) 32 to the logical volume (secondary) 34, the difference bit of the copy difference management table 25A is turned OFF, The volume (secondary) 34 is accessed. If the access is writing, the corresponding difference bit in the copy difference management table 25B is turned off after writing.

  As described above, when two copy difference management tables are used, even if there is an access from the host 11 during the differential copy, the access can be performed without determining whether the copy pointer 24 has passed or not. It becomes possible.

  The magnetic disk system of the present invention has been described above with specific embodiments, but the present invention is not limited to these. Those skilled in the art can make various changes and improvements to the configuration and function of the magnetic disk system in each of the above embodiments without departing from the scope of the present invention.

  For example, the operation performed by the control unit of the present invention has exactly the same operational effects when the computer reads and executes the program on the storage device, and can solve the problem to be solved by the invention.

The magnetic disk system of the present invention is realized by an OS, an application, a database, a network system, etc. built on hardware resources including a computer CPU, memory, auxiliary storage device, display, input device, and the like. Since information processing such as differential copy, logical volume (primary), and (secondary) access control is specifically realized using the above hardware resources, it falls under the technical idea of using the laws of nature. It can be used in the information processing industry that uses a disk system configured with RAID.

1 is a schematic diagram showing an embodiment of a magnetic disk system of the present invention. It is a flowchart of logical volume pair creation. It is a switching flow diagram of logical volumes (primary) and (secondary). FIG. 10 is a flowchart of data update of a RAID group (secondary) 33 and spindown of a RAID group (primary) 31; It is a flowchart of the operation | movement in RAID group (primary) 31 spindown. FIG. 10 is a flow chart of spin-down of a non-copy target logical volume 35. FIG. 10 is a flowchart of data update of a RAID group (primary) 31 and spin down of a RAID group (secondary) 33; (A) It is a schematic diagram showing an overview of all differential copies from a logical volume (primary) to a logical volume (secondary). It is a schematic diagram showing an outline of differential copy from a logical volume (primary) to a logical volume (secondary). 3 is a schematic diagram showing an outline of differential copy from a logical volume (secondary) 34 to a logical volume (primary) 32. FIG. It is a schematic diagram which shows the outline | summary of the difference copy which uses two copy difference management tables.

Explanation of symbols

11 Host 12 PC
20 Control device 21 Cache memory 22 User data 23 Shared memory 23A Control unit 24 Copy pointer 25 Copy difference management table 25A Copy difference management table 25B Copy difference management table 26 On-off timer management table 26A Access monitor timer 27 after re-spin-up 27 Pair management table 28 RAID group status management table 30 Magnetic disk storage device 31 RAID group (primary)
32 Logical volume (primary)
33 RAID group (secondary)
34 Logical volume (secondary)
35 Non-copy target logical volume 37 Physical volume

Claims (4)

A magnetic disk system having a magnetic disk storage device including a plurality of magnetic disk groups constituting a RAID, and a control device for controlling the magnetic disk storage device,
The magnetic disk storage device is powered on in a first time zone that is set and powered off in other time zones, and is powered on in a second time zone that is set. And a second magnetic disk group that is turned off at other times,
The second magnetic disk group is composed of a smaller number of magnetic disks than the first magnetic disk group,
The controller is
An address indicating a position of a track which is a part of a storage area of the first magnetic disk group and the second magnetic disk group, the first magnetic disk group and the second magnetic disk group; A copy pointer portion in which the address of the track that is the target of data copy is written,
A copy difference management table section in which information indicating whether or not there is a difference in data between the first magnetic disk group and the second magnetic disk group is written for each track;
An on / off timer management table section in which the setting information of the first time zone and the setting information of the second time zone are written;
Information indicating that the magnetic disk constituting the magnetic disk storage device belongs to the first magnetic disk group and information indicating that the magnetic disk belongs to the second magnetic disk group are written. A pair management table section;
A RAID group state management table portion in which information relating to the operation of the first magnetic disk group and the second magnetic disk group is written;
A magnetic disk system comprising:   2. The magnetic disk system according to claim 1, wherein the number of magnetic disks constituting the second magnetic disk group is equal to or less than half of the number of magnetic disks constituting the first magnetic disk group. The addresses of all tracks to be copied from the first magnetic disk group to the second magnetic disk group are sequentially written in the copy pointer section,
When copying the track indicated by the address from the first magnetic disk group to the second magnetic disk group in order for each address written in the copy pointer section,
Information indicating that there is a difference in data between the first magnetic disk group and the second magnetic disk group in the track indicated by the address written in the copy pointer section is written in the copy difference management table section. 3. The magnetic disk system according to claim 1, wherein the track is copied from the first magnetic disk group to the second magnetic disk group only when the first magnetic disk group is recorded. When there is a data read or data write access from the host while data is being copied from the first magnetic disk group to the second magnetic disk group,
If the address of the accessed track has already been written in the copy pointer section, the host is made to access the second magnetic disk group, and if the access is a write, the copy difference management table Information indicating that there is a data difference in the track that was accessed
If the address of the accessed track has not yet been written to the copy pointer section, the host is made to access the first magnetic disk group, and if the access is a write, the copy difference management table Information indicating that there is a data difference in the track that was accessed in the part,
The magnetic disk system according to claim 1, wherein the magnetic disk system is a magnetic disk system.