JPH07325674A - Semiconductor memory replacement method and semiconductor disk subsystem control method - Google Patents

Abstract

(57) [Abstract] [Purpose] Performs live line maintenance of any semiconductor memory unit that constitutes the semiconductor memory without duplicating the semiconductor memory. [Structure] A semiconductor disk device 101 connected to a channel 102 includes a semiconductor memory unit 104 including a plurality of memory packages 105 and a spare area 106 having a capacity of one of the memory packages 105. Package information for identifying the attribution of each of the plurality of logical tracks built on the plurality of memory packages 105 to each memory package 105 is set, and by referring to the package information when exchanging any memory package 105, The logical package group on the memory package 105 is selectively saved in the spare area 106, and the memory package 105 is replaced while responding to the access request to the channel 102 by the logical track on the spare area 106.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor memory replacement technology and a semiconductor disk subsystem control technology, and more particularly to a technology effectively applied to replacement of a semiconductor memory package.

[0002]

2. Description of the Related Art In recent years, it has become possible to duplicate a semiconductor memory even in a semiconductor disk due to cost reduction and large scale of the semiconductor memory device. Although this improves the reliability of the device, the bit unit price of semiconductor memory is still considerably higher than that of magnetic disks and magnetic tapes, and simple duplication from the viewpoint of price competitiveness and increased installation space. It is at a disadvantage.

When the semiconductor memory is not duplicated, it is necessary to separate and maintain not only the memory unit but also the entire subsystem when exchanging the package when the semiconductor memory unit fails or performs preventive maintenance, and the system is operated for 24 hours. It is a problem when making it.

As one means for solving this, Japanese Patent Laid-Open No.
As described in Japanese Patent No. 268020, a technique that allows a package to be replaced online by using a maintenance device independent of the semiconductor disk device and substituting a maintenance IC memory provided in the maintenance device. and so on.

Since a semiconductor disk device generally uses a plurality of semiconductor memories by dividing it into a plurality of logical volumes, one semiconductor memory stores data divided into a plurality of logical volumes. That is, when exchanging a semiconductor memory, the conventional technique does not consider a plurality of logical volumes and logical tracks existing in one semiconductor memory. The technical problems related to exclusive control have not been resolved.

Further, due to recent hardware technology,
Even if the device is plugged in and unplugged while the device power is on, live line maintenance is performed that does not physically affect other parts, so an independent power supply should be used when replacing the memory package. There is no need to install it.

[0007]

The above-mentioned conventional technique has a problem that a special maintenance device having an alternative IC memory is required in addition to the semiconductor disk device, and the equipment cost increases. In addition, if the higher-level device accesses during the period when the data of the semiconductor memory to be replaced is being copied to the buffer or the spare memory, which of the copy-destination data is used to process the data It has not been clarified and no consideration was given to copy technology such as management of logical tracks when data is updated.

It is an object of the present invention to provide a semiconductor memory replacement technique capable of hot-line maintenance of any semiconductor memory unit constituting the semiconductor memory without duplicating the semiconductor memory.

Another object of the present invention is a semiconductor disk subsystem capable of performing hot line maintenance of an arbitrary semiconductor memory unit, which is a constituent element of the semiconductor memory, without duplicating the semiconductor memory constituting the semiconductor disk. To provide the control technology of.

Still another object of the present invention is to replace a semiconductor memory unit which is a constituent element of the semiconductor memory without hot duplication of the semiconductor memory constituting the semiconductor disk, and which is a semiconductor memory unit to be replaced. Another object of the present invention is to provide a control technology of a semiconductor disk subsystem capable of accurately performing exclusive control of access during data copy between a spare semiconductor memory unit or a buffer.

[0011]

A method of exchanging a semiconductor memory according to the present invention is a semiconductor memory in which a plurality of semiconductor memory units which can be physically exchanged independently from each other are used and data is stored in a predetermined logical data unit. , Which has attribution identification information for identifying attribution of the logical data unit to each of the individual semiconductor memory units, and selectively separates the logical data unit stored in the semiconductor memory unit prior to replacement of any semiconductor memory unit. By evacuating to the memory location and accessing the logical data unit in the memory location, an arbitrary semiconductor memory unit is replaced during the operation of the semiconductor memory.

In the semiconductor disk subsystem control method of the present invention, the semiconductor memory of the semiconductor disk device is initialized (memory initialization) or the service processor issues an instruction (copy) to replace the memory package (semiconductor memory unit). When the memory package replacement is instructed, the data related to the memory package is created in the shared memory in the track bit map format. If the semiconductor file is expanded and an I / O request is issued from a higher-level device during copying of the semiconductor file, it is determined whether the logical volume for which the I / O request is being made is being copied, and the bit of the shared memory The memory package is accessed based on the information in the map. By a spare of the control to access a memory package or buffer or, in which so as to allow the exclusive control of the copying process.

[0013]

According to the above-described semiconductor memory replacement method of the present invention, the data stored in any semiconductor memory unit to be replaced can be selectively saved without duplicating the capacity of the semiconductor memory. Line maintenance can be performed accurately.

Further, according to the semiconductor disk subsystem control method of the present invention, it is clarified which of the memory package to be replaced, the spare memory package and the buffer to be accessed during the copy process. Since the judgment can be made, the copy status can be managed even when there is an input / output request from the host device, the memory package that constitutes the semiconductor memory can be replaced accurately online, and there is no need to duplicate the semiconductor memory. .

[0015]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an example of an electronic computer system constructed using a semiconductor disk subsystem to which a semiconductor memory replacement method and a semiconductor disk subsystem control method according to an embodiment of the present invention are applied. Is. 2, 3 and 4 are flow charts showing an example of the operation of the control method of the semiconductor disk subsystem of this embodiment. FIG. 5 is a conceptual diagram showing an example of the relationship between the hardware (memory package) and the logical volume in the semiconductor disk subsystem of this embodiment, and FIG. 6 is a conceptual diagram showing an example of the details of the logical volume. 7, 8 and 9 are conceptual diagrams showing an example of package information, volume information and memory start address information of a spare area used in the method of controlling the semiconductor disk subsystem of this embodiment.

As shown in FIG. 1, the electronic computer system according to the present embodiment has a channel 102 for controlling input / output with the outside instead of the central processing unit based on a command from a central processing unit (not shown), and a semiconductor. A semiconductor memory unit 104 including a memory package 105 including a memory and a spare area 106, a microprocessor 107 that operates based on an instruction stored in a CS memory 108,
Semiconductor circuit device 101 including switching circuit 109, channel 102, data transfer circuit 111 for performing data transfer, and shared memory 110, service processor 1
It consists of 03. The microprocessor 107 can access the semiconductor memory unit 104 via the switching circuit 109. In addition, the channel 102 and the semiconductor memory unit 10
4 and a data transfer with the microprocessor 107
Switching circuit 109 and data transfer circuit 11 according to a command from
Data is transferred via 1. Service processor 10
Reference numeral 3 is connected to the semiconductor disk device 101, and various kinds of maintenance can be performed according to instructions from maintenance personnel.

Next, the configuration of the semiconductor memory section 104 and the details of an example of the data storage format will be described with reference to FIGS. The memory package 105 includes a package a (501), a package b (502), and a package c.
It is physically divided in hardware such as (503), and individual packages can be exchanged independently of each other. On the other hand, the memory package 105 is
Logical volume 1 (511) to logical volume 5 (51
It is logically divided into 5). Logical volume 1 (51
1) to the logical volume 5 (515), as shown in FIG. 6, cylinder 0 to cylinder X (603) and E
OF track (604), CE track 0 (604),
CE track 1 (605) and memory surplus (60
7) and. Also, cylinder 0 to cylinder X
Each cylinder of (603) is track 0 to track 14
(608). In addition, the control information area 601
The inline area 602 and the logical volume 0 (5
It is placed in front of 11). Further, in FIG. 5, the package d (504) corresponds to the spare area 106, which serves as a temporary save area for data when the package is replaced, and the memory capacity is one of the packages a (501) to c (503). It is the same as the memory capacity of minutes.

Next, referring to FIG. 7, package information 700
Will be described in detail. The package information 700 exists as many as the number of memory packages 105 installed in the semiconductor memory unit 104 of the semiconductor disk device 101. The package information 700 exists in the shared memory 110. In the case of the present embodiment, there are packages a (501) to package c (503), and the package information of each package is C of the final volume in each package.
It consists of CHH (CC: cylinder number, HH: head (track) number) and CCHH of the final volume in the package. In this embodiment, the package a
As for (501), the contents of the package information 700 are such that the first CCHH (701) of the logical volume 1 in the package a (501) and the end of the logical volume 2 in the package a (501) are as shown in FIG. CC
Holds HH (702). Regarding the package b (502), the contents of the package information 700 are
First CCHH (7 of logical volume 2 in (502)
03) and the logical volume 4 in the package b (502)
Of the final CCHH (704). As for the package c (503), the contents of the package information 700 are as follows: the first CCHH (705) of the logical volume 4 in the package c (503) and the final CCHH (of the logical volume 5 in the package c (503). 706) and. The package information 700 is the semiconductor disk device 10
It is set at IMPL (Initial Micro Program Load) after power-on of 1.

Next, referring to FIG. 8, volume information 800
Will be described in detail. Volume information 800 exists for each logical volume. Also, the volume information 800 exists in the shared memory 110. The volume information 800 includes a busy flag 801, a copy flag 80.
2, all VOL copy flag 803, copy pointer 80
4, a reserve area flag 805, a route 806 responding to a command retry, and a route 807 responding to a device busy. The busy flag 801 indicates that the volume is being used by the channel 102.
The copy flag 802 is a flag indicating that the volume is being copied from the regular package to the spare area 106 or from the spare area 106 to the regular package for package replacement. The all-VOL copy flag 803 indicates that the entire volume is being copied when the package is replaced. Copy pointer 80
When there is a logical volume that extends between packages, 4 stores the CCHH at the beginning of copying and the CCHH at the end of copying. The spare area flag 805 indicates that the spare area 106 already has volume data. A route 806 that responds to the command retry indicates a channel route that responds to the command retry while copying the spare area 106 from the regular package. The route that responded with device busy indicates the route that responded with device busy because the volume is in use.

Next, FIG. 9 will be described. FIG. 9 shows the memory start address of the spare area 106. This information is necessary for copying from the regular package to the spare area 106 and exists in the shared memory 110.

An example of the operation of this embodiment will be described below.

First, the operation illustrated in the flowchart of FIG. 2 will be described with reference to FIGS. In step 201,
It is determined whether or not there is activation from the upper level (channel 102). If there is activation from the upper layer, the process proceeds to step 212 and the processing shown in FIG. 3 is performed. If there is no activation from the host, it is checked in step 202 whether there is an interrupt from the service processor 103. If there is no interrupt from the service processor 103, step 213
If there is an interrupt from the service processor 103,
The processing of step 203 is performed. In step 203, it is checked whether the request is a copy request as a result of communication with the service processor 103 according to a predetermined interface.
If it is not a copy request, the service processor 103 is processed by the determined interface at step 219. If it is a copy request in step 204,
The copy information (package number, package a (50
1) to package c (503)) are fetched. The package information 700 is searched from the copy information (package information) fetched in step 206, and the designated package information (package b (502) in this embodiment) is searched.

The volume information 800 is updated from the package information acquired in step 207. In this embodiment,
Since it is package b (502), logical volume 2
Copy flag 8 in the volume information 800 of (512)
02 is set to “1”, the first CCHH of the logical volume 2 (512) in the package b (502) is stored in the copy pointer 804, and the spare area flag 805 is set to “1”. Further, since the logical volume 3 (503) is all included in the package b (502), the logical volume 3
Copy flag 8 in the volume information 800 of (513)
02 is "1", all VOL copy flag 803 is "1",
And the spare area flag 805 is set to "1". Since the package b (502) is in the middle of the logical volume 4 (514), the copy flag 802 in the volume information 800 of the logical volume 4 (514) is set to "1" and the copy pointer 804 is set in the package b (502). The final CCHH of the logical volume 4 (514) is stored, and the spare area flag 805 is set to '1'. In step 208, the control unit busy is set to perform the copy operation. In step 209, data is copied from the package b (502) to the package d (504) in the spare area 106 based on the package information 700 and the memory start address 900 of the spare area 106. Step two
At 10, it is checked whether the copy process is completed. If the copy process is not completed, the process returns to step 209 to continue the copy process. If the copying is completed in step 210, the processing shown in FIG. 4 is executed in step 211.

In step 213, the copy flag 802 = 0
Check. If the copy flag is “1”, the process jumps to step 216. If step 2
If the copy flag is "1" in step 13, step 214
If there is no route 806 that responded to the command retry in step 216, the process jumps to step 216. Step 21
If the route 806 that responded the command retry in 4 exists, the device end is responded based on the information of the route 806 that responded the command retry in step 215.

In step 216, it is checked whether the busy flag 801 in the volume information 800 of each logical volume is "1", and if the busy flag in the volume information 800 of each logical volume is "1" in step 216.
If so, the process ends. If the in-use flag 801 in the volume information 800 of each logical volume is not "1" in step 216, is there a route 807 that responded to the device busy in the volume information 800 of each logical volume in step 217? Check.
If the device busy in the volume information 800 of each logical volume is returned in step 217, the route 807
If is not present, the process ends. If step 21
When there is a route 807 that responded device busy in the volume information 800 of each logical volume in step 7, route 807 that responded device busy in step 218
The device end is responded based on.

Next, the operation illustrated in the flow chart of FIG. 3 will be described with reference to FIGS. First, step 30
1, the in-use flag 801 in the volume information 800 of the logical volume activated from the upper level (channel 102)
Check if is '1'. If the volume information 80 of the logical volume that was activated from the upper level in step 301
If the in-use flag in 0 is '1', step 3
Jump to 12. If the in-use flag in the volume information 800 of the logical volume activated from the upper level in step 301 is not "1", the copy flag 802 of the volume information 800 of the logical volume activated from the higher level in step 302 is "1". Check if it is 1 '.
If the copy flag 802 in the volume information 800 of the logical volume that was activated from the upper level in step 302
Is not “1”, the process jumps to step 308. If the copy flag 8 in the volume information 800 of the logical volume that was activated from the upper level in step 302
If 02 is "1", it is checked in step 303 whether the all VOL copy flag 803 in the volume information 800 of the logical volume activated from the upper level is "1". If the all-VOL copy flag 803 in the volume information 800 of the logical volume activated from the upper level in step 303 is “1”, step 30
Jump to 6. If the all-VOL copy flag 803 in the volume information 800 of the logical volume that was activated from the upper level in step 303 is not "1",
In step 304, the copy first CCHH of the copy pointer 804 in the volume information 800 of the logical volume that was activated from the upper layer and the CCHH accessed from the upper layer.
Compare with. If the copy first CCHH of the copy pointer 804 in the volume information 800 of the logical volume activated from the upper level in step 304 is larger than the CCHH accessed from the higher level, the CCHH is not the copy target CCHH and the process proceeds to step 308. To jump. If the copy first CCHH of the copy pointer 804 in the volume information 800 of the logical volume activated from the upper level in step 304 is smaller than or equal to the CCHH accessed from the upper level, it is the CCHH to be copied. Step 3 because there is a possibility
Go to 05. In step 305, the CCHH of the copy end of the copy pointer 804 in the volume information 800 of the logical volume activated from the upper level is compared with the CCHH accessed from the upper level. If, in step 305, the CCH of the copy end of the copy pointer 804 in the volume information 800 of the logical volume that was activated from the upper level
If H is smaller than the CCHH accessed from the upper layer, it is not the CCHH to be copied, and thus step 30
Jump to 8. If the final copy CCHH of the copy pointer 804 in the volume information 800 of the logical volume activated from the upper level in step 305 is greater than or equal to the CCHH accessed from the higher level, it is the CCHH to be copied. , Step 3
Proceed to 06. In step 306, the route accessed by the route 806 that responds to the command retry in the volume information 800 of the logical volume accessed from the upper level is stored, and the process proceeds to step 307. Step 307
Then, a command retry is responded to the access from the channel 102.

At step 308, the in-use flag 801 in the volume information 800 of the logical volume that has been activated from the upper level is set to "1". In step 309, the spare area flag 805 in the volume information 800 of the logical volume activated from the upper level is checked. This spare area flag 805 is a flag for discriminating whether to access a regular package or a package d (504) in the spare area 106. If the spare area flag 805 in the volume information 800 of the logical volume activated from the upper level in step 309 is “1”, step 31
Jump to 1. This is because the regular package is being replaced. If the volume information 800 of the logical volume that was activated from the upper level in step 309
When the reserved area flag 805 is not "1",
In step 310, the regular memory package 105 is accessed by the command from the channel 102 and command processing is performed. In step 311, the spare area 106 is accessed by a command from the channel 102 and command processing is performed.

In step 312, the route accessed by the route 807 that responds to the device busy of the volume information 800 of the logical volume that was activated from the upper level is stored, and in step 313, the device busy responds to the access from the channel 102. To do.

Next, the operation illustrated in the flowchart of FIG. 4 will be described with reference to FIGS. In step 401, the copy direction is from the memory package 105 to the spare area 10.
Check if the direction is 6. The information on the copy direction is given by the information from the service processor 103.
If the copy direction is not the direction from the memory package 105 to the spare area 106 in step 401, the process proceeds to step 404. If the copy direction is the direction from the memory package 105 to the spare area 106 in step 401, the process proceeds to step 402, and the spare area flag 805 in the volume information 800 of the logical volume to be copied is set to '1'. In step 403, the service processor 103 is notified of the end of copying.

At step 404, the copy flag 802, the all VOL copy flag 803, the copy pointer 804 and the spare area flag 805 in the volume information 800 of the logical volume to be copied are cleared. This is from the spare area 106 to the memory package 1
Indicates that copying to 05 is complete.

As described above, according to the semiconductor memory replacement method and the semiconductor disk subsystem control method of this embodiment, the attribution of logical data units such as volumes and tracks in each memory package 105 is managed by the package information 700. Therefore, at most, the spare area 10 having the capacity of one memory package 105 to be replaced is required.
By only providing 6, the optional memory package 105 can be replaced in the online state without duplication.

Further, the memory package 105 to be replaced
By managing the copy status of the data from the data area to the spare area 106 by the volume information 800, it is possible to accurately perform the exclusive control for the access request from the channel 102 during the data copy.

In the above embodiment, the in-use flag 801 and the copy flag 8 in the volume information 800 are used.
Needless to say, 02, all VOL copy flag 803, and spare area flag 805 may be either bit unit or byte unit.

In the above embodiment, it is assumed that the package is put in the spare area 106, but the spare area 1
A package originally fixed in 06 may be included. Further, the spare area 106 may be an electronic buffer.

In the above description of the embodiment, the semiconductor disk device and the control microprocessor are integrated, but the semiconductor disk device is divided into the semiconductor disk controller and the semiconductor disk device. It goes without saying that the present invention can be applied even if the above is done.
The present invention can also be applied to the maintenance of electronic buffers such as a cache memory attached to a magnetic disk subsystem and a semiconductor memory in a cartridge magnetic tape device.

[0037]

According to the method of exchanging a semiconductor memory of the present invention, it is possible to carry out hot line maintenance of an arbitrary semiconductor memory unit constituting the semiconductor memory without duplicating the semiconductor memory. .

Further, according to the control method of the semiconductor disk subsystem of the present invention, hot line maintenance of any semiconductor memory unit which is a constituent element of the semiconductor memory is carried out without duplicating the semiconductor memory constituting the semiconductor disk. The effect that can be obtained is obtained. Also, without duplicating the semiconductor memory that constitutes the semiconductor disk,
It is possible to accurately perform exclusive control of access during data copy between the semiconductor memory unit to be replaced and the spare semiconductor memory unit or the buffer during hot line maintenance of any semiconductor memory unit that is a component of the semiconductor memory. The effect of, is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a computer system constructed by using a semiconductor disk subsystem to which a semiconductor memory replacement method and a semiconductor disk subsystem control method according to an embodiment of the present invention are applied. .

FIG. 2 is a flowchart showing an example of the operation of the control method of the semiconductor disk subsystem that is an embodiment of the present invention.

FIG. 3 is a flowchart showing an example of the operation of the control method of the semiconductor disk subsystem that is an embodiment of the present invention.

FIG. 4 is a flowchart showing an example of the operation of the control method of the semiconductor disk subsystem that is an embodiment of the present invention.

FIG. 5 is a conceptual diagram showing an example of a relationship between a memory package and a logical volume in the control method of the semiconductor disk subsystem which is an embodiment of the present invention.

FIG. 6 is a conceptual diagram showing an example of details of a logical volume in a method of controlling a semiconductor disk subsystem according to an embodiment of the present invention.

FIG. 7 is a conceptual diagram showing an example of details of package information in a control method of a semiconductor disk subsystem which is an embodiment of the present invention.

FIG. 8 is a conceptual diagram showing an example of details of volume information in the control method of the semiconductor disk subsystem which is an embodiment of the present invention.

FIG. 9 is a conceptual diagram showing a memory start address of a spare area in the semiconductor disk subsystem control method according to an embodiment of the present invention.

[Explanation of symbols]

101 ... Semiconductor disk device, 102 ... Channel (upper device), 103 ... Service processor, 104 ... Semiconductor memory unit (semiconductor memory), 105 ... Memory package (semiconductor memory unit), 106 ... Spare area, 10
7 ... Microprocessor, 108 ... CS memory, 109
... switching circuit, 110 ... shared memory, 111 ... data transfer circuit, 511-515 ... logical volume, 601 ... control information area, 602 ... inline area, 603 ... cylinder, 608 ... track (logical data unit: logical track), 700 ... Package information (first management information), 8
00 ... Volume information (second management information), 801 ... In-use flag, 802 ... Copy flag, 803 ... All VOL
Copy flag, 804 ... Copy pointer, 805 ... Reserved area flag, 806 ... Route that responded to command retry, 807 ... Route that responded to device busy, 90
0 ... Memory start address.

Claims (3)

[Claims] 1. A semiconductor memory comprising a plurality of semiconductor memory units that can be physically exchanged independently of each other, wherein data is stored in a predetermined logical data unit, and the logical data unit for each of the individual semiconductor memory units. Of the logical data unit stored in the semiconductor memory unit prior to the replacement of any of the semiconductor memory units, and selectively saves the logical data unit to another storage location. In the method of exchanging a semiconductor memory, an arbitrary semiconductor memory unit is exchanged during operation of the semiconductor memory by making access to the logical data unit. 2. A logical track, which corresponds to each track in the disk device and stores predetermined information in a format equivalent to that of the disk device, is constructed in a semiconductor memory and responds to an input / output request from a host device. In the semiconductor disk subsystem for controlling access to the logical track by means of first management information for identifying the attribution of the logical track to each of a plurality of semiconductor memory units that constitute the semiconductor memory and can be exchanged independently of each other. , A control logic for recognizing the logical track stored in each of the semiconductor memory units from the first management information is provided, and when the semiconductor memory unit is replaced, access to the semiconductor memory unit is performed. Avoid and allow access to a spare semiconductor memory unit or buffer According to the method, the input / output operation is continued during the replacement of any of the semiconductor memory units. 3. When the replacement of the semiconductor memory unit is designated by a service processor, control logic for gradually copying data of a logical track belonging to the semiconductor memory unit to a spare semiconductor memory unit or a buffer, and this copy. And a second management information for managing the status, the copied data is made to access a spare semiconductor memory unit or a buffer, and the non-copy data is made to access the semiconductor memory unit, so that the host device A first operation of transferring data access from the semiconductor memory unit to a spare semiconductor memory unit or a buffer during an input / output operation with the semiconductor memory unit, and via a service processor when the copying operation of the semiconductor memory unit is completed. Set the control logic to notify the end of copy processing. , When the maintenance engineer replaces the semiconductor memory unit and receives a designation from the service processor, the backup semiconductor memory unit or the buffer data is copied step by step to the exchanged semiconductor memory unit, and after the copying is completed, the original position is restored. The method of controlling a semiconductor disk subsystem according to claim 2, wherein at least one of a second operation for returning access to the semiconductor memory unit is performed.