JPH0667817A - Method for maintaining semiconductor disk device - Google Patents

Abstract

PURPOSE:To improve the operation efficiency of a device by executing processing without turning off the device at the time of changing memory capac ity or repairing a fault memory. CONSTITUTION:In the maintenance method of a semiconductor disk device 2 to be used by dividing a semiconductor memory part (TM) 8 into plural logical drives DRV0 to DRV1, hierarchical control tables 9 each of which consists of a basic control table and an extended control table are set up in the device 2 at every logical drive, and at the time of changing a memory, the control table 9 is updated at every logical drive, so that memory changing processing can be executed so as to make access logical drives other than the one concerned from a master device 1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of maintaining a semiconductor disk device, which uses a semiconductor memory as a storage medium and virtually executes input / output processing of the magnetic disk device.

[0002]

12 to 14 are explanatory views of a conventional example, FIG. 12A is a block diagram of a semiconductor disk device, and FIG.
2B is an example of a control table. 13A and 13B are explanatory diagrams when the number of cylinders is increased. FIG. 13A shows before formatting and FIG. 13B shows after formatting. Furthermore, FIG.
[FIG. 3] is an explanatory diagram when repairing a memory failure.

12 to 14, 1 is a central processing unit (C
internal Processing Unit: hereinafter referred to as “CPU”, 2 is a semiconductor disk device (Sol)
idState Disk: hereinafter referred to as "SSD"),
3 is a channel adapter (Channel Adapte)
r: hereinafter referred to as “CA”) 4, 5 are control storage (hereinafter “C”)
S), 6 is a service processor (hereinafter referred to as "SV").
7 is a maintenance panel (hereinafter referred to as “PN”).
L ”), 8-track memory (Track Mem)
ory: hereinafter referred to as "TM"), 9 indicates a control table.

Conventionally, a semiconductor memory is used as a storage medium,
An SSD has been known as a device that virtually executes input / output processing of a disk device. FIG. 12 shows an example thereof (see, for example, Japanese Patent Laid-Open No. 61-165153).

Description of SSD configuration: see FIG. 12. As shown in the figure, SSD2 is CA3, SVP6, CS4,
It is composed of CS5, TM8, PNL7, etc., and is used by connecting to the CPU1 which is a host system.

The SSD 2 has a TM having a predetermined capacity.
Data input / output processing by dividing 8 (constructing a semiconductor memory unit) into a plurality of capacities (for example, N1 to N5) and considering them as independent logical drives (DRV0 to DRV4) (emulation function) It is carried out.

The CA3 is a processor that supports the exchange of commands or data (input / output processing) with the CPU1 and the emulation function of the SSD2.
The VP6 is a processor that controls the entire SSD2.

The PNL 7 is a panel operated by an operator or the like, and externally issues an instruction to the SVP 6 to perform maintenance. A control table 9 as shown in FIG. 12B is provided in each of CS4 and CS5.
Is set, and the control table 9 in CS4 is set to CA3.
The SVP 6 uses the control table 9 in the CS 5 to perform various controls (emulation).

TM8 is a memory which constitutes a semiconductor memory unit, and is divided into a plurality of logical drives DRV0 to DRV4 having different capacities to perform control as shown in the figure.
In this case, as described above, the control table 9 is set in CS4 and CS5, and each logical drive DRV0-DRV is set.
The capacity (number of cylinders: CLYNUM) for each 4 is registered.

In this example, the logical drives DRV0-DR
The capacity of V4 is DRV0 = N1, DRV1 =
N2, DRV2 = N3, DRV3 = N4, DRV4 = N
Set like 5.

Description of processing when changing capacity of logical drive (addition in this example) ... See FIGS. 12 and 13. As described above, logical drives DRV0 to DRV in TM8 are used.
The capacity (number of cylinders) of RV4 is N1 to N5, respectively.
Suppose that it was used in the state set to. In this state, the capacity (number of cylinders) is increased as follows.

-1: Mount an additional memory. -2: The CPU 1 issues a specific instruction to the SSD 2. This particular instruction specifies the number of cylinders for all logical tribes.

-3: According to the above instruction, CA3 causes SV
Instruct P6 to change the setting of the control table 9. -4: SVP6 is a subordinate storage area TM8
Each of the logical drives DRV0 to DRV4 in C5 is taken offline, and the control table 9 in CS5 is redefined by the above command.

-5: When the above processing is completed, SV
P6 notifies CA3 of the end of the redefinition process by an interrupt. -6: CA3 accepts the above interrupt and then executes CS
Redefine the control table 9 in 4 according to the above instruction,
The TM8 is reformatted based on the control table 9.

As described above, instead of issuing a specific instruction from the CPU 1, there is also a method of issuing a change instruction from the PNL 7 and performing the processing of -4 to -6.
Through the above processing, the memory capacity is changed. In this case, as shown in FIG. 13, each capacity (number of cylinders) of each logical drive DRV0 to DRV4 before formatting
DRV0 = N1, DRV1 = N1, DRV2 = N
2, DRV3 = N4, DRV4 = N5, and
DRV0 = N9, DRV1 = N10, DRV2 = N11
Add each capacity of.

After the above process is completed, the capacity (number of cylinders) of each logical drive DRV0-DRV4 is DRV0 = N1 + N9, DRV1 = N1 + N1.
0, DRV2 = N2 + N11, DRV3 = N4, DRV
4 = N5.

[0017]: Description of processing at the time of repair in case of memory failure
-Refer to FIG. 12 and FIG. 14. When a failure occurs in TM8, processing is performed as follows. -1: For example, in accordance with an instruction from PNL7, SVP6
Makes the subordinate logical drives DRV0 to DRV4 offline.

-2: Replace the defective memory unit (for example, DRV2) with a good product. -3: The SVP 6 sets the control table 9 in the CS 5 and, at the time when this process ends, the SVP 6
The interruption notifies the end of the redefinition processing to CA3.

-4: After accepting the interrupt, the CA 3 redefines the control table 9 in the CS 4 and reformats the TM 8 based on the control table 9.

[0020]

SUMMARY OF THE INVENTION The above-mentioned conventional devices have the following problems. (1) When changing the capacity of the logical drive (addition of the number of cylinders, etc.), the setting information from the CPU 1 or PNL 7 is used to set / update the number of cylinders of all logical drives.

Therefore, all the drives (entire apparatus) are once brought into an off-line state, and the memory to be added and the memory that has been used until then are all reformatted by CA3.

That is, 1 per logical drive
In the method of controlling with one control table, when changing the capacity, it is necessary to re-format all the memories after disconnecting the SSD 2 from the CPU 1 as described above.

Therefore, since the SSD 2 cannot be continuously used, the operating efficiency (utilization efficiency) of the device (SSD) is lowered. (2) Even when a memory failure occurs, all the logical drives in the SSD 2 must be brought into the offline state once and the repair processing must be performed, as in the case of the above capacity change.

Therefore, also in this case, the operating efficiency (utilization efficiency) of the device (SSD) is lowered. The present invention solves such a conventional problem, and when the memory capacity is changed or the faulty memory is repaired, processing can be performed without taking the device offline, and the operating efficiency (utilization efficiency) of the device is improved. The purpose is to improve.

[0025]

FIG. 1 is a diagram for explaining the principle of the present invention, FIG. 1A is a block diagram of an SSD, and FIG. 1B is a configuration diagram of a control table. In FIG. 1, the same parts as those in FIGS. 12 to 14 are designated by the same reference numerals.

In order to solve the above problems, the present invention has the following configuration. (1) A disk is obtained by using a semiconductor memory (TM8) as a storage medium and dividing the semiconductor memory (TM8) into a plurality of logical drives DRV0, DRV1 ,. In the maintenance method of the semiconductor disk device 2 which virtually executes the input / output processing of the device, in the semiconductor disk device 2, a control table 9 for storing control information of the logical drive is provided in each of the logical drives DRV0, DRV1,・ ・ ・ Set each time,
This control table 9 is set as a hierarchical table, and when the memory is changed, each logical drive DRV0, DRV
By updating the control table 9 for each V1, ..., The memory changing process is performed while the logical drives other than the corresponding logical drive are accessible from the higher-level device 1.

(2) In the above configuration (1), the control table 9 comprises a basic control table for storing initial setting values of control information, and an expansion control table having the same configuration as the basic control table. By setting the addresses of the following control tables in the basic control table and the extension control table, the layers are hierarchized.

(3) In the configuration (1), when a memory is added to the semiconductor disk device that is operating online, the processor (CA3, SVP6) of the semiconductor disk device 2 is instructed from the outside by By referring to the control table 9, the control information of the logical drive to be added is acquired, only the logical drive to be added is taken offline, the format processing of the add-on unit is performed, and the control table 9 of the added logical drive is The value of the control information was redefined, and then the logical drive in the expansion section was brought back online.

(4) In the configuration (1), when the semiconductor disk device operating online is repaired (replaced) with the failed memory, the processor (CA3, SVP6) of the semiconductor disk device 2 is used. ), In accordance with an instruction from the outside, the control table 9 is referred to, the control information of the logical drive to be repaired is acquired, only the logical drive to be repaired is taken offline, and the notification is given to the outside. In the processor (CA3, SVP6) that received the end instruction, the control table 9
With reference to the above, the offline logical drive is formatted, and after the formatting is completed, the corresponding logical drive is brought back online.

[0030]

The operation of the present invention based on the above configuration will be described with reference to FIG. A basic control table is set for each logical drive as shown in FIG. 1B, and a table having the same configuration as this basic control table is set as an expansion control table. The expansion control table is used when adding a drive or cylinder.

Then, the control table is provided with a term of "address of next control table", and the head address of the next table can be set in this term so that the control table is hierarchized.

Further, in the basic control table, "total cylinder number" (for example, N1) can be set, and the presence or absence of the drive can be determined by this value. The hierarchical control table 9 makes it possible to dynamically represent the range of the memory (TM8) belonging to the logical drive.

The hierarchical control table 9 as described above
When the control table 9 is updated by the SVP 6, for example, when the memory is expanded in the semiconductor disk device 2 in which the setting is set, the SVP 6 interrupts the CA 3.

In this interrupt processing, the CA3 recognizes that the control table has been updated, and uses the updated control information of the control table 9 to update the memory to be updated (TM).
8) Format the part.

In this case, only the logical drive to be updated is made offline, and the logical drive not to be updated is in the online state, and the above processing is performed. Also, during the repair process at the time of a memory failure, only the logical drive to be repaired is offline and the other logical drives are
Repair processing can be performed online.

As described above, when the memory is changed (expansion, etc.), it is possible to carry out the processing while keeping the device online while taking only the relevant logical drive offline. Therefore, since the device can be continuously used, the operating efficiency (utilization efficiency) of the device is improved and the processing time can be shortened.

[0037]

Embodiments of the present invention will be described below with reference to the drawings. 2 to 11 are diagrams showing an embodiment of the present invention, FIG. 2 is a block diagram of a semiconductor disk device, FIG. 3 is a diagram showing an example of the relationship between a basic control table and TM, and FIG. Of the control table of FIG. 5 and an example of the relationship between TMs, FIG. 5 shows an example of the control table of Example 1 (when increasing the number of drives), and FIG. 6 shows an example of TM configuration of Example 1. FIG. 7, FIG. 7 is a process flow chart in the case of issuing an extension instruction from the SVP to perform the extension, FIG. 8 is a flow chart in the case of issuing an instruction from the SVP to perform the extension, and FIG. FIG. 10 is a diagram showing an example of the table, FIG. 10 is an explanatory diagram of the second embodiment (an example of the TM configuration), and
Reference numeral 1 is a processing flow chart in the case of performing repair (replacement) when a memory failure occurs.

2 to 11, the same components as those in FIG. 1 and FIGS. 12 to 14 are designated by the same reference numerals. Also, 10
Indicates a control table for comparison (having the same configuration as the control table 9).

Description of SSD (semiconductor disk device) configuration--see FIG. 2 As shown in the figure, SSD2 is CA3, SVP6, CS
4, CS5, TM8, PNL7, etc., and is used by connecting to the CPU1 which is the host system.

Then, the control table 9 is set in the CS4, and the control table 9 and the comparison control table 10 are set in the CS5. The comparison table 10 has the same configuration as the control table 9, and the SVP 6 sets a value based on control information from the outside (such as the PNL 7). Since the configuration other than CS5 is the same as the above-mentioned conventional example, the description thereof will be omitted.

Description of Example of Relation between Basic Control Table and TM8 ... See FIG. 3. In this embodiment, for each logical drive, each control table 9 has a basic control table (first shown in FIG. 3). Set the control table set in. In addition, a table having the same configuration as the basic control table (additional table) is set in each control table 9 as an additional memory.

Each item of the control table 9 shown in FIG.
It is as follows. That is, "Total CYLNU
“M” indicates the total number of cylinders of the relevant logical drive (changed each time it is added) (N1, N2, ...), and “Top”
“Adrs-Basic” is the start address of the logical drive (the numerical value set at the beginning remains) {(a), (b),
...} is shown.

[Basic CYLNUM]
Indicates the number of cylinders set in the control table 9 (the number of cylinders to be added in the extension table) (N1, N2, ...), and “ADD Pointer” indicates the address of the control table next to the control table 9 of the logical drive. Show. In addition, when the control table 9 is only the basic control table, and the final expansion table of the drive is “0”.
And

As described above, the expansion table is used when increasing the number of cylinders, and the "ADD" of the control table 9 is used.
The address of the control table next to the control table is shown in "Pointer", and when "ADD Pointer" is "0", that table becomes the final control table. And using this "ADD Pointer",
The control table 9 is hierarchized.

The presence / absence of the logical drive is determined by "Total CYLNUM" in the basic control table of each drive (when the value is "0", the logical drive is not configured). As a result, logical drive D
The range of TM8 belonging to RV0 to DRV4 can be dynamically revealed.

When the revision of the control table 9 is performed, the update target TM 8 is reformatted according to the updated format of the control table 9 recognized by the CA 3 in the interrupt processing for the CA 3. In this case, the logical drive that is not the update target can continue processing while remaining online.

: Structure of control table after expansion and T
Description of example of relation of M ... See FIG. 4. FIG. 4 shows a configuration example of the control table and TM8 when the control table shown in FIG. 3 is added.

In this example, M is added to the logical drive DRV0.
1 capacity (number of cylinders) is expanded, and logical drive DRV
This is an example in which 2 is newly added. In this case, before expansion (Fig. 3
State), the number of cylinders of DRV0 is N1, the number of cylinders of DRV1 is N2, the number of cylinders of DRV2 is N3 = 0,
The number of cylinders of DRV3 is N4 = 0.

After the expansion (state of FIG. 4), DRV
The number of cylinders of 0 is N1 + M1 (M1: expansion), DRV
The number of cylinders for 1 is N2, and the number of cylinders for DRV2 is N3
(The number is not 0), and the number of cylinders of DRV3 is N4 = 0.

Therefore, the DRV0 "Total CYL"
Since "NUM" adds M1, the value of N1 + M1 is set. Further, "ADD Pointer" of the basic control table of the logical drive DRV0 indicates the start address of the extension table, and "ADD Pointer" other than the basic control table of the logical drive DRV0 does not have the next table, so "0 Is set.

If the drive does not exist,
"0" is set in "TotalCYLNUM" of the basic control table. By the way, as a need for maintenance in a computer system, active maintenance (replacement of parts without turning off the power of the device, maintenance without stopping system operation) is required. , A method of inserting and removing a printed board has been developed.

In this embodiment, S
Addition / deletion or restoration of the memory printed board (logical drive) in SD is performed. Further, there are three methods of adding the memory, namely, a method of increasing the number of logical drives, a method of increasing the number of cylinders of each logical drive, and a method of combining these two.

Specific examples will be described in detail below. : Explanation of Example 1 (method to increase the number of drives)
-See Fig. 5, Fig. 6, Fig. 7 and Fig. -1: Explanation based on the control table and TM ... See Fig. 5 and Fig. 6 In this example, in the SSD with the maximum number of drives 4 KB
yte memory, normal drive number 4 (DRV0-DR
V3) 0x80 Byte (128 per cylinder)
Byte), DRV0-4 cylinders for each drive,
DRV1-6 cylinder, DRV2-12 cylinder, DR
V3-10 cylinders, which were used under the conditions of 2
Addition of KByte memory, DRV4-6 cylinder,
In this example, three drives, DRV5-2 cylinder and DRV6-8 cylinder, are added.

In the control table example shown in FIG. 5, the drives DRV0, DRV1, DRV2, DRV3 are the logical drives (drives before expansion) that have been used until now, and the drives DRV4, DRV5, DRV6 are logical drives to be expanded. Is.

In the control table of the drives DRV0 to DRV3, "Top AdrsBasic" or
If there is a change in "Basic CYLNUM",
It is determined that the setting of the logical drive is to be redone, and the SVP 6 is made to judge that, and the logical drive is set by the same process as the conventional process.

The structure of the TM8 during the above processing is shown in FIG.
As shown in the figure, the added TM8 has DRV4, DRV
5, DRV6 has been added. In addition, this judgment is C
Although it can be executed in the interrupt process of A3, the execution process is smaller in SVP6, so in this example, the judgment is made in SVP6.

In addition, in the control tables of DRV0 to DRV3, "Top Adrs Basic" or
When there is a change in a place other than "Basic CYLNUM", it will be described later.

-2: Explanation based on flow chart
-See Figs. 7 and 8 Figs. 7 and 8 are process flowcharts in the case of issuing an instruction from the SVP to add another unit, which will be described below with reference to Figs. In addition, S1-S14 of a figure show each process number.

(S1): First, the SVP 6 based on the control information from the outside, the comparison control table 10 in the CS5.
Set the value to. (S2): The SVP 6 instructs the expansion.

(S3): The SVP 6 compares the currently set (online) control table 9 with the comparison control table 10 to determine whether there is any change other than in the expansion section. If there is a change, the processing is performed in the same manner as the conventional example (processing from -4 of the above-mentioned conventional example).

(S4): However, if there is a change, it is confirmed by the SVP 6 whether a normal value is set in the comparison control table 10 (whether there is no address duplication, etc.). If there is a duplication, redefine the control table.

(S5): If a normal value is set in the comparison control table 10, the SVP6 determines C
Interrupt A3. (S6): At this time, in SVP6, a value (value of control information) is redefined in the control table 9 in CS5, and normal operation is performed.

(S7): The CA 3 executes the following process for each drive by the above interrupt. (S8): In CA3, it is checked whether or not processing for the basic drive (maximum number of drives) has been performed, and if the processing has been completed, normal operation is performed.

(S9): However, if the processing is not completed, CA3 is set now (online).
How is the control table 9 compared with the control table to be set (how many drives, how many cylinders)
Check whether to format.

That is, it is checked whether or not there is a change in the control table, and if there is no change, the process for the next drive starts. (S10): The CA 3 stores “Tota” in the control table 9.
Check whether "1 CYLNUM" is "0".

(S11): If "Total CYL
If "NUM" is "0", the drive is deleted (not used), and the drive is taken offline. (S12): Then, the value is redefined in the control table 9 in the CS4.

(S13): In addition, "Total CYL
If "NUM" is not "0", only the add-on unit is made offline and the format process is performed only on the add-on unit. (S14): When finished, C for the added drive
The values are redefined in the control table 9 in CS4 so that A3 can emulate.

(S15): CA3 brings the added drive online. By each of the above processes, it is possible to add the memory while formatting only the expansion unit and leaving the others online.

In this example, three drives were added, but the number of drives that can be added is set by changing the number of basic tables in the control table. Also, as long as the number of drives that can be added is not exceeded, it is possible to add more times.

Description of Embodiment 2 (example of increasing the number of cylinders in each drive) ... See FIGS. 9 and 10. FIG. 9 is an example of a control table when adding cylinders. 10 is a configuration example of TM8 according to the second embodiment, FIG. 10A is a configuration example of TM8 before expansion, and FIG.
Shows a configuration example of TM8 after expansion, and FIG. 10C shows the way of thinking (emulation method) of TM8 viewed from CA3.

-1: Control table and explanation by TM. See FIGS. 9 and 10. In this example, in the SSD having the maximum number of drives 5, 4 KB.
yte memory, normal drive number 4 (DRV0-DR
V3) 0x80 Byte (128 per cylinder)
Byte), DRV0-4 cylinders for each drive,
DRV1-6 cylinder, DRV2-12 cylinder, DR
V3-10 cylinders, which were used under the conditions of 2
Add KByte memory, DRV0, DRV1, D
This is an example of adding 6, 2, and 8 cylinders to each of the three drives of RV3.

As shown in the figure, DRV0 to DRV3 are the drives that have been used until now, and among them, D
Add cylinders to RV0, DRV1 and DRV3 respectively. In the control table of DRV0 to DRV3,
"Top Ads Basic" or "Basic
If there is a change in “CYLNUM”, it is judged that the setting of the logical drive is to be redone, and the SVP 6 is made to judge that, and the logical drive is set in the same process as the conventional one (the reason is the same as in the first embodiment described above). ).

-2: Explanation based on flow chart ...
-Refer to FIG. 7 and FIG. 8 Since FIGS. 7 and 8 described above are common to the second embodiment, the processing of the second embodiment will be described below with reference to FIGS. 7 and 8. In addition, S1-S14 of a figure show each process number.

(S1): First, the SVP 6 uses the control information from the outside, and the comparison control table 1 in the CS 5
Set the value to 0. (S2): The SVP 6 instructs the expansion.

(S3): The SVP 6 compares the currently set (online) control table 9 with the comparison control table 10 to determine whether there is any change other than in the expansion section. If there is a change, it is processed in the same way as the conventional example.

(S4): However, if there is no change, S
The VP 6 confirms whether a normal value is set in the comparison control table 10 (whether there is no address duplication, etc.). If there is a duplication, redefine the control table.

(S5): If a normal value is set in the comparison control table 10, CVP from SVP6.
Interrupt A3. (S6): At this time, in SVP6, the value is redefined in the control table 9 in CS5, and normal operation is performed.

(S7): The CA3 executes the following process for each drive by the above interrupt. (S8): In CA3, it is checked whether or not processing for the basic drive (maximum number of drives) has been performed, and if the processing has been completed, normal operation is performed.

(S9): However, if the processing has not been completed, the currently set (online) control table 9 is compared with the control table to be set, and how (how many cylinders are to be added) is compared. Check whether to format.

At this time, the control table 9 "ADD Po
If it is not "0", the next table is searched hierarchically until "0" is reached. Then, it is checked whether or not there is a change in the control table, and if there is no change, the process for the next drive is performed.

(S10): The CA 3 checks whether "Total CYLNUM" in the control table 9 is "0". (S11): If "Total CYLNUM" is "0", the drive is deleted (not used), so the drive is taken offline.

(S12): Then, the value is redefined in the control table 9 in the CS4. (S13): If "Total CYLNUM" is not "0", only the add-on unit is made offline and the format process is performed only on the add-on unit.

(S14): When the process is completed, CS3 is emulated so that CA3 can emulate the added drive.
4. Redefine values in control table 9 in 4. (S15): The added drive is brought online.

By each of the above processes, it is possible to format the expansion section only and to expand the memory while the other sections remain online. In this example, the expansion was done once, but by specifying "ADD Pointer" one after another, it is possible to connect the control tables hierarchically in any number of stages, so that the expansion can be repeated any number of times. Is possible.

In the process of (S14), the TM8 is actually configured as shown in FIGS. 10A and 10B, but when emulating using the control table 9 in the CS4, the CA3 is set as shown in FIG. 10C. To see TM8.

For example, taking DRV0 as an example, DRV0
When attempting to access from the 0x220 byte of, the CA3 emulates access from the 0x1020 address.

Example 3 (Method Combining the Above Two Methods) This example is an example in which the above two methods, that is, the additional drive and the additional cylinder are combined.

The third embodiment can be realized by having a plurality of extra control tables of the second embodiment. That is, since the processing is a combination of the first and second embodiments, the description thereof will be omitted.

Embodiment 4 (Embodiment for Repairing Memory at Memory Failure) ... See FIG. 11 This example is an example of repair processing when a memory failure occurs at a certain address of TM8. . Hereinafter, the process of the fourth embodiment will be described based on the process flowchart of FIG. Note that S21 to S31 in FIG. 11 indicate respective process numbers.

(S21): The address of the faulty memory is designated by the PNL 7 and the SVP 6 is instructed to perform the repair process (for example, the customer engineer instructs). (S22): The SVP 6 checks the control table 9 in the CS 5 to see which drive the address of the faulty memory is assigned to.

(S23): The SVP 6 interrupts the CA 3 to bring the drive checked in (S22) off line. (S24): In the interrupt process of CA3, the designated drive is made offline.

(S25): The CA 3 notifies the SVP 6 that it has gone offline. (S26): When the SVP 6 confirms that it has gone offline, it notifies the person operating the PNL 7 (for example, a customer engineer) through the PNL 7.

(S27): For example, the customer engineer replaces the faulty memory with a new one. (S28): For example, the customer engineer uses PNL7.
Then, the SVP 6 is instructed to bring the drive that has been taken offline to be online (replacement end instruction).

(S29): The SVP 6 interrupts the CA 3 so as to bring the drive online. (S30): In the CA3, in the interrupt process, the drive format process is performed by referring to the control table 9 (control table in CS4) of the drive that has been taken offline.

(S31): After that, the CA 3 brings the formatted drive online. By each of the above processes, the faulty memory can be repaired simply by taking only the drive of the faulty memory unit offline.

[0096]

As described above, the present invention has the following effects. (1) It is possible to add a memory or repair a faulty memory without taking the device offline. Therefore, the semiconductor disk device can be continuously used,
The operating efficiency (utilization efficiency) of the device is improved.

(2) It is possible to add a memory or repair a faulty memory without taking the device offline. Therefore, it is possible to shorten the processing time when the memory is added or when the faulty memory is repaired.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a block diagram of a semiconductor disk device according to an embodiment of the present invention.

FIG. 3 is a basic control table and T in the embodiment of the present invention.
It is an example of the relationship of M.

FIG. 4 is a configuration example of a control table after expansion and a relation example of TM in the embodiment of the present invention.

FIG. 5 is an example of a control table according to the first embodiment (when increasing the number of drives).

FIG. 6 is a configuration example of TM of the first embodiment.

FIG. 7 is a processing flowchart (part 1: SVP processing) in the case of issuing an expansion instruction from the SVP for expansion.

FIG. 8 is a processing flowchart (part 2: CA processing) in the case where an expansion instruction is issued from the SVP and expansion is performed.

FIG. 9 is an example of a control table according to a second embodiment (when increasing the number of cylinders).

FIG. 10 is an explanatory diagram of a second embodiment (a configuration example of TM).

FIG. 11 is a processing flowchart for repairing (replacing) a memory when a memory failure occurs.

FIG. 12 is an explanatory diagram (1) of a conventional example.

FIG. 13 is an explanatory view (No. 2) of the conventional example.

FIG. 14 is an explanatory diagram of a conventional example (No. 3).

[Explanation of symbols]

 1 CPU (Central Processing Unit) 2 SSD (Semiconductor Disk Unit) 3 CA (Channel Adapter) 4, 5 CS (Control Storage) 6 SVP (Service Processor) 7 PNL (Maintenance Panel) 8 TM (Track Memory) 9 Control Table

Claims (4)

[Claims] 1. A semiconductor memory (TM8) as a storage medium
By using this semiconductor memory (TM8) by dividing it into a plurality of logical drives (DRV0, DRV1, ...) Having a predetermined capacity, the input / output processing of the disk device is virtually executed. In the method for maintaining a semiconductor disk device (2), a control table (9) for storing control information of the logical drive is provided inside the semiconductor disk device (2) for each logical drive (DRV0, DRV1, ... ) For each logical drive (DRV0, DRV) when changing the memory by setting the control table (9) as a hierarchical table.
(1 ...) By updating the control table (9) every time, a memory change process is performed in a state in which logical drives other than the corresponding logical drive can be accessed from the higher-level device (1). A method for maintaining a semiconductor disk device characterized by the above. 2. The control table (9) comprises a basic control table for storing initial setting values of control information, and an expansion control table having the same configuration as the basic control table. 2. The maintenance method for a semiconductor disk device according to claim 1, wherein the expansion control table is hierarchized by setting an address of the next control table. 3. In the semiconductor disk device operating online, when a memory is added, the processor (3, 6) of the semiconductor disk device (2) stores the control table (9) according to an instruction from the outside. Referring to the control information of the logical drive to be added, only the logical drive to be added is taken offline to perform the formatting process of the add-on unit, and the control information of the added logical drive is added to the control table (9). The method for maintaining a semiconductor disk device according to claim 1, wherein the value is redefined, and then the logical drive in the additional section is brought back online. 4. Repair (replacement) of a failed memory in the semiconductor disk device operating online
When performing the processing, the processor (3, 6) of the semiconductor disk device (2) refers to the control table (9) according to an instruction from the outside, acquires the control information of the logical drive to be restored, and restores it. Only the logical drive is taken offline, the external is notified, and then the processor (3, 6) that receives the instruction to complete the replacement of the faulty memory refers to the control table (9), and refers to the offline logical drive. Perform the formatting process, and after finishing the formatting process,
The method for maintaining a semiconductor disk device according to claim 1, wherein the method is returned to the online state.