JPH08212014A - Semiconductor storage subsystem - Google Patents

Abstract

PURPOSE: To shorten fault recovery processing time by providing a semiconductor control device with a means for initializing a track table (TTL) at the time of executing an instruction applied from the external or executing a format writing command when the TTL is failed. CONSTITUTION: When a semiconductor storage control device(SSC) 502 receives a format writing command 510 and a TTL fault is generated on an indicated track, the TTL is initialized by TTL initialization 511. When a fault is generated in the TTL 505, the fault generation is informed to the SSC 502 by a fault report 508. After receiving the fault report 508, the SSC 502 informs the fault generation to a CPU 501 by a fault report 509. A maintenance man finds out a failed track based on the report 509 and the initialization of the TTL 505 is requested by a service processor(SVP) 503. Receiving a TTL initializing instruction, the SSC 502 initializes the TTL 505 and track data 506 stored in a semiconductor storage unit(SSU) 504.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor memory device which is one of the memory devices used in a computer system, and more particularly to a semiconductor memory device having means for recovering from a failure in control information. Regarding the device.

[0002]

2. Description of the Related Art As external storage devices of computer systems, there are magnetic disk devices and magnetic tape devices. The storage format (track format) of the magnetic disk device is shown in FIG.

The index marker 101 indicates the beginning of a track. The home address 102 (hereinafter abbreviated as HA) is the first field of the track and indicates the status of the track and the track number.

Counting units 103, 105 and 108 (hereinafter abbreviated as C unit) record records 109, 110 and 111.
Is the first field of the record and indicates the status, position and length of the record. The key section 106 records information that serves as a guide for the subsequent data section, and is used for searching the contents of the data section. The key part does not have to exist in the record. User data is recorded in the data sections 104 and 107. C
A record is composed of a part, a key part, and a data part.

The record immediately after HA is record 0 (10
9), and in the standard format, there is no key part. The records on the track are generally record 0, record 1, record 2, ... Record n. On the initialized track, HA102,
Only record 0 (109) exists.

Semiconductor memory device (hereinafter abbreviated as SSU)
However, normally, a magnetic disk device is emulated and each record is stored in a similar storage format. Unlike magnetic disk devices, SSUs do not have physical restrictions such as rotation of storage media. Therefore, in the record search, it is not necessary to sequentially read the data (C portion) on the track from the beginning, and the search time can be shortened by managing the record information on the track separately from the actual data. it can. This information is a track table (hereinafter abbreviated as TTL).

The record retrieval by the magnetic disk unit and the SSU instructed by the host unit is as follows.
The positioning instruction from the host device is executed by the command group shown in FIG.

The SEEK command 201 designates a cylinder number and a head number to search for a specific track.
The SET SECTOR command 202 (hereinafter abbreviated as SS) designates a position from the beginning of the track and processes the command following this command from that position. SEA
RCH ID command 203 (hereinafter abbreviated as SSI)
Compares the data from the higher-level device with the data in the C section, and positions the record on which the condition is satisfied. T
The IC command 204 is for repeating the SCI when the positioning is unsuccessful in the SCI 203. REA
The D / WRITE command 205 reads or writes a field, record, or track.

Here, it is assumed that the SS202 data is "0" (from the beginning of the track) and the SCI positioning is record 5. In the magnetic disk device, the processing is started from HA as shown in FIG. 3, and the record 0 (302) and the record 1 (30
3), record 2 (304), record 3 (305),
Go to record 4 (306) and record 5 (30
Positioned in 7). Therefore, useless head movement 308 occurs. During this time, the connection with the host device is maintained and unnecessary head movement occurs.

In the case of SSU, when SCI 203 is executed, the target record is retrieved based on TTL. TTL is
The structure shown in FIG. 4 is used. The cylinder / head number indicates the cylinder number and head number of this track. The maximum record number 402 indicates the maximum record number existing in this and the rack. When executing the SCI 203, it is possible to easily determine whether the target record exists on the track by the maximum number of records 402. The maximum sector number 403 indicates the maximum sector number in which a record exists in this track. If the sector number transferred by SS is larger, the processing of the subsequent command is performed from the beginning of the track. Record information 404, 405, 406
Indicates the information of the record existing in this track. In this record information, the sector number of the own record, the position on the memory where the own record starts, etc. are stored. TTL
Is a format write command (WRITE HA, WRITE
The entire track such as the ECKD command or the record is created or erased).

When the SSU retrieves the same data as in FIG. 3, the presence or absence and the position of the target record are obtained by referring to the TTL after receiving the SCI. 5 for the maximum sector number 403
The above values are stored, and it is found that the target record exists. Next, by referring to the record information corresponding to the record 5, the position on the memory is obtained. Positioning to the target record is performed by this value. By this operation, it is possible to avoid unnecessary head movement that has occurred in the magnetic disk device.

When the TTL fails, it becomes impossible to access all the records of this track. For this reason, the occurrence of a TTL failure is reported to the host device and failure recovery is requested. At the time of failure recovery (or after failure recovery), TTL initialization (reformatting) is required. At this time, TTL
And the actual track data must match.

The SSU has the memory initialization (only HA and record 0 exist on the track, and the corresponding TTL is created) immediately after the power is turned on. Conventional SS
U uses this function to initialize the TTL. That is, all the track data of the SSU in which the failure has occurred first is read, and the data is saved to another storage device (such as a magnetic tape). Then, the SSU is restarted manually to initialize the track data and the TTL. Finally, the data saved in this SSU is written. This work requires a lot of time at present because the capacity of the SSU is several gigabytes.

[0014]

In the conventional SSU, the T
When a failure occurs in the TL, it takes a long time for a manual restart process to recover the failure and for saving and writing data, and during that time, the processing capability of the host device is significantly reduced.

It is an object of the present invention to provide a semiconductor control device that shortens the time required for recovery processing when a failure occurs in TTL and prevents the processing capacity of a host device from decreasing.

[0016]

A semiconductor control device according to the present invention is a means for initializing TTL according to an instruction from the outside.
Alternatively, it has means for initializing the TTL when the format write command is executed at the time of the TTL failure.

[0017]

According to the present invention, since only the TTL can be initialized without backing up the track data at the time of the TTL failure, the TTL failure can be recovered in a short time. Also, T
If the format write command is executed at the time of TL failure, T
By initializing the TL, the TTL failure can be recovered only by the error recovery process from the host device.

[0018]

An embodiment of the present invention will be described below.
FIG. 5 is a block diagram of the computer system of this embodiment. This computer system uses track data 5
06 and TTL505 are stored in the semiconductor memory device (SS
U) 504, a semiconductor memory control device 502 (hereinafter abbreviated as SSC) that controls this SSU, and a central processing unit (hereinafter abbreviated as CPU) that is a host device connected to the SSC.
Has 501.

6 and 7 show TTL 505 by SSC.
7 is a flowchart showing the operation of recovery processing in case of failure.

First, the recovery processing for initializing the TTL by the error recovery processing from the host device will be described with reference to FIGS. 5 and 6.

When a failure occurs in the TTL 505 (step 601), the failure occurrence is indicated by the failure report 508 in the SS.
Reported to C502. SSC 502 reports a problem 508
After the receipt of the message, the failure report 509 reports the failure occurrence to the CPU 501 (step 602). In addition, SSC50
2 is the TTL 505 in which a failure has occurred as TTL failure information
The cylinder number and the head number of the memory are stored (step 6).
03).

The CPU 501 which has received the failure report 509 indicates that the failure is due to a track read / write operation of the magnetic disk device.
Since it looks like a permanent error, the format write command 510 is issued as a failure recovery process to instruct the reformatting of the failed track.

SSC 502 sends format write command 510
When the TTL failure is received (step 604), it is determined by referring to the TTL failure information whether or not a TTL failure has occurred in the track for which format writing is instructed (step 60).
5). If a failure has occurred, the TTL is initialized by the TTL initialization 511 (step 606). After the completion of the TTL initialization, the process of the format write command is executed (step 607). By the above operation, the T
TL write processing can be executed.

In the above-described processing, either the WRITE HA command (rewriting the HA and subsequent fields) or the WRITE R0 command (rewriting the R0 and subsequent fields) may be used as the format write command. Since the TTL stores the information from the record 0, the TTL initialization is triggered by the format write command including the record 0. If this condition is satisfied, WRITE HA command, WR
Not only one of the TER commands but also both may be triggered. In addition, WRITE HA command and WR
It may be triggered by both the TER0 command being instructed. In this embodiment, since the WRITE HA command and the WRITE R0 command are issued consecutively, the TTL initialization can be triggered by any of the above methods.

Next, a recovery process for manually performing the TTL initialization will be described with reference to FIGS. 5 and 7.

The SSC 502 of this embodiment has a service processor 503 (hereinafter abbreviated as SVP). S
The VP 503 collects maintenance information of the SSC 502, reports when there is a failure, and the like. The SSC 502 has a function of performing processing according to an instruction from the SVP 503.

When a failure occurs in the TTL 505 (step 701), the failure occurrence is indicated by the failure report 508 in the SS.
Reported to C502. SSC 502 reports a problem 508
After the receipt of the error, the failure report 509 reports the failure occurrence to the CPU 501 (step 702). After this, all processing to this track will report a TTL failure.

According to the trouble report 509, the maintenance staff seeks a trouble track and requests the TTL initialization by the SVP 503. Upon receiving the TTL initialization instruction from the SVP 503 (step 703), the SSC 502 initializes the TTL (step 704) and track data for the SSU 504.

By the above processing, it becomes unnecessary to back up all the data in the SSU, manually restart the SSU, and rewrite the backed up data in the recovery processing at the time of the TTL failure.

[0030]

According to the present invention, during the recovery process at the time of the TTL failure, the initialization is performed in synchronization with the failure recovery process of the CPU, so that the TTL failure recovery time can be greatly shortened and the CPU can be recovered. Additional reduction is possible.

Further, when the maintenance staff instructs the TTL initialization from the SVP, it is possible to confirm the track and optimize the TTL initialization timing in addition to shortening the TTL failure recovery time.

[Brief description of drawings]

FIG. 1 is a diagram showing a track format on a storage medium of a conventional magnetic disk device.

FIG. 2 is a diagram showing a command chain from a higher-level device.

FIG. 3 is a diagram showing a track format on a storage medium of a conventional magnetic disk device.

FIG. 4 is a diagram showing a TTL of SSU.

FIG. 5 is a configuration diagram of a computer system according to an embodiment of the present invention.

FIG. 6 is a flowchart showing an operation of a recovery process for performing TTL initialization by the error recovery process from the host device according to the embodiment of the present invention.

FIG. 7 is a flow chart showing the T
9 is a flowchart showing an operation of a recovery process for performing TL initialization.

[Explanation of symbols]

401: Cylinder / head number 402: Maximum number of records 403: Maximum sector number 404, 405, 406: Record information 501: CPU 502: SSC 50
3: SVP 504: SSU 505: TTL 50
6: Track data 507: TTL failure information 508, 509:
Failure report 510: Format write command 511: TTL initialization

Claims (2)

[Claims] 1. A semiconductor memory device that stores information and management information that manages the information, and a semiconductor memory control device that is connected to a host device and that controls data transfer between the semiconductor memory device and the host device. A semiconductor memory subsystem having the semiconductor memory subsystem, wherein the management information is initialized by an external instruction. 2. The semiconductor storage subsystem according to claim 1, wherein when a failure occurs in the management information, or when a failure occurs in the management information, the semiconductor control device sets a high-ranked area in a higher order. A semiconductor storage subsystem, which reports to a device and initializes the management information when the semiconductor control device receives an initialization process for an area in which the management information has a failure from the higher-level device.