JPH1049394A - System and method for processing fault - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the system and method for fault processing which can improve the degree of relief of an EMU blocking fault without altering software as to a computer system which controls data transfer between an EMU and an MMU by an EMA. SOLUTION: The computer system controls the data transfer between extended storage device(EMU) 7 and main storage device(MMU) by extended storage connection mechanisms(EMA) 5 and 6. If a fault occurs to the EMAs 5 and 6 while arithmetic processing units(EPU) 1 and 2 are executing EMU synchronous transfer instructions, the EPU reports the fault of the EMAs to the software, and a diagnostic control processor(DGP) 4 performs processing such as the gathering of fault information and then restarts the EMAs 5 and 6. The EPU performs processing by software retrial after the restarting of the EMAs is completed if the restarting is in process. Consequently, EMU blocking is not caused immediately in case of an intermittent fault of the EMAs, and the process can be carried on.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure handling system and method, and more particularly, to a failure handling system and method when an extended storage connection mechanism fails.

[0001]

2. Description of the Related Art Conventionally, in this type of failure processing system, an extended storage device (hereinafter, referred to as an EMU) controls the data transfer with a main storage device (hereinafter, referred to as an MMU). If a failure occurs in the EMA during execution of the EMU synchronous transfer instruction in the computer system connected via the EMA, the software retry (retry) the instruction. Closed.

[0002]

The above-mentioned conventional method has a drawback that, even when an intermittent failure occurs in the EMA, the EMU may be blocked if the software retry fails.

An object of the present invention is to provide a computer system in which EMA controls data transfer between an EMU and an MMU, which can improve the degree of remedy for an EMU blockage failure without changing software. Is to provide.

[0004]

According to the present invention, there is provided a system comprising: a main storage device; an extended storage device; an extended storage connection mechanism for controlling data transfer between the main storage device and the extended storage device; A processing unit, and a failure reporting unit that notifies an operating system of the processing unit of the occurrence of the failure when a failure of the extended storage connection mechanism occurs during execution of the extended storage device synchronous transfer instruction by the processing unit; Restart means for restarting the extended storage connection mechanism in response to the occurrence of the failure; and execution of a command retry process in response to a retry instruction from the operating system when the extended storage connection mechanism is in a restart operation. Waiting means until the restart operation is completed.

A method according to the present invention is directed to a computer system including an arithmetic processing unit and an operating system, wherein control of data transfer between a main storage device and an extended storage device is performed by an extended storage connection mechanism. A failure reporting step of notifying the operating system of the arithmetic processing unit of the occurrence of the failure when a failure of the extended storage connection mechanism occurs during execution of the extended storage device synchronous transfer instruction by the device; and responding to the occurrence of the failure. And restarting the extended storage connection mechanism, and when the extended storage connection mechanism is performing a restart operation, executing a command retry process in accordance with a retry instruction from the operating system. Waiting until completion.

An object of the present invention is to provide a computer system in which EMA controls data transfer between an EMU and an MMU.
An object of the present invention is to provide a failure processing method that improves the degree of relief of an EMU blockage failure without changing software.

[0007]

Next, embodiments of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1 showing a first embodiment of the present invention, arithmetic processing units (hereinafter, referred to as EPUs) 1 and 2 sequentially execute software instructions and execute a system control unit (hereinafter, referred to as SCU). ) 15 and 16 are EPU
It reads and writes data to and from the MMU 3 in response to requests from 1 and 2 and an input / output processing device (IOP, not shown).

The diagnostic control processor (hereinafter, referred to as DGP) 4 includes EPUs 1 and 2, SCUs 15 and 16,
In addition, the MMU3 detects failures and collects failure information.
A communication processing mechanism 14 and a failure processing mechanism 13 are provided.

[0010] The communication processing unit 14 is connected to the diagnostic path 40 and the DGP.
4 and activates the fault handling mechanism 13 in response to a notification from a device connected to the diagnostic path 40.

As shown in FIG. 2C, the failure handling mechanism 13 includes a failure information determination unit 29, a device configuration information table reference / update unit 30, a failure information collection unit 31,
MA restart instructing means 32. Failure handling mechanism 13
Starts the failure information determination unit 29, the device configuration information table reference / update unit 30, the failure information collection unit 31, and the EMA restart instruction unit 32 according to the notification from the communication processing unit 14.

The failure information judging means 29 is a communication processing unit 14
The type of the failure is determined based on the notification from.

Device configuration information table reference / update means 30
Performs reference or update of the device configuration information table 12.

The fault information collecting means 31 performs a process of collecting fault information.

The EMA restart instruction means 32 instructs the EMA diagnosis mechanisms 19 and 20 in the EMAs 5 and 6 to perform EMA restart processing.

The EMAs 5 and 6 perform data transfer between the MMU 3 and the EMU 7 in response to a request from the EPUs 1 and 2, and are provided with EMA diagnostic mechanisms 19 and 20 and EMA restart mechanisms 21 and 22. .

EMA diagnostic mechanisms 19 and 20 control the interface between diagnostic path 40 and EMAs 5 and 6,
The EMA fault is detected and its type and fault occurrence location are determined.

The EMA restart mechanisms 21 and 22 are DGP
4 and performs EMA restart processing. When the EMA restart is completed, an interrupt notification is sent to the DGP 4 through the EMA diagnostic mechanisms 19 and 20.

The diagnostic path 40 stores the failure information of each device in the DGP
4 is a route for notifying the user.

The device configuration information table 12 stored in the MMU 3 is a table indicating the status of the devices connected to the system, and exists only on the MMU 3 in the system.

Interrupt control mechanism 8 in EPUs 1 and 2
And 9 are the EMU by the interrupt notification from DGP4.
Activate the transfer mechanisms 10 and 11.

As shown in FIGS. 2A and 2B, the EMU transfer mechanisms 10 and 11 are provided with device configuration information table reference / update means 23 and 24, EMA state determination means 25 and 26, and an EMU transfer request. Issuing means 27 and 28.

Device configuration information table reference / update means 23
And 24 refer to or update the device configuration information table 12.

The EMA state determination means 25 and 26 determine the EMA apparatus state based on the contents of the apparatus configuration information table 12.

The EMU transfer request issuing means 27 and 28 issue a transfer request to the EMA.

The EPU diagnostic devices 17 and 18 are connected to the EPU 1
And 2 and the interface between the diagnostic path 40 and
EPU failure is detected and its type and failure location are determined.

The communication path 41 includes the EPUs 1 and 2 and the MMU.
3 is a path for transferring data between the EMAs 5 and 6.

Next, the operation of the embodiment of the present invention will be described.

The EMU synchronous transfer command issued by the software (OS) is executed by the EPU 1 and the EMU transfer mechanism 10
Start Here, referring to FIG. 3, the device configuration information table referring / updating unit 23 in the EMU transfer mechanism 10 refers to the device information of each EMA in the device configuration information table 12 (step-D1), and the EMA state determination unit 25 checks whether or not each EMA can be used (step-D2). That is, if there is an available EMA, the configuration information table referring / updating unit 23 checks the E of the device configuration information table 12 for the EMA.
Update the MA device information to “in use” (step-D
3), the transfer request issuing means 27 issues a transfer request via the communication path 40 via the system controller 15 (step-D4). The EM that received the request
A starts data transfer between EMU 7 and MMU 3 or IOP.

On the other hand, EMA in a usable state
If there is no, the EMA state determination means 25
Check whether the MA is in use by another EPU2 (step-
D5) If in use, return to the beginning of the processing of EMU transfer mechanisms 10 and 11.

On the other hand, if a certain EMA is not being used by another EPU 2, the EMA state determination means 25 sets
It is checked whether the MA is restarting (step D6). If the MA is restarting, the configuration information table reference / update means 2
3 refers to the device information of the EMA in the device configuration information table 12, waits until the restart of the EMA is completed and the EMA becomes usable, and then the transfer request issuing unit 27 sends the request via the system control device 15. Communication path 4
Issue a transfer request via Step 1 (Step-D7,
D8 and D9).

On the other hand, if the EMA is not being restarted, there is no usable EMA, and the OS is notified that the transfer command has been abnormally terminated (step D10).

Upon detecting an EMA failure, the EMA diagnostic mechanisms 19 and 20 collect failure information, and notify the DGP 4 of the occurrence of the EMA failure via the diagnostic path 40 as an interrupt. The communication processing mechanism 14 in the DGP 4
Upon receiving a failure notification from the MA diagnosis mechanism 19 or 20, the failure processing mechanism 13 is activated.

Referring to FIG. 4, in the failure handling mechanism 13, the device configuration information table reference / update means 30 refers to the EMA device information in the device configuration information table 12 (step-C1). The failure information determination means 29 checks whether or not the failure is an EMA failure (step-C2). EM
If the fault is A, the fault information collecting means 31 collects the fault information of the EMA from the EMA diagnostic mechanism 19 or 20 (step-C3). The device configuration information table referring / updating unit 30 stores the E in the device configuration information table 12.
The device information of the MA is changed from “in use” to “restarting” (step-C4). The EMA restart instruction means 32 issues a restart instruction to the EMA diagnostic mechanism of the EMA (step-C5). On the other hand, if it is not a failure of the EMA, it is checked whether it is a notification of the completion of the EMA restart (step-
C6) If the notification is the completion of the restart, the device configuration information table reference / update means 30 changes the device status of the EMA in the device configuration information table 12 from "restarting" to "usable" (step-C7). . If it is not the notification of the completion of the EMA restart, the processing shifts to another failure processing.

The EMA diagnostic mechanism 19 or 20 is a DGP4
EMA restart mechanism 2 upon receiving an EMA restart instruction from
1 or 22 is started. The EMA restart mechanism 21 or 22 performs EMA restart processing, and when the restart is completed, the EMA diagnostic mechanism 19 or 20 sends the EGP to the DGP 4.
An interrupt notification is sent via the diagnostic path 40 that the MA restart has been completed.

Next, the operation of the embodiment of the present invention will be described in more detail with reference to FIGS. 3, 4 and 5.

When the OS (software) issues an EMU synchronous transfer instruction (sequence-A1), and this is executed by the EPU1, the EMU transfer mechanism 10 determines that the EMA5 is usable and transfers the EMU to the EMA5. A request is issued (sequence-A2, steps-D1 to D4).

EMA5 which received the transfer request is EMU
Data transfer is started. If a failure occurs in the EMA 5, the EMA diagnosis mechanism 19 of the EMA 5 notifies the DGP 4 of the occurrence of the failure by interruption, and furthermore, the EPU 1
Also notifies that the transfer has failed (sequence-A
3).

The DGP 4 that has received the notification from the EMA diagnosis mechanism 19 of the EMA 5 activates the failure processing mechanism 13. The device configuration information table reference / update means 30 in the failure processing mechanism 13 refers to the device information of the EMA5 in the device configuration information table 12 (step-C1). Failure information determination means 29
Checks whether there is an EMA failure based on the notification content from the EMA diagnosis mechanism 19 (step-C2). EMA
In the case of the fault, the fault information collecting means 31 collects the fault information of the EMA 5 from the EMA diagnostic mechanism 19 (step-C3), and refers to the device configuration information table / update means 30.
Changes the device information of the EMA 5 in the device configuration information table 12 from “in use” to “restarting” (step-C).
4). The EMA restart instruction means 32 issues a restart instruction to the EMA diagnosis mechanism 19 of the EMA 5 (step-C).
5, Sequence-A4).

The EPU 1 receives the transfer failure notification from the EMA 5 and reports to the OS (software) that the synchronous transfer command has been abnormally terminated (sequence-A 5).

The OS (software) issues a synchronous transfer instruction again (hereinafter, software retry) (sequence-A6) because the synchronous transfer instruction has been abnormally terminated.

When the EMU synchronous transfer command is reissued by software retry, the device configuration information table reference / update means 23 in the EMU transfer mechanism 10 can use the E
Check for MA, but EMA5 is restarting and EMA6 is out of order.
Wait for the transfer request to be issued (Step-D
1, D2, D5, D6, D7, D8, sequence-A
7).

The EMA restart mechanism 21 performs a restart process of the EMA 5, and when the restart is completed, the EMA diagnostic mechanism 19
Notifies the DGP 4 of the completion of the EMA restart via the diagnostic path 40 (sequence-A8).

The communication processing unit 14 that has received the notification from the EMA diagnosis unit 19 activates the failure processing unit 13, and the device configuration information table reference / update unit 30 in the failure processing unit 13 reads the device information of the device configuration information table 12. With reference to (Step-C1), the failure information determining means 29 determines the type of the failure and checks whether the failure is an EMA failure (Step-C1).
2) When it is determined that the notification is the restart completion notification from the EMA 5 (step-C6), the device configuration information table reference / update unit 30 causes the EMA of the device configuration information table 12 to be updated.
The device status of No. 5 is changed from "restarting" to "usable" (step-C6, sequence-A9).

The device configuration information table reference / update means 23 in the EMU transfer mechanism 10 refers to the EMA device status in the device configuration information table 12 and waits for processing until the restart of the EMA 5 is completed and the EMA 5 becomes usable. However, since the EMA 5 has become available, the transfer request issuing means 27 issues a transfer request via the communication path via the system controller (steps-D7, D7).
8 and D9, sequence-A10).

EMA5 receiving the transfer request is EMU
The transfer is started, and when the transfer is completed, the EPU is notified that the transfer is completed (sequence-A11).

The EPU receives the notification from the EMA 5 and
The software (OS) is notified that the U data transfer has been completed (sequence-A12).

[0048]

The effect of the present invention is that when an EMU synchronous transfer fails due to an intermittent failure of the EMA in a system configuration in which the number of EMAs becomes one due to a failure or the like, the EMU blocking failure can be performed without changing the software. That is, the degree of relief can be improved.

The reason is that when an EMA intermittent failure occurs, the EM
A restart and the EMA is the only EMA in the system
If, the EM is restarted until the restart of the EMA is completed.
This is to wait for the U transfer process in the EPU.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram showing details of each unit in FIG. 1;

FIG. 3 is a flowchart illustrating an operation of an EMU transfer mechanism of the EPU according to the present invention.

FIG. 4 is a flowchart showing the operation of a DGP failure handling mechanism according to the present invention.

FIG. 5 is an EMU synchronous transfer operation sequence diagram showing an example of a failure process according to the present invention.

[Description of Signs] 1-2 arithmetic processing unit (EPU) 3 main storage unit (MMU) 4 diagnostic processing unit (DGP) 5-6 extended storage connection mechanism (EMA) 7 extended storage unit (EMU) 8-9 interrupt control Mechanisms 10 to 11 EMU transfer mechanism 12 Device configuration information table 13 Fault handling mechanism 14 Communication processing mechanism 15 to 16 System control unit (SCU) 17 to 18 EPU diagnostic device 19 to 20 EMA diagnostic mechanism 21 to 22 EMA restart mechanism 40 Diagnosis Path 41 Communication path 10-11 EMU transfer mechanism 23-24 Device configuration information table reference / update unit 25-26 EMA state determination unit 27-28 Transfer request issuance unit 29 Failure information determination unit 30 Device configuration information table reference / update unit 31 Fault information collecting means 32 EMA restart instruction means

Claims (3)

[Claims] A main storage device, an extended storage device, an extended storage connection mechanism for controlling data transfer between the main storage device and the extended storage device, an arithmetic processing device, and expansion by the arithmetic processing device. A failure reporting unit for notifying the operating system of the arithmetic processing unit of the occurrence of the failure when a failure of the extended storage connection mechanism occurs during execution of the storage device synchronous transfer instruction; and Restart means for restarting a storage connection mechanism; and when the extended storage connection mechanism is in a restart operation, causing the execution of a retry process of an instruction in response to a retry instruction from the operating system to wait until the restart operation is completed. A fault handling system comprising a waiting means. 2. The system according to claim 1, further comprising a plurality of said extended storage connection mechanisms, wherein said restarting means examines a state of said plurality of extended storage connection mechanisms and performs said restart based on a result of the investigation. Fault handling system as described. 3. A computer system including an arithmetic processing unit and an operating system, wherein control of data transfer between a main storage device and an extended storage device is performed by an extended storage connection mechanism. A failure reporting step of notifying the operating system of the arithmetic processing unit of the occurrence of the failure when a failure of the extended storage connection mechanism occurs during execution of the transfer instruction; and A restarting step of restarting a mechanism; and a waiting step of, when the extended storage connection mechanism is performing a restarting operation, waiting for completion of the restarting operation until execution of a retry process of an instruction according to a retry instruction from the operating system is completed. And a failure handling method.