JP2002358170A - Disk storage device, computer system provided with the disk storage device, and error notification method at retry processing in the computer system - Google Patents

Abstract

(57) [Summary] [Problem] To suppress the occurrence of a timeout error in a host system. A microprocessor provided in a control unit of an HDD, when an error occurs in execution of an access command given from a host system,
A retry process including a retry step group is started.
When the host system 10 sends an access command,
The response from the DD 20 is monitored with the set monitoring time as the upper limit, and if a timeout occurs or an abnormal end is notified within the monitoring time, a reset request is sent first, and then the same command as before is sent to the HDD 20. When receiving the reset request during the reset process, the microprocessor 221 turns on the reset flag during retry in the retry management area 224, and notifies the host system 10 of the specific error code when receiving the same command as the previous command in this state. I do. The host system 10 extends the monitoring time according to the notification.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system having a disk storage device, and more particularly, to a computer system provided by a host system for executing an access request for data in the disk storage device which is given to the disk storage device. The present invention relates to a disk storage device suitable when an error occurs and retry processing is executed, a computer system including the disk storage device, and an error notification method at the time of retry processing in the computer system.

[0002]

2. Description of the Related Art In a computer system provided with a disk storage device, for example, a magnetic disk device (hereinafter, referred to as an HDD), an HDD and a host system using the HDD have a host interface, for example, a SCSI (Small Co.).
Generally, they are interconnected by a bus (mputer System Interface).

In such a computer system, the HD
When the access request is given from the host system, D executes the requested access processing. If the requested access processing is not completed normally and an error (internal error) occurs, an error retry is performed inside the HDD according to a predetermined procedure. The error retry performed inside the HDD is called HDD retry.

In recent years, HDD error retries (H
The DD retry function has been greatly enhanced. In other words, recent HDDs are configured to execute an enormous number of retries in a predetermined order. Therefore, in the access processing requested by the host system, the HD
Even if an error occurs inside D, there is a high possibility that the error will be remedied by this enhanced error retry.

[0005] On the other hand, when the host system issues an access request to the HDD, the host system monitors the response of the access request from the HDD in, for example, a status phase for a time. That is, the host system is HDD
Monitor the processing time of. If no response is returned from the HDD even after a certain monitoring time has elapsed, the host system detects a timeout. In this case, the host system first sends a reset request (for example, SCSI reset) for forcibly resetting (initializing) the HDD to the HDD. The HDD initializes the HDD in response to a reset request from the host system.

When the host system resets the HDD in response to the detection of the timeout, the host system performs a retry for issuing the same access request as the timeout has occurred to the HDD. This retry of the host system is called host retry. The HDD executes the access request issued again from the host system.

[0007]

As described above, in recent HDDs, the error retry function has been strengthened, and as a result, there is a higher probability that errors will be relieved as compared with earlier HDDs.

However, the number of retry steps is significantly increased, for example, due to the enhancement of the error retry function of the HDD, and the processing time required for retry is also increasing. In the worst case, that is, when the retry (HDD retry) is performed up to the final step, the processing time is on the order of tens of seconds to tens of seconds.

Therefore, in a computer system having an HDD with an enhanced error retry function, a timeout (timeout error) may be detected in the host system during HDD retry processing. If a timeout error occurs, a reset request is sent from the host system and subsequent host retry
The HDD again executes the same access processing as that in which the error occurred first. In this case, the same location (address) as before
May cause an error. If an error occurs again, the HDD takes a long time to start the retry process from the first retry step, and there is a possibility that a timeout error occurs again.

As described above, in recent years, the error retry function of the HDD has been enhanced to improve the probability that an error will be rescued. However, since the time required for the retry process increases, it takes more processing time than the monitoring time of the host system. In some cases, a timeout error may be detected. When a timeout error is detected, the host system performs a retry and issues the same request to the magnetic disk device again. However, there is a possibility that a retry occurs at the same location and a timeout error occurs again.

Therefore, recently, for example, Japanese Patent Application Laid-Open No. H10-161
As described in Japanese Patent Publication No. 818, an access request from a host system, an address specified by the request, and a retry step (number) being executed when a reset request is transmitted upon detection of a timeout error in the host system
Is stored in the HDD, and when the same access request is issued from the host system due to the host retry, normal processing is performed until immediately before the address being executed at the time of the previous reset, and There has been proposed a technique for shortening the retry processing time by executing the HDD retry from the retry step that was being executed at the time of the previous reset (hereinafter, referred to as prior art).

However, in the above-mentioned prior art, although the time required for the retry processing again when an error occurs during execution of the access request again can be shortened, depending on the monitoring time set on the host system side. During the retry process, a timeout error may be detected again in the host system.

The cause of the detection of the timeout error in the host system may be (1) the case where the HDD retry takes a long time, and (2) the case where the HDD itself is in failure or is in a locked state. However, in the above-described prior art, there is a problem that a factor for detecting a timeout error cannot be specified.

Further, in order to reduce the occurrence of a timeout error, the monitoring time on the host system side is set to be sufficiently longer than the time necessary for the retry processing on the HDD side to be reliably performed until the final retry step. It is possible. However, the cause of the HDD not responding within the monitoring time may be that the HDD itself has failed as described above. Even in such a case, it is useless to set the monitoring time to a fixed length.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to detect a time-out error in a host system during a retry process in a disk storage device and to issue a reset request from the host system to the disk storage device. If given, the event is recorded, and in that state, if the same access request as the one that made the retry earlier is given from the host system to the disk storage device, A disk storage device having a configuration in which a specific error code is notified to the system to extend the monitoring time in the host system to suppress occurrence of timeout error detection, a computer system including the disk storage device, and the computer An object of the present invention is to provide an error notification method at the time of retry processing in a system.

[0016]

A disk storage device according to the present invention is a means for executing an access request received from a host system, and executes a predetermined retry step when an error occurs in execution of the access request. Means for executing a retry process consisting of a group, storage means for securing a specific area, and a time-out error in which a response from the disk storage device regarding execution of an access request sent by the host system was not returned within the set monitoring time. When a reset request sent to the disk storage device is received from the host system during execution of the retry process by being detected by the host system, retry-in-progress reset information indicating that fact is stored in the specific area of the storage means. The retry information management means to be registered and the access request sent from the host system Is executed by the access request executing means. If the access request is the same request as the previous request and the reset information during retry is registered in a specific area of the storage means, execution of retry processing is performed. Means for notifying the host system of a specific error code indicating that the reset request has been received and requesting the host system to extend the monitoring time, and causing the access request execution means to execute the access request. It is characterized by having.

Here, when the initialization processing according to the reset request is performed in a state where the reset information during retry is registered in the specific area of the storage means, at least the specific area is excluded and the reset information during retry is It is recommended that the configuration be saved.

The same request (the same access request as the previous one) is not limited to the same access request including the start address and the size of the access request in which the previous error occurred, but at least the last retry processing. Refers to an access request that includes the address at which the request was made.

In the disk storage device having such a configuration,
When an error occurs during execution of an access request given from the host, a retry process including a predetermined retry step group is started. In the recent retry processing, as described in the section of the related art, the processing time is increasing due to the strengthening of the internal retry of the disk storage device. Therefore, the time required for the retry process often exceeds the monitoring time set in the host system. In such a case, a timeout error is likely to be detected in the host system during execution of the retry processing. The host system is
When a time-out error is detected or the retry processing is executed to the end in the disk storage device, that is, until the final retry step, the error has not been resolved and the disk storage device reports an abnormal end indicating an error factor. In this case, a reset request is sent to the disk storage device, and thereafter, a host retry for sending the same access request to the disk storage device as before is executed.

In the disk storage device, when a reset request is received from the host system during execution of the reset process, retry-in-progress reset information as log information indicating this is held in a specific area. In this state, when the disk storage device receives the same access request from the host system as before, it indicates that the disk storage device received the reset request during the execution of the retry process and extends the monitoring time. Is notified. The host system is notified by the specific error code notification to the reset request that the disk storage device is in the process of retrying at the time of the reset request, that is, monitoring is performed because a timeout error occurs during the retry process because the currently set monitoring time is short. It is determined that extension of the time is required, and the monitoring time is set to be extended. The extended setting of the monitoring time makes it possible to reduce the occurrence of a timeout error in the host system. Further, the host system can determine that the disk storage device itself has not failed due to the short monitoring time due to the cause of the timeout error, based on the notification of the specific error code from the disk storage device accompanying the detection of the timeout error of the host system. Conversely, if there is no specific error code notification from the disk storage device due to the detection of the timeout error, the host system can determine that the disk storage device may have failed (or locked).

As described above, according to the present invention, when a reset request is received from the host system during the retry processing in the disk storage device, the occurrence of the event is recorded as a log. Is received, a specific error code is notified from the disk storage device to the host system, and the monitoring time is extended to the host system, so that the occurrence of a timeout error in the host system can be reduced, and the timeout error can be reduced. The factors can be easily determined by the host system.

Here, the access request and the address subjected to the retry processing and the retry step identification information indicating the retry step being executed are registered in the specific area, and the above-mentioned specification is performed each time the retry step is switched in the retry processing. The retry step identification information in the area is updated, and when the same access request as the target of the retry processing is received from the host system, it becomes the target of the previous retry processing registered in the specific area. Access to the specified address is performed by a retry process from the retry step indicated by the retry step identification information on the specific area,
That is, if the retry is started in the middle of the previous retry processing stopped by the reset request from the host system, the processing time can be reduced while increasing the probability of error recovery.

[0023]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a computer system according to one embodiment of the present invention.

The computer system shown in FIG. 1 includes a host system (host computer) 10 and a disk storage device used as an external storage device of the host system 10 by the host system 10, for example, an HDD (magnetic disk device) 20. Be composed. Host system 10 and HD
D20 is, for example, S as a host interface bus.
They are interconnected by a CSI bus 30. Note that SC
Fiber Channel 1 instead of SI bus 30
(Fibre Channel) or the like can also be used.

The HDD 20 includes a drive unit 21 including a disk drive and a control unit 2 for controlling the drive unit 21.
2 is provided. Note that a host interface and the like that are connected to the SCSI bus 30 and serve as an interface with the host system 10 are omitted.

The control unit 22 includes a microprocessor 221 which forms the center of the control unit 22, a ROM (Read Only Memory) 222 in which a control program is stored in advance, and a RAM which provides a work area for the microprocessor 221.
(Random Access Memory) 223.

The microprocessor 221 performs read / write access to the drive unit 21 specified by the read / write command from the host system 10 by R.
Control is performed according to a control program stored in the OM 222. The microprocessor 221 executes a retry process including a group of predetermined retry steps when an error occurs in read / write access. Further, when the SCSI bus reset is given from the host system 10 during the retry processing, the microprocessor 221 records the fact in the RAM 223, and in this state, the host system 10 gives the command that caused the retry. In this case, a specific error code is notified to the host system 10.

The RAM 223 has a retry management area 2
24 are secured. The retry management area 224 is shown in FIG.
Field 224a in which the type (command type) of the retried access command is set as shown in FIG.
And a field 224b in which a start address specified by the command is set, a field 224c in which the size is set similarly, and a field 224d in which a retry (error) address (error address) is set.
And field 2 in which the number of the retry step is set
24e and a field 224f in which a reset flag during retry is set. The reset flag during retry indicates that the host system 10
D20 indicates whether a SCSI bus reset (reset request called as a reset request) has been received (if ON) or not (if OFF).

Next, the operation of the computer system shown in FIG. 1 will be described with reference to flowcharts shown in FIGS. First, it is assumed that an access command has been transmitted from the host system 10 to the HDD 20. This command includes the start address (start address) of the access destination
And information on the size of the access data.

Upon receiving the access command from the host system 10, the microprocessor 221 provided in the control unit 22 in the HDD 20 refers to the retry management area 224 secured on the RAM 223 (Step S1). Then, the microprocessor 221 determines whether or not the received command is the same as the command registered in the retry management area 224, and more specifically, whether or not the received command is the same as the command for which the retry processing was previously performed. Is determined (step S2). Here, it is assumed that the command is the same when the command type, the start address, and the size match.

If the commands are not the same, the microprocessor 221 clears the information in the retry management area 224 (step S3), executes the received command, and executes the start address (start sector block) specified by the command. Drive section 21 starting with address)
Control to read out data of the specified size from the area of, or control to write data of the specified size to the area of the drive unit 21 starting from the specified start address (step S4). When the command has been executed normally (step S5), the microprocessor 221 determines that the SCS
The host system 10 is notified of the normal end using the status phase of the I bus 30 (step S6).

On the other hand, when an error occurs during execution of a command, that is, when an error occurs during access to a certain address (sector block address) (step S
5), the microprocessor 221 registers the information of the command being executed and the address at which the error occurred (error address) in the retry management area 224 (step S).
7). Here, the command information includes a command type, a start address, and a size, the command type is in a field 224a of the retry management area 224, the start address is in a field 224b, and the size is in a field 2
24c are registered. The error address is registered in the field 224d of the retry management area 224.

Next, the microprocessor 221 sets a variable N indicating the number of the retry step to an initial value, for example, 1 (step S8). And microprocessor 2
21 is a field 224 of the retry management area 224
The content of e (retry step number) is updated to N (step S9), and the retry step N, that is, the Nth retry step in the retry process is executed (step S9).
10). As is apparent, during the execution of the retry processing, the information (command type, start address, and size) of the command to be retried and the error address (the target of the retry processing) are stored in the retry management area 224. Address) and the number N of the retry step being executed.

If the error has not been resolved by executing the retry step N for accessing the error address (step S11), the microprocessor 2
21 increments N by 1 (step S1).
3) The processing after step S9 is executed again. The microprocessor 221 repeatedly executes the processing from step S9 onward until the retry succeeds (step S11). However, if the retry is not successful even after executing the retry processing up to the predetermined retry step (steps S11 and S12), the microprocessor 221 uses the status phase to make an error to the host system 10. The end is notified (responded) (step S17). This abnormal end notification is accompanied by information of an error location (error address).

On the other hand, when the retry step N can be executed normally, that is, when the error is eliminated by the retry step N (step S11), the microprocessor 221 clears the contents of the retry management area 224 (step S11). Step S14). Next, if there is an unexecuted address among the accesses in the address range specified by the command requested from the host system 10 (step S15), the microprocessor 221 targets the address following the address succeeding the retry. And execute the command (step S16). Then, if the command execution has been performed normally (step S5), the microprocessor 221 notifies the host system 10 of the normal end in step S6 and ends the processing (step S6). If it is (step S5), the processing after step S7 is performed.

On the other hand, if the access command from the host system 10 is the same as the command for which the retry process was performed last time (step S2), the microprocessor 2
21 is a field 224f of the retry management area 224
It is determined whether or not the retry-in-reset flag set to is ON (step S21).

If the reset flag during retry is in the ON state (step S21), the microprocessor 221 sends the host system 1 during the last reset processing.
From 0, it is determined that a SCSI bus reset has been received. In this case, the microprocessor 221 turns off the reset flag during the retry (step S22) and notifies the host system 10 of a predetermined specific error code (step S23).
Proceed to 24. On the other hand, if the reset flag during the retry is not in the ON state (step S21), the microprocessor 221 determines that the SCSI bus reset has not been received from the host system 10 during the previous reset process, that is, the host system Judge that the last reset step has been executed within the monitoring time at 10,
Returning to step S3, the same processing as when the commands are not the same is performed.

The microprocessor 221 determines in step S2
4 refers to the retry management area 224, and executes the retry step executed when the address (error address) A set in the field 224d and the SCSI bus reset set in the field 224e are received. And the number N. Then, the microprocessor 221 executes the newly received normal access processing designated by the same command as the one previously retried from the start address to the address A-1 immediately before the detected address A (step S).
25). Next, the microprocessor 221 executes access to the address A in which the error has occurred last time, not through normal processing, but through retry processing, and from the retry step N indicated by the detected retry step number N (step S26). That is, the microprocessor 221
Indicates that an error has occurred in the access to the address A among the addresses designated by the command received from the host system 10 last time, a retry process is performed, and a SCSI bus reset is received during the execution of the retry step N in the retry process. When the same command is executed next, the access to the address A is executed by the retry processing from the retry step N.

For example, the previous command is a read command, and the address specified by the command is # 100 to # 1.
Assuming that a retry occurs at the time of reading from the address # 105 and a SCSI bus reset is received during the retry step # 15 of the retry steps # 1 to # 30 (of the retry step sequence), Next, when the same command is received from the host system 10, the command is executed as follows. First,
Access to addresses # 100 to # 104 is normally performed. Next, access to the address # 105 is executed from the retry step # 15.

If the access to the address # 105 is normally performed unlike the present embodiment, the probability of occurrence of a retry is high. Further, even if the retry processing is performed, there is little possibility of recovery from retry processing from retry step # 1 to retry step # 14. On the other hand, in the present embodiment, since the retry processing is performed from the retry step # 15 which is being executed last time, the probability of error recovery increases and the processing time is shortened.

After the above step S26, the above-described processing after step S11 is performed.

Next, on the HDD 20 side, the host system 10
The operation when a SCSI bus reset is received from the CPU will be described. First, the microprocessor 221 determines whether a retry process is being performed (step S31). if,
If the retry processing is being performed, the microprocessor 221 stops the retry processing, and sets the reset flag during retry in the field 224f of the retry management area 224 to O.
N (Step S32). As a result, a log indicating that a SCSI bus reset has been received from the host system 10 during the retry process, that is, a log indicating that a timeout error has been detected in the host system 10, has been collected in the field 224f.

The microprocessor 221 determines in step S
32, except for the retry management area 224,
The initialization process for the HDD 20 is performed in response to the SCSI bus reset from the host system 10 (step S33). As a result, the information of the retry management area 224, that is, the information and error address of the command that was the target of the retry processing when the SCSI bus reset was received, and the number of the retry step executed at that time are not reset. Will be left. On the other hand, if the host system 10
When a bus reset is received, the microprocessor 2
21 performs a normal initialization process for the HDD 20 in response to the SCSI bus reset (step S34). In this case, the information of the retry management area 224 is also initialized (cleared). The microprocessor 221 determines in step S
When the initialization processing of step S33 or S34 is performed, the state changes to the READY state, and a command from the host system 10 is waited.

Next, retry in the host system 10
That is, the host retry will be described. When transmitting the access command to the HDD 20, the host system 10 monitors a response to the command from the HDD 20 for a predetermined monitoring time. If an abnormal end is notified from the HDD 20 during the monitoring time,
Alternatively, if a timeout error occurs without returning a response from the HDD 20 even after the monitoring time has elapsed, the host system 10 transmits a SCSI bus reset to the HDD 20, and then executes a host retry described below.

First, the host system 10 sends the same command as the one that caused the abnormal termination or the timeout error to H
The data is transmitted to the DD 20 (step S41). Then, the host system 10 starts monitoring the response from the HDD 20 (step S42).

As described above, on the HDD 20 side, the reset flag during retry indicates that the reset flag during retry is ON, that is, the SCSI bus reset is received from the host system 10 during the retry process. In this state, when the same access command as the one in which the retry processing was performed last time is received from the host system 10, the specific error code is notified from the HDD 20 to the host system 10.

When a specific error code is notified from the HDD 20 in response to the transmission of the access command to the HDD 20 (step S43), the host system 10 sets the monitoring time to be extended by a predetermined time (step S4).
4).

The host system 10 monitors the response from the HDD 20 with the set monitoring time as an upper limit (steps S45 and S46). If there is no response from the HDD 20 even after the monitoring time has elapsed, the host system 10 determines a timeout error in the host retry (step S46).

When the host system 10 determines that a timeout error has occurred, it determines whether a timeout error has occurred consecutively M times (M is a predetermined integer) without notification of an error code (steps S47 and S4).
8). If a timeout error occurs consecutively M times without notification of an error code, the host system 1
A value of 0 determines that the HDD 20 is in a failure or locked state (step S49). In this case, the host system 10 notifies the user (system administrator) of the failure or the locked state of the HDD 20 via a display screen or the like.
Request exchange of DD20, etc.

On the other hand, if the number of consecutive occurrences of the timeout error is less than M, or if the error is a timeout error with the error code being notified even if the number of consecutive occurrences is M, the host system 10 It is determined that the timeout error is due to the monitoring time being shorter than the time required for the retry processing in the above, and is not the timeout error due to the failure or the locked state of the HDD 20. In this case, the host system 10
, A SCSI bus reset is transmitted again (step S50), and then the host retry is executed again.

On the other hand, if the HDD 20 returns a response to the access command from the host system 10 in the status phase before the timeout occurs, that is, within the monitoring time (step S45), the host system 1
It is determined whether 0 is a response indicating a normal end or a response indicating an abnormal end from the response content (step S5).
1). If the response is an abnormal end response, the host system 10 determines whether an abnormal end response has been returned L times (L is a predetermined integer) continuously (step S5).
2). If the number of consecutive abnormal termination responses is less than L, the host system 10 determines that the HDD 20 may terminate normally by retrying the host again.
, A SCSI bus reset is transmitted again (step S50), and then the host retry is executed again.
On the other hand, if an abnormal end response is returned L times consecutively, the host system 10 determines that further host retries are useless and performs error processing based on the information of the error location indicated by the abnormal end response. Perform (step S53).
Note that the error processing itself is not directly related to the present invention, and a description thereof will be omitted. When a normal end response is returned from the HDD 20, the host system 10 ends the host retry.

In the embodiment described above, when the command type, the start address, and the size of the received command are the same as those of the command for which the previous retry processing was performed, the received command is subjected to the previous retry processing. Although it is assumed that the command is the same as that described above, the present invention is not limited to this. For example, an access command requesting access to an address range including at least the address where an error has occurred last time may be the same command as the last time.

The present invention is not limited to the above-described embodiment, and can be variously modified in an implementation stage without departing from the gist of the invention. Further, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some components are deleted from all the components shown in the embodiment, the problem described in the column of the problem to be solved by the invention can be solved, and the effects described in the column of the effect of the invention can be solved. If you get
A configuration from which this configuration requirement is deleted can be extracted as an invention.

[0054]

As described above in detail, according to the present invention, when a reset request accompanying a timeout error detection in the host system is received during a retry process in the disk storage device, the event is recorded and the event is recorded. In this state, when the same access request as that which was retried earlier to the disk storage device is given from the host system, the disk storage device notifies the host system of a specific error code and notifies the host system. Since the monitoring time is extended, the occurrence of timeout error detection in the host system can be suppressed. Further, according to the present invention, the cause of the timeout error can be easily determined in the host system based on whether a specific error code is notified from the disk storage device to the host system in response to the reset request accompanying the detection of the timeout error.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a computer system according to an embodiment of the present invention.

FIG. 2 is a view showing an example of a data structure of a retry management area 224 in FIG. 1;

FIG. 3 is an exemplary flowchart showing a part of a processing procedure for receiving an access command from the HDD 20 according to the embodiment;

FIG. 4 is a view showing another part of the flowchart for describing the processing procedure when the access command is received.

FIG. 5 is a view showing the rest of the flowchart for explaining the processing procedure when the access command is received.

FIG. 6 is an exemplary flowchart illustrating a processing procedure when the HDD 20 receives a SCSI bus reset in the embodiment.

FIG. 7 is an exemplary flowchart for explaining the procedure of a host retry process of the host system 10 in the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Host system 20 ... HDD (magnetic disk device, disk storage device) 21 ... Drive part 22 ... Control part 221 ... Microprocessor 222 ... ROM 223 ... RAM 224 ... Retry management area (specific area)

Claims (7)

[Claims] 1. A disk storage device used as an external storage device of a host system by said host system, wherein said means executes an access request received from said host system, and execution of said access request results in an error. A means for executing a retry process including a predetermined retry step group; a storage means having a specific area secured; and a response from the disk storage device regarding execution of the access request sent by the host system is a set monitoring time. If a timeout error not returned within the host system is detected by the host system, and a reset request for initializing the disk storage device sent from the host system during execution of the retry process is received, The retry-in-reset information indicating that fact is stored in the storage device. Retry information management means for registering in the specific area of a stage, means for causing the access request execution means to execute an access request sent from the host system,
If the access request is the same request as the previous request and the reset information during retry is registered in the specific area of the storage unit, it indicates that the reset request was received during execution of the retry process, and Means for notifying the host system of a specific error code requesting the host system to extend the monitoring time and causing the access request execution means to execute the access request. . 2. The retry information management means registers, in the specific area, an access request and an address subjected to the retry processing by the access request execution means and retry step identification information indicating a retry step being executed. At the same time, each time the retry step is switched, the retry step identification information on the specific area is updated, and the access request execution means receives from the host system the same access request as the target of the retry processing. In this case, an access to an address which has been previously subjected to the retry processing registered in the specific area is executed by a retry processing from a retry step indicated by the retry step identification information on the specific area. The disk storage device according to claim 1, wherein 3. A means for performing initialization processing of the disk storage device in response to the reset request from the host system, wherein when the reset request is received during execution of the retry processing, at least the 2. The disk storage device according to claim 1, further comprising means for performing an initialization process excluding a specific area. 4. A computer system comprising a host system and a disk storage device used by the host system as an external storage device of the host system, wherein the disk storage device receives an access request received from the host system. Means for executing, when an error occurs during execution of the access request, means for executing a retry process including a predetermined retry step group; storage means for securing a specific area; execution of the retry process When receiving a reset request for initializing the disk storage device from the host system during retry information management means for registering retry-in-reset information indicating the fact in the specific area of the storage means, Execution of the access request sent from the system And means for executing by the step,
If the access request is the same request as the previous request, and the reset information during retry is registered in the specific area of the storage unit, a specific request indicating that the reset request was received during execution of the retry process Means for notifying an error code to the host system and causing the access request execution means to execute the access request. When the host system sends an access request to the disk storage device, the host system executes the access request. Monitoring means for detecting a timeout error by monitoring a response from the disk storage device with respect to a currently set monitoring time as an upper limit, wherein the monitoring is performed when the error code is notified from the host system. Computer having monitoring means for monitoring for extended time Stem. 5. The host device determines a failure of the disk storage device when the timeout error is continuously detected a predetermined number of times for the same access request in a state where the error code is not notified. 5. The computer system according to claim 4, further comprising: 6. The retry information management means registers, in the specific area, an access request and an address subjected to the retry processing by the access request execution means and retry step identification information indicating a retry step being executed. At the same time, each time the retry step is switched, the retry step identification information on the specific area is updated, and the access request execution means receives from the host system the same access request as the target of the retry processing. In this case, an access to an address which has been previously subjected to the retry processing registered in the specific area is executed by a retry processing from a retry step indicated by the retry step identification information on the specific area. The computer system according to claim 4, wherein 7. An error notification method at the time of a retry process in a computer system including a host system and a disk storage device used by the host system as an external storage device of the host system, wherein When an error occurs in execution of the access request given to the disk storage device, the disk storage device executes an internal retry process including a predetermined retry step group; and the host system accesses the disk storage device. Detecting a time-out error by monitoring the response from the disk storage device regarding the execution of the access request with the currently set monitoring time as an upper limit when the request is transmitted; and Detection or said data Upon host retry of sending the same access request from the host system to the disk storage device again in response to the abnormal termination notification from the disk storage device, the host system initializes the storage device to the disk storage device. Sending a reset request for executing the internal retry process of the disk storage device, and when the disk storage device receives the reset request, holding retry-in-progress reset information indicating the reset request, Initializing the inside of the disk storage device except for at least the reset information during retry in response to a reset request; and the disk storage device receives an access request from the host system in a state of holding the reset information during retry. If the disk storage device Notifying the host system of a specific error code; and extending the monitoring time by the host system according to the error code. .