JP2830592B2 - Route failure processing method for external storage device in information processing system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing system, more specifically, an OS running on a host processor.
Relates to a path failure processing method for managing the state of an external storage device connected to a host processor via a plurality of paths.

[0002]

2. Description of the Related Art In an information processing system, when an external storage device detects some kind of failure and abnormally terminates an input / output operation, it generates detailed information indicating the state of the failure and sends it to an O / O.
Report to S. Conventionally, it has been general to indicate in this detailed information whether the failure is a path failure or a failure in a common part independent of the path. The OS determines the information of the common unit failure or the path failure, and retry the abnormally terminated input / output operation as follows.

First, in the case of a common unit failure, the input / output operation is retried a specified number of times on the failure path. Next, in the case of a path failure, retries are performed a specified number of times on the failed path, and if this cannot be remedied, retries are performed a specified number of times on the alternate path. If there is no available alternative route, the same failure processing as the former is performed. If the remedy succeeds on the alternate route, the operation of the alternate route is normal, and multiple attempts have already failed on the failed route, and the failed route was used in subsequent I / O operations. In such a case, since the probability of occurrence of the same type of fault is high, the fault route is closed so as not to be used in subsequent input / output operations. The purpose of this path blockage is to stop using a path with a high probability of failure and avoid an increase in system overhead due to failure processing.

[0004]

By the way, although the line failure is intermittent, in a repetitive environment, as the frequency of occurrence of the failure increases, the number of times the failed line is retried increases accordingly, and the performance of the system is reduced. There was a problem of doing.
In addition, since the fault route is remedied by retrying, the conventional method does not block the fault route and continues to use the fault route in subsequent input / output operations.
SUMMARY OF THE INVENTION An object of the present invention is to solve the above-described problem, and to provide a path failure processing method for an external storage device that reduces the chance of using an unstable path and prevents a decrease in system performance due to useless retry processing. is there.

[0005]

In order to achieve the above object, an external storage device in an information processing system according to the present invention is provided.
Path failure processing method of, OS running on the host processor, an information processing system odor that performs state management of the external storage device connected by a plurality of paths to the host processor
Te, be counted a path failure cumulative number of times the external storage device
A path failure accumulation counter means that, when a path failure occurs after the cumulative number of path failure occurrences exceeds a threshold value, when the I / O operation abnormal end is reported to the OS, the path failure accumulation number counts the threshold value. and a control means for reporting a call beyond,
If the OS receives the report exists the alternative path, on this alternative path, if the alternative path exists so as to retry the input and output operations in the current failure path, over the alternative path retried
If the I / O operation succeeds, the initial failure
It is configured to block the road .

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in more detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of an information processing system employing a line fault processing method for an external storage device according to the present invention. The information processing system of this embodiment is physically composed of a host processor 10, an input / output channel group 20, and an external storage subsystem 30. OS
A program 12 runs on the host processor 10 and manages the entire information processing system such as the job management unit 13 and the resource management unit 14. A resource status table 17 is provided in a part of the resource management unit 14 of the OS 12, and the OS 12
The status of each resource in the information processing system that can use the information 12 is controlled.

The input / output channel group 20 is a connection unit for connecting the host processor 10 and the external storage subsystem 30, and is used as a part of an access path from the OS 12 to the external storage subsystem 30. The external storage subsystem 30 includes a plurality of external storage devices 33 and a plurality of external storage control devices 31 for exchanging data between the external storage device 33 and the input / output channel 20. The external storage device 33 is connected to the OS 12 by a plurality of connection paths due to its importance, and is configured to be able to continue input / output operations on an alternate path even if a certain path becomes unavailable due to a failure. The external storage control device 31 is arranged on each path.

An input / output request issued by the application program 11 is executed under the control of an input / output control unit 16 in the OS 12 via a file management unit 15 in the OS 12. The input / output control unit 16 stores the resource status table 1
7, an available path is selected from a plurality of paths to the external storage device 33 to be accessed, and an input / output instruction is issued to the input / output channel 20 on the selected path. The issued I / O instruction is executed by the combined operation of the I / O channel 20 and the external storage controller 31,
Detailed information indicating the presence / absence of a failure and the state of the failure is returned to the input / output control unit 16 as the end result of the input / output command. A failure in the input / output operation occurs in the input / output channel 20, the external storage control device 31, or the external storage device 33,
Among them, detailed information is generated in the external storage control device 31 with respect to a failure that has occurred in the external storage control device 31 and the external storage device 33.

The external storage controller 31 categorizes the generated fault into a common part fault and a path fault in the external storage subsystem 30, and if the fault is a path fault, a path fault accumulation counter installed in the external storage controller 31. 32 is incremented. In the detailed information generated by the external storage control device 31, there is provided a portion for presenting a type of a fault processing request to the input / output control unit 16, and “path faults whose route fault cumulative counter is less than a threshold value”, “path faults”. Either "path failure whose cumulative counter is equal to or greater than the threshold value" or "not a path failure" is stored. Upon receiving this report, the input / output control unit 16 selects a failure process according to the failure process request type and the resource status table 17.

FIG. 2 is a flowchart for explaining the operation procedure of the input / output control unit. First, it is determined whether or not there is an alternate route (step 1). If the route is unavailable, retry is performed on the failed route regardless of the type of the failure processing request (step 7). On the other hand, when the alternative route is available, the type of the failure processing request is determined (step 2), and the following failure processing is executed. If "not a path failure", retry is performed on the failed path (step 7). In the case of "path failure", it is determined whether or not "path failure with a path failure cumulative counter smaller than a threshold value" (step 3). If it is "route fault whose route fault cumulative counter is less than the threshold value", after retrying the fault route a specified number of times (step 4), it is determined whether or not rescue is possible (step 5). If the remedy can be made, the retry is completed successfully. If it cannot be remedied, the process proceeds to the retry on the alternate route (step 9). In the case of "route fault whose route fault cumulative counter is equal to or greater than the threshold value", the process immediately proceeds to step 9 to retry on an alternate route.

When the fault is remedied by either the retry of the fault route or the retry of the alternate route, the input / output control unit 16 notifies the application program 11 that is the input / output requester that the input / output operation has been normally completed. I do. Also, the input / output control unit 16 retrys the alternate route in the course of the above-mentioned fault processing and relieves on the alternate route.
Makes the initial failure route unavailable (step 1
1). On the other hand, the external storage control device 31 executes each input / output instruction, and determines whether each input / output instruction is an initial instruction, an instruction at the time of retrying a failed path, or an instruction at the time of retrying an alternate path. Is not determined. Therefore, the failure processing request type is just a request, and the failure processing type is determined by the input / output control unit 16.

[0012]

As described above, according to the present invention,
Even if the path failure is intermittent but occurs repeatedly, the system is configured so that alternate path retry can be promoted and the failure path can be closed early, so that system performance due to useless retry processing is improved. In addition, there is an effect that the occurrence of a more serious failure which is likely to occur by continuously using the unstable path can be prevented.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a route failure processing method for an external storage device according to the present invention.

FIG. 2 is a flowchart illustrating a failure process of an input / output control unit in FIG. 1;

[Explanation of symbols]

 REFERENCE SIGNS LIST 10 host processor 11 application program 12 OS (operation system) 13 job management unit 14 resource management unit 15 file management unit 16 input / output control unit 17 resource status table 20 input / output channel 30 external storage Subsystem 31 ... Storage control device 32 ... Path failure accumulation counter 33 ... External storage device

Claims (1)

(57) [Claims] 1. An OS running on a host processor,
In an information processing system for managing a state of an external storage device connected to a host processor via a plurality of paths, a process for counting the number of times of path failure occurrence accumulation in the external storage device is performed.
A path failure accumulation counter means for, when a path failure occurs after the cumulative number of path failures exceeds a threshold value, reporting the I / O operation abnormal end to the OS, the cumulative number of path failure occurrences exceeding the threshold value; reporting and octopus
A control means for, if the report received OS is it exists the alternative path, on this alternative path, so as to retry the input and output operations in the current failure path if the alternative path exists, the Retryed on alternate route and the I / O operation was successful
A path failure processing method for an external storage device in an information processing system, wherein an initial failure path is closed in a case .