JPH03265960A - Channel fault processing system - Google Patents

Abstract

PURPOSE:To prevent the transformation of data by separating a faulty channel temporarily from an input/output processor, applying a function test to the channel from a service processor, registering the error information to the service processor if the abnormality is through, and hereafter keeping the channel in an unavailable state. CONSTITUTION:An input/output processing part 101 instructs a channel connection control part 103 to separate a faulty channel from an input/output processor. Thus the part 103 turns off the logical connection at a channel connection part 104. Both a channel error log register part 106 and a channel function test (T&D) monitor 107 check the executing results and fetch the error information from a T&D executing part 108 as long as the abnormality is through. The error information is registered to the part 106 through a log transmission line 113. Then an instruction is given to the part 101 through a CH state setting instruction line 112 to keep the separated faulty channel in an unavailable state. Thus it is possible to prevent the occurrence of the transformation of data.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a channel failure handling method, particularly to a channel failure handling method in an information processing apparatus having an input/output processing device having a plurality of channels and a service processor.

[Conventional technology]

In conventional channel failure processing, when a failure occurs in a channel, the input/output processing unit logically disconnects the failed channel and reports the details of the failure to the service processor. The number of failures is counted, and if the number of failure occurrences is less than n times in a certain period of time, the error state of the failed channel is released and then the channel is incorporated into the input/output processing device again.

Further, when a failure occurs on the same channel n times in a certain period of time, it is determined that a fixed failure has occurred, and the channel is left disconnected from the input/output processing device.

[Problem to be solved by the invention]

The above-described conventional channel failure processing method treats failures occurring within n times on the same channel as intermittent failures and unconditionally incorporates them into the input/output processing device, which has the following drawbacks.

First, even a channel in which a fixed failure has occurred continues to be used until the failure occurs n times, so there is a risk that garbled data may occur.

Second, unless a fault occurs once within a certain period of time, it is not determined to be a fixed fault, so if a fixed fault occurs in a channel with a low input/output load, it is treated as an intermittent fault and the channel is not isolated. Sometimes.

Third, when a channel failure occurs, a log is collected to confirm the location of the failure, but the cause of the failure may not be grasped from the log information alone. In such a case, in order to trace the cause of the failure, one possible method is to run an operational test and isolate the pattern in which the failure occurs, but the system environment at the time of the failure is special, and unless it is that environment, the failure will occur. If you cannot see the problem, it may be difficult to reproduce the same problem later.

[Means to solve the problem]

The method of the present invention provides an information processing device that includes an input/output processing device to which a plurality of channels are connected and a service processor, and the service processor has a function of instructing the input/output processing device to set the state of the channel. In the channel fault handling method, when a fault occurs in the channel, means for logically separating the channel from the input/output processing device and reporting the details of the fault to the service processor; means for performing a functional test of the channel when a fault report is received; and, when the functional test is successfully completed, releasing the error state of the faulty channel and logically connecting it to the input/output processing device; , when an error is detected in the functional test, registering error information in the service processor;
and means for keeping the channel disconnected.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

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

The input/output processing unit 101 controls the operation of the input/output processing device,
The channel failure detection unit 102 detects the occurrence of a channel failure and sends a channel error log to the service processor. The channel connection unit control unit 103 connects/disconnects between the input/output processing device and the channel 105.

The channel connection unit 104 connects the input/output processing device to the channel 10.
Channel 105 provides data transfer between peripheral devices and input/output processing devices.

The log registration unit 106 registers a channel error log and error information during execution of a channel function test (hereinafter referred to as T&D). The T&D monitor 107 activates T&D for a channel when a channel failure occurs. The T&D execution unit 108 executes T&D of the instructed channel.

The failure detection unit 109° 111 notifies the input/output processing unit 101 and the T&D monitor 107 of the channel number where the failure has occurred.

Log sending 110 sends the error log of channel 105 from channel failure detection section 102 to log registration section 106 .

The CH state setting instruction 112 sends a channel state setting instruction from the T&D monitor 107 to the input/output processing section 101.

Next, the operation of the channel failure handling method of the present invention will be explained.

When a failure occurs in channel 105, the channel number is notified to channel failure detection section 102. Upon receiving the failure notification, the channel failure detection unit 102 detects the failure in the channel 105.
After reading the error log of the error log, the fault channel number is sent to the input/output processing unit 101 . T&D
The monitor 107 is notified. Then, the error log of the channel is registered in the log registration unit 106 through the log sending 110.

The input/output processing unit 101 that has received the channel failure notification instructs the channel connection control unit 103 to disconnect the failed channel from the input/output processing device, and the channel connection control unit 103
turns off the logical connection in the channel connection unit 104
Make it. Further, the T&D monitor 107 is the T&D execution unit 10
8 to activate T&D for the faulty channel.

When the T&D is completed, the T&D execution unit 108 reports the T&D execution results to the T&D monitor 107.

The T&D monitor 107 checks the execution result, and if the execution has ended normally, it instructs the input/output processing section 101 to put the faulty channel into a built-in usable state through a CH state setting instruction 112. If the termination is abnormal, the T&D execution unit 10
After fetching error information from 8 and registering it in the log registration unit 106 through log sending 113, CH status setting instruction 112
This command instructs the input/output processing unit 101 to keep the faulty channel disconnected and make it unusable.

When the input/output processing unit 101 receives a channel usable state setting instruction from the T&D monitor 107, it releases the error state of the faulty channel and issues a channel connection instruction to the channel connection control unit 103, thereby making the faulty channel available to the input/output processing device again. 101.

Further, upon receiving an instruction to set a channel unusable state from the T&D monitor 107, the input/output processing unit 101 sets the faulty channel to the 5TOP state for channel state management, and from now on,
It responds to input/output requests issued to that channel with a report of channel 5 TOP.

FIG. 2 shows the operations described above in relation to the passage of time.

〔Effect of the invention〕

As explained above, in the present invention, a faulty channel is once disconnected from an input/output processing device, a service processor performs a functional test on the channel, and if the result is normal, the channel is reinserted. The system is built into the output processing device, and if an abnormal termination occurs, error information is registered in the service processor, and the channel is rendered unusable in the future, resulting in the following effects.

First, the results of tests that cover all functions of the channel determine whether the fault that occurs is fixed or intermittent, so there is no need to worry about the method of unconditionally incorporating the channel until the fault occurs n times. Data corruption can be prevented.

Second, even if a fault occurs in a channel with a low input/output load, it is possible to quickly determine whether the fault is fixed or intermittent.

Thirdly, in an unmanned system, even if a failure occurs due to a special system environment, the system can immediately and automatically
If &D is executed and it terminates abnormally, the error information is registered in the service processor, so when a long time has passed since the failure occurred, you can perform a reproduction test in a completely different system environment than when the failure occurred. , can solve problems that are almost impossible to reproduce.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, and FIG. 2 is a diagram showing the operation of the processing method of the present invention in relation to the passage of time. 101... Input/output processing unit, 102...
Channel failure detection unit, 103...Channel connection control unit, 104...Channel connection unit, 105...
...Channel, 106...Log registration section,
107...T&D monitor, 108...
T&D execution unit, 109... Failure detection, 110.
...Log sending, 111...Failure detection, 1
12...CH status setting instruction.

Claims (1)

[Scope of Claims] An information processing device comprising an input/output processing device to which a plurality of channels are connected and a service processor, wherein the service processor has a function of instructing the input/output processing device to set the state of the channel. In the channel fault handling method, when a fault occurs in the channel, means for logically separating the channel from the input/output processing device and reporting the details of the fault to the service processor; means for performing a functional test of the channel when the report is received; and when the functional test is successfully completed, releasing the error state of the faulty channel and logically connecting it to the input/output processing device; When an error is detected in the functional test, registering error information in the service processor;
means for keeping the channel disconnected.