JPH0690693B2 - Channel failure recovery controller - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention is used for fault handling in a computer system.
In particular, it relates to channel failure recovery control due to failure of the input / output control device and peripheral control device.

〔Overview〕

In the I / O controller and the system connected to it by channels for I / O processing to peripheral devices, a channel status storage table is provided to store the failure status for each channel. When a fixed failure is stored in this table, channel initialization is performed. By prohibiting execution, the delay in failure processing time when a failure occurs is prevented.

[Conventional technology]

A failure that conventionally includes an input / output control device and a peripheral control device connected to the input / output control device by a channel, and detects a fixed failure or an intermittent failure in the input / output control device or the peripheral control device. There is known a channel failure recovery control device provided with a detection means and a channel initialization means for executing the initialization of the channel by sending an intermittent failure to the detection output of the failure detection means. In this device, when the failure detection means reports a failure of an intermittent failure, the channel initialization instruction is executed, whereby the channel initialization processing is activated and a temporary failure such as an intermittent failure is resolved. There is.

[Problems to be solved by the invention]

In the above-described conventional technique, the channel initialization instruction is executed even when the fixed failure is reported immediately after the intermittent failure is reported. That is, even if a fixed fault is detected in the hardware, the channel initialization process is activated by this, and it is necessary to wait for a certain period of time to detect the initialization failure. There was a defect such as the operation of being paused.

An object of the present invention is to improve this and to provide a failure recovery control device that does not execute unnecessary initialization instructions.

[Means for solving problems]

The present invention provides a channel status storage table that stores the detection output of the failure detection means for each channel by classifying it into a fixed failure or an intermittent failure, and the channel initialization means refers to the channel status storage table after activation, When a fixed failure is stored for a channel, it includes means for notifying the failure recovery means on the host immediately of the failure of initialization of the channel, and the failure recovery means immediately receives the notification of the failure of initialization of the channel. It is characterized in that it includes means for separating.

[Work]

When a fixed failure continues after the intermittent failure of the input / output control device or the peripheral control device, the fixed failure state is set in the channel status storage table by the failure detection means. As a result, even if a channel initialization instruction is generated, it is referred to before activation, so that the operation is immediately prohibited and abnormally terminated, and unnecessary channel initialization processing can be suppressed.

〔Example〕

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

FIG. 1 is a block diagram showing the overall configuration of the present invention. This device includes an input / output control device 1 that performs input / output processing with respect to the peripheral device 3, a peripheral control device 2 connected to the input / output control device 1 by a channel, and a peripheral device 3 connected under the peripheral control device 2. A fault detection means 4 for detecting a fault in the input / output control device 1 or the peripheral control device 2, a fault reporting means 5 for reporting a fault to the fault recovery means 7, and a detection output of the fault detection means 4. A channel status storage table 6 that stores each channel separately for fixed failure or intermittent failure, failure recovery means 7 that performs appropriate recovery processing according to the type of failure, and channel initialization means 8 that executes channel initialization. , A channel 9 connecting the input / output control device 1 and the peripheral control device 2, and a diagnostic means 10 for diagnosing the input / output control device 1 and the peripheral control device 2.

FIG. 2 is a control flowchart of the main part of the present invention.

When a failure occurs in the input / output control device 1 or the peripheral control device 2, the failure detection means 4 sets in the channel status storage table 6 whether the failure is a fixed failure or an intermittent failure, and reports the failure to the failure reporting means 5. Instruct them to do so. The fault reporting means 5 refers to the channel status storage table 6, determines whether to report a fixed fault or an intermittent fault, and reports the fault to the fault recovery means 7.

The failure recovery means 7 checks the status of the reported failure,
If a fixed fault is reported, recovery processing is not performed for the channel for which the fault has been reported, and the peripheral control device 2 connected to the channel and the peripheral device 3 connected thereunder are disconnected from the system. To do. Then, after the device is repaired, the diagnostic means 10 diagnoses the I / O controller 1 and the peripheral controller 2, and if it is confirmed that the failure is repaired, the fixed failure status of the channel status storage table 6 is determined. Reset.

If the reported failure is an intermittent failure, the failure recovery means 7 executes a channel initialization command to the channel initialization means 8 to initialize the channel and attempts recovery from the failure.

The channel initialization means 8 refers to the channel status storage table 6 to check the status of the corresponding channel before executing the initialization of the channel. At this time, if the fixed failure state is set in the channel status storage table 6, the channel initialization means 8 immediately abnormally terminates the initialization command. If the initialization command is abnormally terminated, the failure recovery means 7
As in the case where the fixed failure is reported, the peripheral control device 2 connected to the channel and the peripheral device 3 connected thereunder are disconnected from the system. If the intermittent failure state is set in the channel status storage table 6, the channel initialization means 8 requests the peripheral controller 2 for initialization and normally terminates the initialization command.
When the initialization command ends normally, the failure recovery means 7 waits for an initialization completion event that is reported after the peripheral controller 2 completes initialization.

When the peripheral control device 2 receives the initialization request from the channel initialization means 8, it resets the fault state, and when the initialization of the registers is successful, it notifies the channel initialization means 8 that the initialization was successful. . As a result, the channel initialization means 8 reports the initialization completion event to the failure recovery means 7.

If the initialization fails, the peripheral controller 2 cannot notify the initialization failure to the channel initialization means 8. Therefore, the failure recovery means 7 waits for the initialization success event for a fixed time after the initialization command is normally completed. If no initialization success event is reported within this period, it is regarded as a fixed failure, and the peripheral control device 2 connected to the relevant channel and the peripheral device 3 connected thereunder are disconnected from the system. To do.

When the failure recovery means 7 receives the initialization completion event from the channel initialization means 8, the failure recovery means 7 determines that the failure has been recovered, the peripheral control device 2 connected to the channel, and the peripheral device 3 subordinate to the peripheral control device 2. Reuse and complete the disaster recovery process.

The failure detected by the failure detection means 4 is initially an intermittent failure, and the failure reporting means 5 reports an intermittent failure to the failure recovery means 7, and the failure recovery means 7 issues an initialization command to the channel initialization means 8. If the fault detecting means 4 detects a fixed fault before executing, the fault detecting means 4 immediately sets the state of the fixed fault in the channel status storage table 6 so that the channel initializing means 8 can set the channel status. A fixed failure is detected when the storage table 6 is referred to, and the initialization command is immediately terminated abnormally. The failure recovery means 7 can immediately disconnect the failed channel, the peripheral control device 2 connected to the channel and the peripheral device 3 under the peripheral control device 2 without waiting for the initialization success event.

〔The invention's effect〕

As described above, according to the present invention, when the fixed failure continues after the intermittent failure of the input / output control device or the peripheral control device, the failure recovery process automatically performs the recovery of the first intermittent failure. It is possible to detect a fixed fault and prevent a delay in fault processing time due to unnecessary initialization processing, thereby improving the operating rate of the system.

The present invention is extremely effective when used for channel failure recovery control of a computer system.

[Brief description of drawings]

FIG. 1 is a block diagram showing the overall configuration of the present invention. FIG. 2 is a control flowchart of the main part of the present invention. 1 ... Input / output control device, 2 ... Peripheral control device, 3 ... Peripheral device, 4 ... Fault detection means, 5 ... Fault reporting means, 6
...... Channel status, 7 ... Failure recovery means, 8 ... Channel initialization means, 9 ... Channel connecting input / output control device and peripheral control device, 10 ... Diagnostic means.

Claims (1)

[Claims] 1. An input / output control device and a peripheral control device connected to the input / output control device by a channel, wherein a fixed fault or an intermittent fault occurs in the input / output control device or the peripheral control device. Failure detection means for detecting, failure recovery means for receiving the detection output of the failure detection means, and channel initialization means for receiving a channel initialization request from the failure recovery means and returning an initialization result to the failure recovery means. In the provided channel failure recovery control device, a channel status storage table is provided, in which the detection output of the failure detection means is divided into fixed failures and intermittent failures and stored for each channel, and the channel initialization means is provided for If a fixed fault is stored for the channel by referring to the status storage table, initialization of the channel fails. Immediately comprises means for notifying to said fault recovery means, the above failure recovery means channel failure recovery control unit which comprises a means for disconnecting the immediate channel receives the notice of initialization unsuccessful.