JPS634212B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to means for reporting failure details to a higher-level device when a failure occurs in a control device having an interface with a higher-level device.
By logically connecting two microprocessors, even in the event of a failure in which the control device cannot report the details of the failure to the higher-level device, error reporting allows other control devices to report the exact details of the failure to the higher-level device. Regarding the method.

In conventional control devices, when a failure occurs in the control device, the contents of the failure are edited in order to report the details of the failure to the higher-level device during processing or at the end of processing, and the edited failure details are sent to the higher-level device. Trying to report. However, if the fault occurs in a central part that causes a significant loss of function of the control device, or if it is an interface part with a host device, it becomes impossible to report the details of the fault to the host device. If such a failure occurs, a retry from the higher-level device may not be performed correctly, and a large amount of time will be required for recovery work.

In addition, in systems that are designed to be redundant in order to reduce the impact of control device failures and increase system reliability, when a failure occurs in one control device, the details of the failure are reported to the higher-level device. In this case, the OS (operating system) executes a retry via an alternate path that has no failure, and blocks the failed path.

However, even in a system designed for duplication,
If the failure details of the control device cannot be reported to the higher-level device, the OS will not perform appropriate retries, resulting in a delay in recovery.

An object of the present invention is to use a microprocessor to establish a path between each control device, and analyze and edit the details of the error using the microprocessor. Even for failures that cannot be reported, the correct details of the failure can be reported through other control devices.
The primary purpose is to encourage the OS to retry appropriately and to minimize the impact on the system.

Furthermore, the second purpose is to restore the failure control device in a short time based on the failure details reported to the host device using the above method.

The present invention logically connects each control device via a microprocessor, and when a failure occurs in a control device and the control device determines that the failure cannot be reported to a higher-level device, or when the control device functions. In the event of a failure that significantly impairs the control device, the microprocessor receives a signal from the control device to receive the details of the fault in the faulty control device, or analyzes the fault and edits the fault details. A normally operating control device that generates an interrupt signal that retains the fault details to the device and receives a response signal from the control device will send an interrupt signal to the microprocessor when it confirms the interrupt signal at the end of processing. Then, in order to receive the details of the fault in the faulty device, a transfer command is issued to the microprocessor, and the fault details are received. The control device that receives the fault details reports the fault details to the higher-level device along with a code indicating the control device in which the fault has occurred.

In this way, even failures that could not be reported in the past can be reported to the host device, thereby eliminating the conventional drawbacks and significantly improving the reliability of the system.

The present invention will be described in detail below using a magnetic disk control device as an example. At present, storage devices such as magnetic disks are not only used as mere input/output devices, but are also indispensable as auxiliary storage devices for systems.

For this reason, high reliability is required as a file subsystem.

Use of the present invention in such a file control device is extremely effective in increasing reliability.

An embodiment of the present invention will be described. FIG. 1 shows part of a conventional file subsystem, in which 1 is a CPU, 2 is a channel 3 is a control device 4
The cross-call adapter 5 is a disk device, and the path for accessing the disk device from the CPU is completely duplicated.

FIG. 2 shows an embodiment of the file subsystem according to the present invention, in which 1 to 5 are the same as in FIG. The difference is in how they are connected.

FIG. 3 shows the interface and configuration between the control devices 3, 33 and the microprocessor 6, 10 and 11 are adapters for connecting the control device and the microprocessor, and 12 is an adapter for connecting the control device and the microprocessor.
A common bus 13 for the microprocessor is a microprocessor, and 14 is a memory for controlling the microprocessor.

The interface between the control device and the microprocessor consists of information lines and control lines.

In normal operation, CPU 1 selects and issues commands to channel 2, controller 3, and adapter 4 to access disk 5. However, when a failure occurs in the control device 3, the control device 3
analyzes the content of the fault, edits the fault content, and
Report to CPU1. CPU1 analyzes the failure details and
Channel 22 to perform a replacement pass retry;
Control device 33, select adapter 4, and select disk 5
access. By duplicating the paths to the disks in this way, the reliability of the system is increased.

However, if the control device 3 is unable to report a failure occurring in the control device 3 to the CPU 1, an appropriate alternate path retry is not performed. Furthermore, since the cause of the failure in the control device 3 is not known, a large amount of time is required for recovery.

Therefore, a microprocessor 6 is connected between the control devices 3 and 33 as shown in FIG.

If a failure occurs in the control device 3 while the CPU 1 is accessing the disk 5 through the control device 3, the contents of the failure are analyzed and edited during or at the end of the process depending on the nature of the failure. Attempt to report to device. At this time, if the fault in the control device 3 significantly impairs the function, an interrupt signal is generated to the microprocessor 6 when the fault occurs, and the microprocessor 6 is requested to analyze and edit the fault.

That is, when the control device 3 detects an error during data transfer, etc., it performs a retry, but if the error is still not resolved, it notifies the host device of the details of the failure. However, in the case of a hardware failure in the interface section with the host device, for example, notification to the host device will not work (for example, even if a signal is sent to the host device, a response to it cannot be received). In such a case, in the present invention, the control device 3 generates an interrupt signal to the microprocessor 6. In addition, in the control device 3,
While the interface section with the host device is relatively complex using high-speed elements, the interface section with the microprocessor 6 may be medium or low speed and should be relatively simple. Can be done. Therefore, the failure rate of the interface section with the microprocessor 6 can be lowered compared to the interface section with the host device.

Furthermore, it is considered that the probability of both failures occurring at the same time is sufficiently small.

Now, the microprocessor 6 that received the interrupt
causes the control device 3 to analyze and edit the details of the failure, and transfers the results to the microprocessor 6.

When the microprocessor 6 retains the details of the fault in the control device 3, the microprocessor 6 generates an interrupt signal to the control device 33. Control device 3
3 confirms the interrupt signal from the microprocessor 6, it issues a command to the microprocessor 6 to transfer the fault details, and the control device 33
Receive the details of the problem. The control device 33 displays the details of the fault in the control device 3 indicating that it is the fault of the control device 3.
It attaches an ID and reports it to the higher-level device, prompting an appropriate retry.

In this way, having a microprocessor that analyzes failures in control devices and logically connecting control devices through that microprocessor is extremely important in reporting accurate failure details to higher-level equipment. This is an effective method, and has great effects in encouraging appropriate retries by the OS, shortening recovery time for failure control devices, and increasing system reliability.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a configuration example of a conventional file system, FIG. 2 is a configuration block diagram of a file system according to an embodiment of the present invention, and FIG. 3 is a partially detailed block diagram of FIG. In the figure, 1 is the CPU, 2, 22 are the channels, 3,
33 is a disk control device, 4 is a cross-call adapter, 5 is a disk device, and 6 is a microprocessor.

Claims (1)

[Claims] 1. In a system in which a common lower-level device is controlled by multiple control devices having an interface with a higher-level device, a microprocessor that is independent of the multiple control devices, and an interface between the microprocessor and the multiple control devices If a fault occurs in the control device, and the faulty control device itself cannot report the details of the fault to the higher-level device, a microprocessor independent of the control device handles the fault through an interface with the faulty control device. The content of the fault is received or analyzed, and the fault content is sent to the normal control equipment through the interface between the microprocessor that is independent of the control equipment and the normally operating control equipment. An error reporting method characterized by a control device reporting to a higher-level device.