JPH0476136B2 - - Google Patents

Description

[Detailed Description of the Invention] [Summary] Regarding a device for notifying a host of an abnormal state of a subsystem including redundant equipment, for example, the present invention reduces the processing burden caused by unnecessarily frequent abnormality notifications. The purpose is to avoid this by adding only a few elements, including an abnormality detection means for detecting an abnormal state in the subsystem, a timekeeping means, and a timekeeping means for the above-mentioned timekeeping means by detecting the abnormality state by the abnormality detection means. a time measurement starting means for starting the time measurement; a time measurement interruption means for interrupting the time measurement by the counter during the period of access from the host; a time excess detection means for detecting that the time value of the counter has reached a predetermined value;
When accessed by the host, if the abnormality detection means detects an abnormality and the time exceedance detection means detects that the time value of the timer exceeds the predetermined value, the host and a notification means for notifying detailed information of the abnormal state to the
Configured an abnormality notification device in the subsystem.

[Industrial application field]

The present invention relates to a device for notifying a host of an abnormal state of a subsystem including, for example, duplicated equipment.

[Conventional technology and its problems]

If an abnormal state within a subsystem can be detected by the host during access, there is no need to provide a means for the subsystem to report the abnormal state to the host; If there is a failure on one side of the device being used, this subsystem continues to operate normally, so it is impossible to identify the abnormality from the processing details. Since abnormalities are not detected, it is necessary for the subsystem to report to the host at a frequency that does not allow the subsystem to overlook the host system or the operator, and in a manner that does not significantly affect access to and from the host. be.

Conventional methods for notifying abnormal conditions in subsystems include providing display means such as failure indicators and warning buzzers on the subsystem itself without reporting to the host. If the device is installed at a remote location, recognition cannot be performed reliably.

In addition, there is a system that continuously displays a failure on the host, and when a maintenance person or operator confirms it and operates a switch, the failure display is stopped and possible processing is continued. However, in this case, not only the process is temporarily stopped but also manual intervention is required, and there is a possibility that the occurrence of the abnormal state may be forgotten.

Another method is to notify the host of an abnormality every time the subsystem is accessed by the host, but this abnormality report places unnecessary load on the host, especially for devices that are accessed frequently. There is a problem that the processing efficiency of the entire system decreases. For example, if there is an abnormality in one of the duplexed devices, the problem will not occur as long as the other device is operating normally, so it is necessary to notify the host of the abnormality in one device every time the host accesses it. It is clear that there is no such thing.

Furthermore, Japanese Patent Application Laid-Open No. 53-142143 states that when an abnormality is detected in a subsystem while the subsystem is not in use and is free, the host is immediately notified of the abnormality by means such as an interrupt. However, it is not desirable to perform such interrupt processing because it complicates the processing.

An object of the present invention is to obtain an abnormality notification device in a subsystem that can avoid processing burden due to unnecessary frequent abnormality notifications by simply adding a few elements.

[Means for solving problems]

an abnormality detection means for detecting an abnormal state within a subsystem; a timekeeping means; a timekeeping start means for causing the timekeeping means to start timing upon detection of an abnormality by the abnormality detection means; and a period of time during which the host is accessing. , a notice interrupting means for interrupting time measurement by the counter, and a time excess detection means for detecting that the time measurement value of the counter has reached a predetermined value;
When accessed by the host, if the abnormality detection means detects an abnormality and the time exceedance detection means detects that the time value of the timer exceeds the predetermined value, the host and a notification means for notifying detailed information of the abnormal state to the
Configured an abnormality notification device in the subsystem.

[Effect]

Detailed information on the abnormality is displayed on the host when accessed by the host after the cumulative value, excluding the time during which the host is accessing the abnormality after it is detected by the above abnormality detection means, reaches a predetermined value. As a result, devices that are not directly accessed by the host or devices that are accessed infrequently are reported by a relatively small number of accesses, and devices that are frequently accessed are reported by a relatively high number of accesses. Depending on the number of times, the abnormality will be reported for the first time.

This makes it possible to significantly reduce the unnecessary load on the host due to abnormality reports from devices that are accessed frequently, especially compared to the conventional example in which an abnormality notification is issued for each access from the host.

In addition, since this error report is performed when there is an access from the host, there is no need for the host to access the subsystem just to report the error, which greatly reduces the impact that the error report has on host processing. A special effect can be obtained in that it can be made smaller.

〔Example〕

The figure shows an embodiment of a subsystem according to the present invention, and only important parts related to the present invention are shown.

The microprocessor 1 has the function of periodically monitoring abnormalities within the subsystem and detecting and storing detailed information.
For example, in the case of a semiconductor disk device equipped with a semiconductor memory and a micro magnetic disk device that loads or saves data to and from the semiconductor memory when the power is turned on and off, information stored in the entire area of the semiconductor memory is read out and stored in the ECC code. Abnormality detection is carried out by correcting a 1-bit error and writing again, or abnormality detection of the micro magnetic disk device by attempting to write to and read from a specific area of the micro magnetic disk, etc., is periodically and repeatedly performed. As a result of this monitoring, if an abnormality is detected, the first flip-flop 2
is set, and if no abnormality is detected, this flip-flop 2 is reset.

The idle counter 3 measures time by counting the output clock of the clock generator 4.
When a signal is applied to the start terminal St., counting starts, when a signal is ignited to the hold terminal H, counting is stopped while holding the counted value, and when a signal is applied to the reset terminal R, it is reset. be done.
When the count value of the idle counter 3 overflows, its output sets the second flip-flop 5.

Now, if the microprocessor 1 detects an abnormality as a result of monitoring, the first flip-flop 2 is set, but if the count value of the idle counter 3 does not overflow, the second flip-flop is set. 5 is not set, the 3-input AND circuit 6 does not produce an output.

In this state, the idle counter 3 continues to count the clocks, and when the count value overflows, the overflow output sets the second flip-flop 5. However, if an access request is made from the host to this subsystem during the period until this overflow output occurs, the access acceptance signal output from the microprocessor 1 to the conductor 10 will be held at the idle counter 3 during that period. Since it is applied to the terminal H, the count value does not increase during this access request period, and the count value of the counter 3 corresponds to a period in which there is no access request from the host. This counter 3
In the case of the above-mentioned semiconductor disk device, it may take about 15 minutes until overflow occurs.

In this way, when the first and second flip-flops 2 and 5 are both set, that is, when there is an abnormality in this subsystem and the cumulative amount of time that the subsystem is not accessed reaches a predetermined abnormal value. , when there is an access from the host via this subsystem, the microprocessor detects this access and outputs an access acceptance signal on the conductor 10, so the 3-input AND circuit 6 generates an output and switches the multiplexer 7, Detailed information about the abnormality from the microprocessor 1 is sent to the host in place of the information requested by the host. Note that after the microprocessor 1 finishes sending out this detailed information, it sends a reset signal onto the conductor 11, resets the idle counter 3 and the second flip-flop 5, and then applies a start signal onto the conductor 12. Start lever counter 3 again.

〔Effect of the invention〕

According to the present invention, if there are few accesses from the host, the report is made more frequently than the number of accesses, and if there are many accesses from the host, the report is made less frequently than the number of accesses.
An efficient abnormal condition reporting method can be obtained even though it has little effect on access between the host and the subsystem.

[Brief explanation of the drawing]

The figure shows an embodiment of the present invention, in which 1 is a microprocessor that has the function of periodically diagnosing and storing devices and functions within the subsystem, and 3 is an idler that cumulatively counts the time that is not accessed by the host. It is a counter.

Claims (1)

[Scope of Claims] 1. Abnormality detecting means for detecting an abnormal state in a subsystem; Time measuring means; Timing starting means for causing the time measuring means to start measuring time when the abnormal state is detected by the abnormality detecting means; time measurement interrupting means for interrupting time measurement by the counter during a period of access from the host; time-exceeding detection means for detecting that the time measurement value of the counter has reached a predetermined value; If the detection means detects an abnormality and the time excess detection means detects that the time value of the timer exceeds the predetermined value, notify the host of detailed information about the abnormality. An abnormality notification device in a subsystem, comprising: a notification means for: