JPH04112238A - Event log output system - Google Patents

Abstract

PURPOSE:To output the log of events without omission or overlap by saving a pair of recovery information of an actual I/O and the event number at the time of writing the event saved in a non-volatile memory in an event log store and writing the event based on these saved significant information at the time of restoration from crush or the like. CONSTITUTION:A resource management group 1 which detects events which occur in a system and an event report accepting part 3 are provided, and this part 3 gives an unequivocal event number to the event reported from the group 1 to save it in a non-volatile memory 4 and requests the write of this saved event to an external storage device (event log store) 6 and saves a pair of recovery information and the event number, which are returned correspondingly to this write request, in the memory 4. Consequently, recovery information of the actual I/O and the event number are saved and the effective display number is switched when the event saved in the memory 4 is written in the external storage device 6. Thus, the log of events is outputted to the external storage device 6 with-out omission and redundancy based on these saved significant information at the time of restoration from crush or the like.

Description

[Detailed Description of the Invention] [Summary] Regarding the event log output method for outputting a log of events that occur in the system, when writing events saved in non-volatile memory to an event log store (external storage device), recovery of the real I10 is performed. The purpose of this system is to save information and event numbers in pairs, and when the system recovers from a crash, etc., to write events based on the saved significant information, and to output event logs without omission or duplication. A resource management group that detects events that occur in
A unique event number is assigned to the event notified from this resource management group, and the non-volatile memory S: save, this save 5
A request to write the event to the external storage device and an event notification reception in which the recovery information (medium location where the event was written) and η/element number that are returned in response to this write request are saved as a pair to the non-continuous memory. The non-volatile memory card
If the system crashes while outputting a log of an event to an external storage device, after recovery, the non-volatile memory is The system is configured to retrieve an event with a corresponding event number from the memory, write it from the corresponding medium position in the external storage device, and output it as a log.

[Industrial application field]

The present invention relates to an event log output method for outputting a log of events occurring in a system. In order to efficiently log events that occur within a computer system,
There is a need for a mechanism to identify redundant event records that exist in a system due to a system crash or the like.

=Problems to be solved by the conventional technology and invention] Conventionally, as shown in the configuration of FIG. After receiving this notification, the event notification receiving unit 23 temporarily stores the events in the non-volatile memory (FS 5) 24, and then stores the events, for example, events A, B, and C in the buffer 25 and stores them in an external storage device (event log). Store) Request for writing to 26,
I was trying to output logs. At this time, the event notification receiving unit 23 recognizes the output position (writing position) to the external storage device 26 based on the last written information, so if the system crashes at the position (4), for example, There is a problem in that temporary events are written to the external storage device 26 and redundant event records are output to the log. The operation of the configuration shown in FIG. 5 will be briefly explained below.

(2) The notifier (resource management group 21) notifies the event notification reception unit 23 of the detected event (for example, disk abnormality) (FIG. 5 (1)).

■ The event notification reception unit 23 sends the notified event to the FSS.
(nonvolatile memory 24) (FIG. 5(2)).

■ Extract the saved event and store it in the buffer 25L,
Write to the external storage device 26 (Figure 5 +3+>6 ■ In response to the fact that the real I10 is returned from the event log store (external storage device 26), the event notification receiving unit 23 writes the recovery information and the last data of the real I10. The event number is saved in the FSS (non-volatile memory 24) ((5) in FIG. 5).

(e) Delete the events saved in the FSS24.

At this time, the event was stored in the buffer 25, but as shown in FIG.
4) If the system crashes while the data has not yet been written to the external storage device 26, the events stored in the buffer 24 will be lost. Accordingly, when the system is restored, the events saved in the non-volatile memory 24 are stored in the buffer 25 and written to the external storage device 26, so events are written redundantly and redundant event records are created. There was a problem that the log was output.

The present invention provides a method for writing actual ■/○
The purpose is to save the recovery information and event number as a - pair, and when the system recovers from a crash, etc., it writes the event based on the saved significant information, and outputs the event as a log without omission or duplication. It is said that

[Means to solve the problem]

Means for solving the problem will be explained with reference to FIG.

In FIG. 1, resource management group 1 is for detecting events occurring in the system.

The event notification receiving unit 3 assigns a unique event number to the notified event, saves it to the nonvolatile memory 4, requests to write the saved event to the external storage device 6, and returns it in response to this write request. The recovery information (medium location where the event was written) and event number are saved in a non-volatile memory as a pair, and the effective display number is switched.

The nonvolatile memory 4 saves events in association with arbitrary event numbers, saves the event numbers and recovery information that have been logged to the external storage device 6, saves valid display numbers, etc. It is something to do.

The external storage device 6 outputs and stores events as a log.

[Effect]

As shown in FIG. 1, in the present invention, the event notification receiving unit 3 assigns a unique event number to the event notified from the resource management group 1 and saves the saved event in the nonvolatile memory 4. A write request is made to the external storage device 6, and the correspondingly returned recovery information (medium location it where the event was written) and event number are saved as a pair to the non-volatile memory 4, and the effective display number is switched. Stay on your knees.

At this time, if the system is in trouble while outputting event logs from the non-volatile memory 4 to the external storage device 6, after recovery, significant recovery information corresponding to the valid display number saved in the non-volatile memory 4 will be saved. Based on the event number and the event number, the event with the corresponding event number is retrieved from the nonvolatile memory 4 and written from the corresponding medium position in the external storage device 6 using a printer.

Therefore, when writing the event saved in the non-volatile memory 4 to the event log store (external storage device 6), the actual I1
To save the recovery information and event number of 0 and to switch the effective display number, and to output a log to an external storage device 6 without omission or redundancy of the event based on the saved significant information when recovering from a crash or the like. I can do it.

〔Example〕

Next, the configuration and operation of one embodiment of the present invention will be sequentially explained in detail using FIGS. 1 to 4V.

In the 111th ffl, the resource management group 1 detects an event that occurs in the system and notifies the event notification reception unit 3. Here, the detected event notification person,
Event notification B, event notification C, etc. are sequentially notified to the event notification receiving unit 3.

The event notification reception unit 3 sequentially assigns unique event numbers to the events notified from the resource management group 1 and saves them in a non-volatile memory (FSS) 4, stores the saved events in a buffer 5, and stores them in external storage. A write request is made to the device 6, and the recovery information of the real I10 (medium position where the event was written) and the event number returned in response to this write request are saved as a pair in the non-volatile memory 6, and the valid display number is switched. etc.

A non-volatile memory (FSS) 4 is a memory whose contents do not disappear even when the power is turned off. The event number and recovery information are saved as a pair, and a valid display number indicating which of the saved event number and recovery information is valid (significant) is saved.

The buffer 5 is a buffer for writing events to the external storage device 6, and is used to write events to the external storage device 6 when the event becomes full.

The external storage device (event log store) 6 outputs and stores events as a log.

First, the normal operation of the configuration shown in FIG. 1 will be described in detail in accordance with the order shown in the flowchart of FIG.

■ Resource management group (event notifier) 1 detects an event (abnormality) that occurs in the system, and sends event notification A, event notification B, etc.
, is notified as event notification C to the event notification receiving unit 3 configuring the event control unit 2 ((1) in FIG. 1, 0 in FIG. 2).

■ The event notification reception unit 3 that received the event notification in ■ stores the event assigned the event number of
4 (Fig. 1 (2), Fig. 2 0).

■ The event notification reception unit 3 stores the events retrieved from the nonvolatile memory 4, such as event A, event B, and event C, in the buffer 5, and requests writing to the file (external storage device 6) (see FIG. 1). 3), Figure 2 0, @YES, ■).

■ In response to the request for writing to the external storage device 6 in step 3, the recovery information of the real I 10 (medium position where the event was written) and the last event number returned to the event notification reception unit 3 are stored in the non-volatile memory. 4 (Fig. 1 (5), Fig. 2 @) YES, @).

■ As shown in Figure 4, after saving the recovery information and the last event number as a pair to the non-volatile memory 4, switch the valid display number (for example, 1 (previous recovery information and event number are valid)) to 2 (current recovery information and event number are valid) (Figure 1 (6), Figure 2 [phase]
).

■ Since the log output of the event to the external storage device 6 has been completed, the event is deleted from the non-volatile memory 4 (Fig. 1 (7)
, Fig. 2 [phase]).

As described above, a unique event number is assigned to the event notified from the resource management group 1, and the event is saved in the non-volatile memory 4, and is also retrieved and output as a log to the external storage device 6. After the log is output, the event number and the last event are The numbers are saved in pairs in the nonvolatile memory 4 and the valid display numbers are switched. As a result, even if the system crashes at the time point (■) in Figure 1, the event number pointed to by the valid display number and the recovery information will be stored in the non-volatile memory 4 at the time of recovery by following the steps shown in the flowchart in Figure 3, which will be explained next. Based on this, it becomes possible to extract all events, including those that have not been output, from the nonvolatile memory 4 and output them to the log from the corresponding location in the external storage device 6 without omission or duplication. Hereinafter, in accordance with the order shown in the flowchart of FIG. 3, and with reference to FIG. 4, the operation at the time of restoration of the configuration of FIG. 1 will be explained in detail.

(1) A crash occurs at the time when the event notification receiving unit 3, which comes at the time point (■) in FIG. Figure ■ Crash occurrence, Figure 3 ■).

+2) (After the crash of 11 occurs and the system is restored, the effective display number saved in the non-volatile memory 4 in FIG. 4 is retrieved (FIG. 3, 0).

(3) (Valid display number retrieved with 2+, for example “1
” (the number indicating that the previous event number and recovery information are valid) are the recovery information and event number pointed to by, for example, the recovery information “!” from Figure 4. -m (the position on the medium of the event that completed the log output to the external storage device 6 last time) and the event number "n-k" (the last event number that was last logged to the external storage device 6) are retrieved ( Figure 3 O).

(41) The magnetic head is positioned to the file on the external storage device 6 based on the recovery information (position on the medium) retrieved in step 31 (FIG. 3 [phase]).

+51 Write starting from the event corresponding to the event number extracted in (3) from the position located in +41.

As described above, by writing to the external storage device 6 based on the recovery information and event number pointed to by the valid display number saved in the non-volatile memory 4, the event that was notified and saved in the non-volatile memory 4 Even if the system crashes, it is possible to write to the external storage device 6 without omission or duplication upon recovery.

FIG. 4 shows an example of saving event number/recovery information according to the present invention. Here, the pair of event number and recovery information is the one that was saved in the nonvolatile memory 4 when the event was written to the external storage device 6 and the actual ■/○ was returned (see Figure 1 (5).
)). The valid display number is a number indicating the current valid event number and recovery information. For example, in FIG.
This is for switching to point to the last event number and recovery information pair that was saved when 10 was returned to the event notification receiving unit 3. Therefore, the pair of event number and recovery information indicated by this valid display number indicates that the event up to the event number has been written up to the position of the medium indicated by the recovery information on the external storage device 6.

〔Effect of the invention〕

As explained above, according to the present invention, when writing the event saved in the nonvolatile memory 4 to the event log store (external storage device 6), the recovery information and event number of the actual ■/○ are saved and displayed as valid. Since we have adopted a configuration in which events are written based on the saved significant information when the system switches to the non-volatile memory 4 and recovers from a crash, etc., the events saved in the non-volatile memory 4 after receiving notifications are recorded without omission or duplication. The log can be output to the external storage value W6 without having to do so. Thereby, it is possible to improve the usage efficiency of files on the external storage device 6.

In the diagram, 1: Resource management group (event notifier) 2: Event control section 3: Event notification reception section 4: Non-volatile memory (FSS) 5: Hasofa 6: External storage device (event log store)

Claims (1)

[Claims] In an event log output method that outputs a log of events that occur in a system, a resource management group (1) that detects events that occur in the system, etc.
), assigns a unique event number to the event notified from this resource management group (1), and saves it to the non-volatile memory (4).
Request to write this saved event to external storage device (6),
and an event notification reception unit (3) that pairs the recovery information (medium location where the event was written) returned in response to this write request with the event number and saves it to the nonvolatile memory (4). If the system crashes while outputting an event log from the memory (4) to the external storage device (6), significant pairs of recovery information and event numbers are saved in the non-volatile memory (4) after recovery. Based on this, the event with the corresponding event number is extracted from the non-volatile memory (4), written from the corresponding medium position in the external storage device (6), and output as a log. Event log output method.