JPH02143333A - Fault recover device - Google Patents

Abstract

PURPOSE:To realize high speed fault recovery even for faults occurring frequently from the same information processing device without deteriorating the resolution of faulty data by using a log area as circulating it every time the fault occurs. CONSTITUTION:An information processing device group 1 consisting of plural information processing devices 11,12,... in and a fault recovering device 2 to process the fault at the time of the occurrence of the fault in the information processing devices 11 to 1n are provided. Besides, the device is constituted by providing a service processor (SVP) 3 which receives log data in response to a request from the fault recovering device 2 and stores it in an auxiliary storage device. The log area is secured at every occurrence of the fault, and is used as being circulated. Thus, the fault recovering can be executed at high speed by raising the duty rate of the log area.

Description

[Detailed Description of the Invention] [Field of Application] The present invention provides a service processor that is connected to a plurality of information processing devices and a service processor. Collect data at the time of failure, temporarily store the error message in the log area, process the failure using the log data at the time of failure, and when the failure handling is complete, send the log data to the service processor. Regarding the handover failure handling device.

[Conventional technology]

Conventionally, when a failure processing device receives a report of a failure from an information processing device, it sequentially extracts all internal information (log data) from the information processing device where the failure occurred first.
Log data is temporarily stored in a log area specified in advance for each information processing device on the memory in the fault processing device, and after fault processing for the information processing device in which the fault has occurred is performed based on the log data, the magnetic storage device A request to retrieve log data was issued to a service processor (SVP) having a storage device such as the above, and the log data was added to the SVP.

There are two types of failures: intermittent and fixed, but the failure processing device makes it possible to recover from intermittent failures, and if it is possible to re-execute an instruction (instruction retry) based on log data, it will retry the instruction. However, if the failure is fixed rather than intermittent, the failure processing will receive the failure report again and increase all the internal information of the information processing device, even if it is due to the same cause. The separation process was being performed. In addition, if a failure occurs again even if the instruction retry has been completed, the failure handling device will transfer the log data to the log area until it becomes free, if the log data has not yet been handed over to the SvP. The system had to wait for the log data to be stored, and when the log area became empty, it started troubleshooting and collected the log data.

[Three Problems to be Solved by the Invention] In the conventional failure processing device mentioned above, a log area is determined for each information processing device, and log data is received by an SVP that has a storage device such as a magnetic storage device. Therefore, if a failure occurs frequently in the same information processing device in the nl where the log data is taken over by SVP and the specified log area becomes full, failure processing for the same information processing device will have to wait until the log area becomes free, and the same Even if a failure occurs from an information processing device due to the same cause, all log data from the failed device is collected regardless of the cause of the failure, so most of the time spent on troubleshooting is spent on collecting log data! ! This method has the disadvantage that high-speed failure processing cannot be performed because of the large amount of data.

[11 Steps to Solving the Problem] The first fault handling device of the present invention includes: a log area management table that holds the base address of log data every time log data is stored in the log area; Obtain the latest base address and oldest base address of the log area currently in use, determine whether the log area can be secured based on the data collected as log data this time, and if it is determined that the log area can be secured, log Stores the base address of the next log area to be secured in the management table and instructs the log data collection unit to collect log data. When the service processor completes receiving the log data, it stores the base address of the collected log data. The log area usage status management section to be deleted from the log area management table is deleted.

The second failure processing device of the present invention includes a log area base address of log data stored in the log area, a device code of the information processing device corresponding to the log data, and an indication of whether the log data is the latest log. A log area management table containing the V bit, a device status management table containing various information for each information processing device including information indicating whether an instruction is being retried, and log data collected depending on the failure occurrence status of the information processing device. Device status management refers to the defined log data definition table and the device status and deaf management table to determine whether the information processing device that has received a failure report has already experienced a failure and is retrying commands. and, if the command is not being retried, determine whether log data of the same information processing device still remains in the log area or not by referring to the device code in the log area management table;
In addition, there is a log area usage state management unit that determines whether the log data is the latest log data using the V bit and updates the log area management table, and if it is determined that there is a failure during an instruction retry, The data required to determine the cause of the failure defined in the log data definition table is collected from the information processing device where the failure occurred, and it is determined whether the cause is the same as the failure before the command retry, and the cause of the failure to be retried is determined. If a failure occurs again, collect all remaining log data from the failed information processing device, and if the command is not being retried and no log data from the same information processing device remains in the log area. collects all log data defined in the log data definition table from the failed information processing device, and if the command is not being retried and log data from the same information processing device remains in the log area. Collect the failure cause determination log data defined in the log data definition table from the information processing device that caused the failure, read out the log data remaining in the 20-group area, compare both, and determine whether the cause is the same and the cause is the same as the one between the two sides. If the fault is not due to the same cause, collect all remaining log data, and if the failure is due to the same cause, add the log data necessary for command retry defined in the log data definition table. It has a log data collection unit that collects data from the information processing device that has caused the failure.

[For production]

In the first fault processing device, a log area is secured every time a fault occurs and is used cyclically, so that the usage rate of the log area can be increased and fault "Jf" processing can be performed at high speed.

In the second failure processing device, log data collection is performed based on the status of the information processing device (is the instruction being retried?), whether log data from the same information processing device remains in the log area, and whether the cause of the failure is the same. Since the control is more finely controlled, high-speed failure processing can be performed even for frequently occurring failures from the same information processing device.

〔Example〕

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

FIG. 1 is a configuration diagram of an information processing system including first to embodiments of the failure processing device of the present invention, FIG. 2 is a diagram showing how to use the log area 20 in this embodiment, and FIG. 3 is a diagram showing log data. FIG. 4 is a flowchart showing the collection processing of the log area, and FIG. 4 is a flowchart of the log area release processing.

This information processing system includes a plurality of information processing devices 11.1.
2. When a failure occurs in the information processing device n1 consisting of . It consists of 5VP3 stored in the auxiliary storage device.

The failure processing device 2 has a log area 20 and a log area 20.
FIFO-structured log area management table 2I that holds the base address every time log data is stored in
, a log data collection unit 22 that collects log data from the failure occurrence +i'r M processing device, a log data storage unit 23 that stores the collected log data in the log area 20,
Obtain the latest base address and oldest base address of the currently used log area 20 from the log area management table 21, determine whether the log area can be secured based on the amount of data to be collected as log data this time, and set the log area. If it is determined that the log can be secured, the log management table 21
When storing the base address of the next log area to be secured, Party B instructs the log data collection unit 22 to collect log data, and when the collection of log data is completed by 5vP3, the base address of the collected log data is stored. When the log area usage state management unit 24 deletes from the log area management table 21 and the failure process is completed, 5
It is composed of a fault processing section 25 that issues a request to take over log data to the VP3. Next, the operation of this embodiment will be explained with reference to FIGS. 2 to 4.

When a failure occurs in any of the information processing devices in the information processing device group 1, the failure processing device 2 receives the report and starts failure processing.

First, the log area usage state management unit 24 obtains the latest base address and oldest base address of the currently used log area from the log area management table 21, and uses the data size to be collected as log data this time to determine the log area. It is determined whether fi1 can be secured (process 30). When it is determined that the log area can be secured, the base address of the next log area to be secured is stored in the log area management table 21, the log area management table 21 is updated (31), and the log data collection unit starts collecting the log data. 22 (process 32).

After this, the log data collection unit 22
For example, if the usable log area starts near the limit of the log area and returns to the lower limit, when storing the log data, the data is collected from the lower limit of the contact log area that has stored data up to the limit. Subsequent log data is stored (area A in FIG. 2 corresponds to this).

When the failure processing unit 25 completes the failure handling, it requests the 5VP3 to take over the log data.

However, even if a failure report occurs from the information processing device group 1 while the log data is being received by the 5VP3, the base address of the new log area 20 is determined in process 30, log collection is performed in process 31, and failure handling is delayed. (Area B in Figure 2).

In this way, even if a failure occurs in the information processing device before the log data is received by the 5VP3, failure processing can be performed until the log area is full. When the 5VP3 completes receiving the log data, the log area usage state management unit 24 checks which log has been collected (40), deletes the corresponding log area pass address from the log area management table 21, and deletes the log area from the log area management table 21. is released (process 41) to prepare for the next failure.

By circulating the log area and using it in this way, it is possible to increase the usage rate of the log area and perform fault processing at high speed.

FIG. 5 is a configuration diagram of an information processing system including a second embodiment of the failure handling device of the present invention, FIG. 6 is a format diagram of the log data definition table 56, and FIG. 7 is a format diagram of the log area management table 51. , FIG. 8 is a format diagram of the device status management table 55.

This information processing system includes a plurality of information processing devices If, 1
2. When a failure occurs in the information processing device group 1 consisting of -, In and the information processing devices 11 to 1n, the failure processing device 5 that handles IIX damage processing and the log processing device 5 upon request from the failure processing device 5 It consists of 5VP3 that receives data and stores it in an auxiliary storage device.

The failure processing device 5 has a log area 50 and a log area 50.
Area base address of log data stored in,
A log area management table 51 (FIG. 7) including the device code of the information processing device corresponding to the log data, the V bit indicating whether the log data is the latest log, and information indicating whether the instruction is being retried. A device status management table 55 (FIG. 8) containing various information for each information processing device included; a log data definition table 56 (FIG. 6) defining log data to be collected depending on the failure occurrence status of the information processing device; The device status management table 5 determines whether the information processing device that has received the failure report has already experienced a failure and is retrying the command.
The device status management unit 57 identifies the information processing device by referring to 5, and the device code in the log area management table 51 determines whether log data of the same information processing device still remains in the log area when the instruction is not being retried. to determine whether the logo data is the latest log data.
Judgment based on bits, log area management table 51
The log area usage state management unit 54, which updates the log area, determines whether the faulty information processing device is retrying instructions, whether log data of the faulty information processing device remains in the log area, and whether the cause of the fault is a bait. from the log data collection unit 52 that collects log data from the information processing device that has caused the failure depending on whether it is the same as or not, and the failure processing unit 58 that performs failure processing and requests the 5VP3 to retrieve the log data when the processing is completed. It is configured.

Next, the operation of the tree embodiment will be explained with reference to FIG.

When a failure occurs in one of the information processing devices in the information processing device group 1, the failure processing device 5 receives the alarm and starts failure processing. First, it is determined whether the faulty device has already caused a fault and is retrying the command by referring to the device status management section 57 or the device status management table 55 (process 61). If the command is not being retried, the log area usage state management unit 54 stores the log data of the same information processing device at 5V in the log area 50 on the failure processing device 5.
It is determined by the log area management table 51 whether or not the log area has not been handed over to P3 and still remains (processing 62).

If it remains in the log area 50 on the failure processing device 5, the log data collection unit 52 from the information processing device that caused the failure
collects all the log information defined in the log data definition table 56 as failure information (process 63). At this time, the log area usage state management unit 54 updates the V bit indicating the latest data log. Then, processing such as command retry processing is performed depending on the content of the failure (processing 71), and a log data collection request is issued to the lfi SVP 3 to flush out the log data (processing 73), and the failure processing is performed. end. On the other hand, in process 62, the log data of the device that caused the failure is stored on the failure processing device 5.
If the log data remains in the log area 50, the log data collecting unit 52 collects the fault determination log data from the information processing device that has caused the fault by referring to the log data definition table 56 (IA management 64). Then, the log area usage state management unit 54 refers to the V bit from the log area management table 51, selects the latest log data from the log data that has already been collected, and compares it with the failure determination information. It is recognized whether or not there is a fault (processing 65). If the failure is not due to the same cause, the remaining log data is collected (process 66). At this time, the log area usage state management unit 54 updates the V bit indicating the latest data log. Then, processing such as instruction retry processing is performed depending on the content of the failure (processing 71). Conversely, if the cause is the same, the log data collection unit 22 collects the data necessary for handling the failure in process 71 as additional data.
f! ! 67), and passes control to process 71. Next, process 6
If it is determined in step 1 that the failure is occurring during an instruction retry, the log data collection unit 22 collects the data necessary for determining the failure and collects the information about the cause of the failure. Process 68) is collected from the A physical device, and it is determined whether the cause is the same as the failure of instruction retry rn (process 69).

If a failure occurs again due to a cause different from the instruction retry moe, the remaining log data is collected (process 70). At this time, the log area usage state management unit 54 updates the V bit indicating the latest data log. Then, trouble processing is performed based on the collected log data (processing 72).
The process 73) instructs the VP3 to clean out the log data, and the process is completed. If the cause is the same, the process is directly passed to process 72.

〔Effect of the invention〕

As explained above, the present invention increases the usage rate of the log area and speeds up failure processing by circulating the log area each time a failure occurs. By controlling the presence or absence of residual log data and the cause of failure in detail, it is possible to perform high-speed failure processing even for frequent failures from the same information processing device without reducing the resolution of failure data. .

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of an information processing system including a first embodiment of the failure processing device of the present invention, FIG. 2 is a diagram showing how to use the log area 20 in this embodiment, and FIG. 3 is a diagram showing log data. FIG. 4 is a flowchart showing log area release processing, FIG. 5 is a configuration diagram of an information processing system including the second embodiment of the failure handling device of the present invention, and FIG. 6 is a flowchart showing log data collection processing. FIG. 7 is a format diagram of the definition table 56, FIG. 7 is a format diagram of the log area management table 51, FIG. 8 is a format diagram of the device status management table 55, and FIG. 91J is a flowchart showing the operation of the second embodiment. . Lid......+'? i-information processing device group, 11~]
n... Information processing device, 2.5... Failure processing device, 20.50... Log area, 21.51... Log area management table, 22.52
...Log data collection section, 23.53...Log data storage section, 24.54...Log area usage state management section. 25, 58...fault processing unit, 55...device status management table, 56...
...Log data definition table, 57...
・・・・・・Device status management department, 3・・・・・・・・・
・svp.

Claims (1)

[Claims] 1. When a plurality of information processing devices and a service processor are connected and a failure occurs in the information processing device, data at the time of failure is collected from the information processing device and temporarily stored in the log area. A fault handling device that stores fault information automatically, performs fault processing based on log data at the time of the fault occurrence, and when the fault processing is completed, issues a log data takeover request to the service processor and hands over the log data to the service processor. , every time log data is stored in the log area, the log area management table that holds the base address and the latest base address and oldest base address of the currently used log area are obtained from the log area management table. Then, it is determined whether the log area can be secured based on the amount of data to be collected as log data this time, and if it is determined that the log area can be secured, the base address of the next log area to be secured is stored in the log management table, and the log data is and a log area usage state management unit that instructs the log data collection unit to collect the log data and, when the service processor completes the collection of the log data, deletes the base address of the collected log data from the log area management table. Characteristic failure handling device. 2. When multiple information processing devices and service processors are connected and a failure occurs in the information processing device, data is collected from the information processing device at the time of failure, and the failure information is temporarily stored in the log area. Then, trouble processing is performed using the log data at the time of the fault occurrence, and when the fault processing is completed, the fault processing device issues a log data collection request to the service processor and stores the log data in the log area. A log area management table including a log area base address of the log data currently being processed, a device code of the information processing device corresponding to the log data, a V bit indicating whether the log data is the latest log, and whether or not an instruction is being retried. A device status management table that contains various information for each information processing device, including information indicating whether the information processing device A device status management unit that refers to the device status management table to determine whether the information processing device has already experienced a failure and is currently retrying an instruction; Determine whether log data of the information processing device still remains by referring to the device code in the log area management table;
In addition, there is a log area usage state management unit that determines whether the log data is the latest log data using the V bit and updates the log area management table, and if it is determined that there is a failure during an instruction retry, Collect the data required to determine the cause of the failure defined in the log data definition table from the information processing device where the failure occurred, determine whether the cause is the same as the failure before retrying the instruction, and check if the cause is different from the cause before retrying the instruction. If the problem occurs again,
If all the remaining log data is collected from the information processing device that caused the failure, and if the instruction is not being retried and no log data from the same information processing device remains in the log area, the log data defined in the log data definition table is Collect all log data from the failed information processing device, and if the command is not being retried and log data from the same information processing device remains in the log area, check the log data defined in the log data definition table. Collect the failure cause determination log data from the information processing device that caused the failure, read the log data remaining in the log area, compare the two, determine whether the cause is the same as the previous one, and if the failure is not due to the same cause. collects all remaining log data, and in the case of a failure due to the same cause, collects the log data from the information processing device that caused the failure as additional data with the log data required for command retry defined in the log data definition table. A failure processing device characterized by having a data collection section.