JPH02190945A - Extraction system for fault analyzing information - Google Patents

Abstract

PURPOSE:To extract the fault information of a device by allowing a device relating to abnormality to execute logout (storage of internal information in a storage area) when another unit or device detects abnormality. CONSTITUTION:A program in a CPU 10 monitors time required up to the end of data transfer, and when the data transfer is not ended within a fixed time, supposes the generation of a fault in any one of the CPU 10, a CHC 13 and an I/O 14 and commands the CPU 10 to forcedly stop the operation. The CPU 10 sets up a halt pause(HP) or clear pause(CP) bit in a subchannel 15 of a main storage device 11 and the CHC 13 disconnects an interface with the I/O 14. When an internal fault is generated in the CHC 13, a SVP 12 receiving the report of the internal fault requests the extraction of the internal information to the CHC 13 and the CHC 13 extracts the internal information in a log data area 17, and after the completion of the internal information extraction, the SVP 12 stores the information obtained from the log data area 17 in a logout area 16. Consequently, the internal information of a device or unit which does not detect a fault by itself can be extracted.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a fault analysis information collection method in a computer system.

[Conventional technology]

In conventional devices, when the hardware detects an internal failure,
It stores hardware information deemed necessary for analyzing the cause of a failure in a pre-specified storage area, and also resets the internal state to prepare for the next operation and prevent failure information from being lost. Additionally, a method was used to clarify the cause by analyzing the fault information in the storage area.

For information on how to store fault information in the storage area when a fault is detected, see Sanpo Publishing's "Computer Course: Computer System Design"
It is described in Chapter 4.4 Error Control (pages 216-2°18) in Chapter 4 Availability Design, written by Kozo Kayashima, first published June 20, 1972.

[Problem to be solved by the invention]

In the conventional technology described above, only internal information within the device itself is stored in the storage area when an abnormality is detected. That is, even if the cause of the abnormality is in a related device other than the own device, only the internal information of the own device is stored. If the cause of the abnormality is in a related device other than the own device, and the device does not treat it as an abnormality or is unaware that it is an abnormality, conventional technology stores the internal information of the device that caused the abnormality in its storage area. There was no way to take it out. Therefore, there was a problem in that information useful for failure analysis could not be obtained.

An object of the present invention is to store internal information in a device related to the abnormality from the outside even if the abnormality is not detected internally when another unit or device including a program detects an abnormality. Store in area (log out from now on)
The purpose is to carry out the following.

[Means to solve the problem]

The above purpose is to explicitly issue a logout instruction to a device related to the abnormality when another unit or device detects an abnormality, and to collect log information of the related device.

Alternatively, an instruction means can be set in advance so that the related device collects log information in response to special processing from another unit or device, or logs out in response to processing performed by another unit or device. This is achieved by providing log information within the device and collecting log information when an event indicated therein occurs or is performed in the related device.

[Effect]

The device or unit that detects the failure issues explicit logout instructions to the devices or units that are believed to be involved. Or, perform special processing to trigger the device or unit to log out, or provide a setting means in advance for the device or unit, and log out when an event specified by the setting means occurs in the device or unit. Have them do this. You received an explicit logout instruction.

Alternatively, a device or unit that has undergone special processing specified as an opportunity to log out, or that has experienced an event that causes logout specified in advance by the setting means, can specify internal information even if it has not detected any abnormality. The data is stored in the specified storage area. As described above, it is possible to collect internal information of a device or unit that has not detected a failure itself.

〔Example〕

An embodiment of the present invention will be described using FIGS. 1 to 6.

FIG. 1 is a configuration diagram of a computer system including the present invention.

When a program in the central processing unit (hereinafter referred to as CPU) 10 starts up and transfers data to the input/output device e (hereinafter referred to as 10) 14, the program transmits 5SCH (ST) to the CPU 10.
ART 5UBCH) command. The CPU 10 sends the CCW address (CCWA
DR) and the start pending (sp) bit indicating 5SCH activation. Input/output processing unit 13 (hereinafter referred to as CHC)
periodically checks the subchannels in the main memory 11, and if there is a subchannel whose SP bit is set, it activates the IO corresponding to that subchannel.

When the SP bit of subchannel 15 is set,
CHCl3 is.

Activate the corresponding l014. Upon successful activation, CHCl3 resets the SP bit and sets the subchannel active (SA) bit.

Also, when the data transfer is completed, the SA bit is reset and the status pending (STP) bit is set. The program in the CPU 10 monitors the time from when the CPU 10 is started until the data transfer is completed, and if the data transfer is not completed within a certain period of time, that is, S 'l'
If the P bit is not set by CHCl3,
CPU10. (1: It is assumed that a failure has occurred in one of Hc131014, and HALT or CL is applied to CPU10.)
Issues an EAR command to instruct forced stop of operation. CPU
When l0 receives a HALT or CLEAR command, it sets the halt pending (HP) or clear pending (cp) bit in the main storage subchannel. CHCl3 disconnects the interface with 1014 if the HP or CP bit is set in the subchannel 15 like SP.

Also, if an internal failure occurs in CHCl 3, the
An internal fault is reported to the S_P 12 by the interface 21 to the S_P 12. The 5VP 12 uses the interface 22 to request the CHCl3 to collect internal information. Interface 21 is a 1-bit interface that indicates to SVP 12 that an internal failure has occurred in CHCl3, whereas interface 22 is a 1-bit interface that indicates to SVP 12 that an internal failure has occurred in CHCl3.
It is a bidirectional interface that can exchange data with P12. The command format on the interface 22 will be shown later using FIGS. 4 and 5. Upon receiving a request to collect internal information from the 5VP 12, CHCl3 collects the internal information in the log data area 17 and indicates that the internal information has been collected. 5VP 12 using the interface 22.

The 5VP 12, which has received the completion of internal information collection, stores the information in the log data area 17 sent from the CHCl 3 in the logout area 16.

FIG. 2 is a configuration diagram of the subchannel 15.

The SP, HP, and cp bits are set by the program in CPU10 to set 5SCH, HALT, and CLE to the corresponding subchannel.
It issues AR commands and indicates that %CHCl3 has not yet processed them.

The SA bit indicates that 1014 is in the process of data transfer. STP pinto indicates that data transfer has ended.

CCWADR is L indicating the address on main memory where the channel command word (CCW) for data transfer is located.
OGP 201 and LOGCTL 208 are bits and areas newly set in the present invention, and their purpose will be shown later.

The operations corresponding to each complaint shown in the claims are shown below.

1. Due to some failure, data transfer ends at 5TP20.
When 6 cannot be set, the program is HALT according to the previous explanation.
In the present invention, a new command called a logout command is issued prior to this, in which the operation is interrupted by issuing a CLEAR command. When the program issues this logout command, CPU10 will log out from the LO in subchannel 15.
Set GP201. CHCl3 detects subchannel 15 where LOGP201 is set, and
After resetting GP201, the sixth
Collect data using the configuration shown in the figure.

Figure 6 shows that when a failure occurs in the conventional CHCl3, the CHC
It consists of internal information 601 collected by l3 and 5UBCH information 602 in subchannel 15 newly added in the present invention. CHC that collected data in log data area 17
l3 uses the interface 22 to issue a request in the data format shown in FIG. 5 504 to the 5VP 12.

Furthermore, Figure 5 shows the format of the command issued from CHCl3 to 5VP12.
Following 504, P12 stores the data in the log data area 17 transferred from CHCl3 in the logout area 16 as LOG data that can be collected according to the present invention. In addition, if a failure occurs within CHC13,
5VP12 data in the format shown in Figure 4 401
is issued to CHCl3 to only collect internal information. Incidentally, FIG. 4 shows the format of the command issued from 5vP12 to CHCl3. When CHCl3 collects the internal information data, it sends a request for the data format shown in 503 to 5V using the interface 22.
P12 and the fault 1 information collected in the log data area 17 is stored in the logout area 16 by data transfer.

2. The program performs special processing such as forced stop of operation.
When issuing a HALT or CLEAR command to the HC, the CPU 10 sets HP203 or CP2O4 in the subchannel.

When CHCl3 detects a subchannel in which HP203 or CP2O4 is set, conventional HAL'r or CL
In the present invention, CHCl3 is separated from HP2 as defined by the EAR instruction.
When detecting 03 or CP2O4, CHC at the time of detection
The internal information of l3 and the information of subchannel 15 are transferred to CHCl.
3 is collected in the log data area 17. After that, the conventional I
Performs interface separation processing with on4.
The method of transferring data from the log data area 17 to the logout area 16 is the same as described above.

3.5VP12 uses interface 22 to connect CHC
Reading and writing of subchannel 15 is possible through l3. 5VP12 to CHCl3 Fig. 4 402
When you send a request in the format shown in , CHCl3 returns 40
The subchannel indicated by the 5UBCH number of 2 is read from the main storage device, and following the formatting of 502, the subchannel information is transferred to the 5VP12. Similarly format 4
Following the request shown in 03, 5VP12 sends the subchannel information to be written to CHCl3, and CHCl3 sends 40
Using this read and write interface that writes the data sent from the 5VP12 to the subchannel in the main memory indicated by the 5UBCH number of 3, the present invention sets/resets the LOGCTL 208 in the subchannel on the main memory to the SVP. LOGCTL starting from 12
The format of 208 is shown in Figure 3. If bit O or bit 1 is set, it means that the HALT and CLEAR commands are issued to the corresponding subchannel, that is, the HP and CP bits of the corresponding subchannel are set. When the CHC confirmed the LOGP bit explained in Claim 1 and the HP and CP bit explained in Claim 2, the internal information and subchannel information were collected as LOG information in the same way as Collect internal information and subchannel information.

After that, l0 by conventional HALT and CLEAR commands.
Performs interface separation processing with 14. Note that the method of transferring data from the log data area 17 to the logout area 16 is the same as described above.

4. During startup and data transfer between CHCl3 and l014, when CHCl3 detects an abnormality in the CCW used for data transfer, or when l014 detects an abnormality, CHCl
3 and 1014 interrupt the processing, notify the program of an abnormality by L, program check, unit check, etc., and perform recovery processing by the program.0 In the present invention, bits 5 and 6.7 in LOGCTL 208 are This is the LOGOUT control bit for IL, program check, and unit check, and when the CHC detects that the corresponding event has occurred, the internal information and subchannel information of the CHC are collected in the same manner as described above. Note that the method of transferring data from the log data area 17 to the logout area 16 is the same as described above.

According to this embodiment, the following effects are achieved.

(B) Since logout processing is performed using existing logout hardware, there is no need to provide special hardware.

(b) When the program detects an abnormal state that appears to be a hardware failure, it only needs to issue a newly provided logout command, and even after issuing this command, it is possible to continue processing as before.

(c) HALT/C issued by the program when an abnormal condition is detected, even if the program does not issue a special command.
It is also possible to log out using the LEAR command as a trigger.

(d) If a logout is performed unconditionally when a HALT/CLEAR command is issued, an unnecessary logout may occur if these commands are issued when there is no abnormal situation.

To avoid this, it is also possible to log out only when HALT/CLEAR is issued to a preset subchannel.

(e) By making settings in advance, it is also possible to log out in response to an exceptional event (IL, program check, unit check, etc.) detected within the input/output processing device.

In any of cases (c) to (e), the operation after logout can be continued.

〔Effect of the invention〕

According to the present invention, when another unit or device detects an abnormality, it is possible to log out the device related to the abnormality, so it is possible to collect failure information of the device that could not be collected conventionally, and to analyze the failure. It has the effect of improving abilities.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of an embodiment of the present invention, FIG. 2 is a configuration diagram of the subchannel 15 in FIG. 1, and FIG. 3 is a configuration diagram of the subchannel 15 in FIG. 2.
The configuration diagram of 208, Figure 4 is from Figure 1 5VP12 to CHC
An explanatory diagram of commands to l3. FIG. 5 is an explanatory diagram of commands from CHCl3 to 5vP12 in FIG. 1, and FIG. 6 is a configuration diagram of L0GOUT information collected by CHCl3. 10... Central processing unit, 11... Main storage device, 12
... Service processor, 13... Input/output processing device. 14... Input/output device! , 15... subchannel, 1
6... Logout area, 17... Log data area,
2L: Internal failure detection signal of input/output processing device. Shuzushu Zushima! //-- Main storage device/Z-1・5VP 13--・I/O processing device! ＋・・－Input/output device for the cow picture a"4-

Claims (1)

[Claims] 1. In a device that is equipped with a means for logging out internal information when a failure is detected and a storage means for storing the logged out information, it is also possible to log out by an explicit logout instruction from another unit or device including a program. A method for collecting failure analysis information characterized by: