JPH0497445A - Diagnostic system for information processor - Google Patents

Abstract

PURPOSE:To easily detect a defective position at the time of generating a fault and to execute processing within a short period by providing this diagnostic system with a means for automatically pointing out the fault position. CONSTITUTION:An information processor l is provided with a start small test means 1111 for executing a test by the minimum number of hardwares and successively expanding a test range, a test ID setting means 1112 for setting up a test number in a software (SF) visible register 12 prior to the execution of respective tests in a test program and a logout requesting means 1113 for outputting an error log sampling request to a service processor 3 connected to the processor l through a diagnostic pass 2. The service processor 3 is provided with an error logging means 311 for sampling an error log in an error log file 41 when a request is outputted from the means 113 and an error log analyzing means 312 for displaying a defective parts name based upon a test ID. Consequently, a defective position can easily be detected at the time of generating a fault and the fault can be repaired within a short time.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for diagnosing an information processing device, and particularly to a method for diagnosing an information processing device using a test program called Patrol JOB.

[Conventional technology]

Conventionally, when a comparison error is detected in diagnosing an information processing device using a patrol job, the system is stopped, an error log is collected, and a failure location is manually pointed out from the error log information.

[Problem to be solved by the invention]

As mentioned above, in the conventional diagnostic method for information processing equipment, the location of the failure was manually pointed out. Although it is possible to detect the occurrence of an error, it takes time to point out the location of the failure, making repairs difficult. The problem was that it lasted a long time.

SUMMARY OF THE INVENTION An object of the present invention is to provide a diagnostic method for an information processing device that can easily identify the location of a failure and repair it in a short time when a failure occurs.

[Means to solve the problem]

A diagnostic method for an information processing device according to the present invention is a diagnostic method for an information processing device in which a test program is executed while an OS is running to detect a failure, in which a test is performed using the minimum hardware of the information processing device and the test range is sequentially determined. a start small test means for expanding the test program; a test ID setting/setting means for setting a corresponding test number in a software visible register before executing each test of the test program; The information processing device includes logout request means for issuing an error log collection request to a service processor connected to the device via a diagnostic path, and when a request is received from the logout request means, the error log is converted into an error log file. This is a refinement in which the service processor is equipped with an error logging means for collecting data on the test ID, and an error logging means for displaying the name of the failed part based on the test ID.

〔Example〕

Next, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a system configuration diagram of a diagnostic method for an information processing device showing an embodiment of the present invention. It shows that The main storage unit 11 of the information processing device 1 includes a start small test means 11.
11, a test ID setting means 1112, and a logout requesting means IJ13. Perform test 1. Furthermore, Figure 1 shows 5VP
3, error logging means 311 and error logging means 3
12. FIG. 2 is a diagram showing the correspondence between each test of the test program by start small execution and the hardware. Figure 3 is a patrol job
2 is a flowchart showing an example of a test method. FIG. 4 is a diagram showing the processing of the DIAG command for maintenance diagnosis. FIG. 5 shows an error logging means 311 and an error logging means 312.
2 is a flowchart showing an example of processing of a maintenance program consisting of. FIG. 6 is a diagram showing an example of an FRU analysis table for analyzing parts replacement units.

Next, using FIG. 2, the correspondence between each test of the test program and the hardware (HW) to be tested will be explained. Each test in the test program has its own test number, and the tests are executed sequentially starting from test number 1. The test numbers are arranged in such a way that as the number increases, the number of pieces of hardware to be tested in the device under test increases. First, when test #1 in FIG. 2 is executed, test #1 executes a test of hardware A. Next, when test #2 is executed, test #2 tests hardware A and hardware B. If a comparison error is detected when test #2 is executed, it can be seen that a comparison error has been detected by hardware B (the shaded area). In this way, start small with 111
By executing the 6th test, it is possible to point out the location of the failure.

In addition, Patrol JOB III is started when ○5112 is started up, tests are performed sequentially from test #1, and when test #5 is successfully executed, the same patrol JOB 111 is started again after 1 hour.○5112 request. From then on, patrol JOB 141 will be executed every hour.

Next, a test method for Patrol JOBI 11 will be explained according to the flowchart shown in FIG. Patrol JOB III is
○When it is started from 5112, the test number # is displayed first.
The SW visible register (VR
) 12 (process 61) and execute test #1 (process 62). Test #1 has an expected value, and after the test is executed, the expected value and the actual value are compared (63). When a comparison error is detected, the Patrol JOBI 11 executes the DIAG command to request logout. 66). If no comparison error is detected, Patrol J
OBIII determines whether all tests have been performed or not.6
4) If there is a test that has not been executed yet, the process returns to step 61 and repeats the process of storing the next test number to be executed in the VR 12. 1 at 03112 when all tests are run
A request is made to restart the device after a certain period of time (process 65).

FIG. 4 shows the hardware execution flow of the DIAG instruction. When the DIAG instruction is executed in the patrol JOB 111, the hardware sets the bit of the error indicating flip-flop (EIF) 13 to 1, indicating that a comparison error has occurred (processing 71). When the bit of EIFlB is set to 1, the hardware stops the clock (process 72) and starts the maintenance program 31 on the 5VPB through diagnostic path 2 (process 73).

When the maintenance program 31 is activated, it analyzes the FRU according to the flow shown in FIG. 5 and outputs an error log. First, the maintenance program 31 determines whether the bit of EIFlB is set to 1 (81),
If 1 is set, the contents of EIFlB and VR12 are output from 5VPB to the error logging file 41 of disk 4 along with error log information (process 82). Furthermore, the maintenance program 31 extracts the test number from the VR 12 and analyzes the FRTJ based on the FRU analysis table 42 on the TIF 4 (processing 83). here,
Assuming that VR12 contains number 2, the hardware to be tested is NW A and HW from Figure 6.
It becomes B. Since HW A has been tested in test number 1, the hardware that caused the comparison error is HW
It turns out that it is B. When the FRU is analyzed, the maintenance program 31 outputs the FRTJ on the console 5 (process 84).

In this manner, the information processing apparatus diagnostic method according to the present invention can point out the failure location using the patrol job.

〔Effect of the invention〕

As explained above, the present invention provides a means for automatically pointing out a failure point in diagnosis of an information processing device using patrol JOB, so that when a failure occurs, the failure point can be easily identified in a short time. Able to handle repairs.

[Brief explanation of drawings]

FIG. 1 is a system configuration diagram of a diagnostic method for an information processing device showing an embodiment of the present invention, FIG. 2 is a diagram showing the correspondence between each test of a test program by start small execution and hardware, and FIG. FIG. 4 is a flowchart showing an example of a patrol JOB test method; FIG. 4 is a flowchart showing DIAG command processing; FIG. 5 is a flowchart showing an example of maintenance program processing;
The figure is a diagram showing an example of an FRU analysis table. 1... Information processing device, 11... Main storage unit, 111.
... Patrol JOB, 1111 ... Start small test means, 1112 ... Test ID setting means, 111
3...Logout request means, 112...OS, 12
...SW visible register, 13...EIF, 2.
...Diagnostic path, 3...SVP, 31...Maintenance program, 311...Error logging means, 312...
Error logging means, 4... Disk, 41... Error logging file, 42... FRU analysis table, 5... Console.

Claims (1)

[Claims] In a diagnostic method for an information processing device that detects a failure by executing a test program while an OS is running, a test is started with the minimum amount of hardware of the information processing device and the scope of the test is gradually expanded. a small test means; a test ID setting means for setting a corresponding test number in a software visible register before executing each test of the test program; The information processing device includes a logout requesting means for issuing an error log collection request to a connected service processor, and an error logging means for collecting an error log into an error log file when a request is received from the logout requesting means. , an error log analysis means for displaying a failed component name based on the test ID, the service processor comprising: a diagnostic method for an information processing device;