JPH096652A - Trace collection method and computer system - Google Patents

Abstract

PURPOSE: To provide a trace collection method and a computer system which can keep the continuity of the trace data despite high load of a CPU of an execution program. CONSTITUTION: A trace collection program 105 is called out of an optional module of an execution program 103 during execution of the program 103 and a load monitor program 104. Then the program 105 refers to a trace importance number 114 that is updated in steps and according to the CPU load acquired from a trace collection management table 106 and retrieves a trace data collection range 107 that is previously set to the number 114. Thus the program 105 collects the trace when a module name is set and then does not collect the trace when no module name is set respectively.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system,
It relates to a method of collecting the program operation history (trace).

[0002]

2. Description of the Related Art In the trace data collection in a computer system, a trace collection program started by an online program to be traced stores the program name and calling parameters in a trace data area to obtain a program operation history. There is. in this case,
Japanese Patent Laid-Open No. 62-24336 is known in which a dynamic processing unit is provided in the trace collection control unit and the number of journal (trace data) collections is automatically adjusted depending on the load on the online processing system. ing.

[0003]

In the above-mentioned prior art, the number of times trace data is output is adjusted by changing the output waiting time to the external storage device according to the load state. However, when the high load state continues. Had a problem of lack of data continuity.

An object of the present invention is to provide a trace collection method and a computer system capable of maintaining the continuity of trace data even when the CPU load of the online control program is high.

[0005]

An object of the present invention described above is to monitor a CPU load amount or a transaction amount in a trace collecting method for collecting and storing a trace which is a program operation history and cope with a change in the amount. The trace collection range is set in advance so that the trace amount to be collected can be increased or decreased stepwise, and the trace collection process called from any module of the program being executed is the CPU load amount or transaction amount at that time and This is achieved by referring to the trace collection range to determine whether or not the trace collection of the module is necessary, and collecting the trace if necessary.

The trace collection range is set to be wide for low importance and narrow for high importance according to a plurality of importance according to the purpose, and the trace collection processing corresponds to the CPU load amount or transaction amount in stages. The feature is that the trace collection range to be referred to is selected according to the corresponding importance level.

The trace collection range is classified according to the importance,
A feature is that the call module name of trace collection is registered.

An object of the present invention is to provide a plurality of trace importance levels in a computer system comprising a trace collecting means for collecting and storing a trace of a program operation history by a call from a running program and a CPU load monitoring means. Separately, a trace collection range storage means for registering a wide range where the trace amount increases in low importance level and a narrow range where the trace amount decreases in high importance level in a stepwise manner, and fluctuation of the CPU load by the monitoring means. A computer system comprising switching means for changing the trace collection range according to the degree of importance in accordance with the above.

The trace collection range storage means is configured to be registered with a module name that calls the trace collection means, and the trace collection means refers to the trace collection range changed by the switching means. Then, the presence or absence of registration of the calling module name of the program being executed is determined, and if registered, the trace data is collected.

[0010]

According to the trace collection method of the present invention, the trace collection range setting table for changing the trace collection range (trace amount) according to the importance of the trace is provided to monitor the fluctuation of the CPU load (or transaction amount). Since the online trace collection range is switched according to the load amount, even when the load is high, the collection of highly important traces is maintained without adversely affecting the online processing. As a result, continuity of important traces (for example, traces of lock and unlock modules) such as investigation of the cause of system failure can be ensured, and the reliability of highly important trace data can be improved.

According to the computer system of the present invention,
The degree of importance of trace and its collection range can be arbitrarily set according to the state and purpose of the system, so that the generation and maintenance of the system can be facilitated.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of a computer system having a trace collection function of this embodiment. It is composed of a CPU 101, a main storage device 102, a system bus 109, a system bus control device 110, and auxiliary storage devices 111 and 112.

On the main storage device 102, a trace collection management table 106 having a trace collection range number 114 and a trace collection presence / absence flag 115, a trace collection range setting table 107, an arbitrary execution program 103 to be traced, and a trace processing program 104. , A trace definition program 105, and a trace data collection area 108 on the main memory 102. The trace data collection area 113 is also provided on the auxiliary storage device 112 such as a hard disk that can be permanently stored, and the trace data collected in the area 108 on the main storage is stored in the execution program 1
It is transferred and stored at the end of 03.

FIG. 2 shows the format of the trace collection range setting table. Trace collection range setting table 1
In 07, a module (program) name for trace collection is registered for each trace importance number. In the case of this example, 3 from importance level 3 to importance level 1 from low load to high load
It becomes a stage. The table of importance number 3 is open, c
lose, times, read, write, loc
Modules such as k and unlock are registered, and the trace amount in this range is large. Importance, number 2 is read, w
The write, lock, and unlock modules are registered, and the trace amount is medium. The importance number 1 is loc
The k and unlock modules are registered, and the trace amount is small.

The range of the trace importance can be arbitrarily set by the operator and is changed depending on the system status and the purpose of the trace. In this example, loc of shared resource
This is a case in which a system failure occurs due to the omission of the unlock process for the k (occupied) process. In order to investigate which program has the cause of the failure, the importance of the lock / unlock processing is set to be high, and the continuous traces of those are acquired. Generally, the importance of the module having a high influence on the system is increased. In the case of retrospective investigation for system failure or program debugging, the demand range can be set more purposefully.

FIG. 3 is a flow chart for explaining the processing outline of the computer system according to this embodiment. Arbitrary execution program 103 is open module 301, wri
The te module 302 is executed. The open module 301 calls the trace processing program 104 before the open, and the write module 302 calls the trace processing program 104 before the write. Hereinafter, a method of collecting trace data according to the CPU load will be described in detail.

FIG. 4 is a processing flow chart of the CPU load monitoring program according to this embodiment. This process is a delay
It is periodically executed at the monitoring cycle determined by the process (401). First, the CPU load measurement program 116 is started (402), and the trace importance number 114 and the trace collection presence / absence flag 115 of the trace management table 106 are updated according to the measured load. That is, if the load is less than 60% (403) importance number = 3 (407), if the load is less than 60% to 80% (404) importance number =
2 (408), if the load is less than 80% to 95% (4
05) Importance number = 1 (409) is set, and trace collection presence / absence flag = present. In this example, when the load is 95% or more, the trace collection presence / absence flag = none (406) is set, but the importance number = 1 may be included including this period.

FIG. 5 is a processing flow chart of the trace processing program according to this embodiment. This processing 104 is performed by the open module 301 in the execution program 103,
It is activated from the write module 302. First, the trace collection management table 106 is referenced (510).
Here, it is checked whether or not there is a trace collection presence / absence flag (511). If there is a flag, the trace importance number is referred to.

When the trace importance number = 3 (51
2), referring to the trace collection range setting table 107,
It is determined whether or not the calling module of the program 104 is registered in the trace collection range of the importance number 3 (515). If registered, collect the trace data of the calling module and trace data collection area 1
It is stored in 08 (516). On the other hand, if it is not registered, trace collection of the calling module at this point is not performed.

Similarly, when the trace importance number = 2 (513) and the trace importance number = 1 (514)
Also, referring to the trace collection range setting table 107,
Check if the calling module is registered (517,
519), collect and store trace data if registered (518, 520).

As described above, in the present embodiment, the trace collection range is dynamically switched according to the increase / decrease of the online CPU load. Therefore, in the high load state, the trace data amount is reduced to affect the execution program. While avoiding
Maintain collection of important trace data. This ensures continuous collection of trace data that is important for investigating the cause of system abnormalities such as deadlocks, or trace data that is important for some purposes, thus improving the reliability of trace data and maintaining the system. To facilitate.

In the above embodiment, the collection range of the trace data is switched according to the CPU load, but it may be switched by monitoring the transaction amount.

[0024]

According to the present invention, since the trace collection range divided according to the importance of the trace is dynamically switched according to the increase or decrease of the CPU load, the continuity of the important trace is maintained even in the high load state. , Has the effect of improving the reliability of data.

According to the present invention, the degree of importance of traces and the collection range thereof can be set according to the status and purpose of the system, so that there is an effect that the system can be easily generated and maintained.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a computer system having a trace collection function according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a setting example of a trace collection range setting table.

FIG. 3 is a flowchart illustrating the operation of the computer system of this embodiment.

FIG. 4 is a flowchart illustrating a CPU load monitoring process according to the present embodiment.

FIG. 5 is a flowchart illustrating a trace process of this embodiment.

[Explanation of symbols]

101 ... CPU, 102 ... Main storage device, 103 ... Execution program, 104 ... Load monitoring program, 105 ... Trace collection program, 106 ... Trace collection management table, 107 ... Trace collection range setting table, 108 ...
Trace data collection area, 109 ... System bus, 1
12 ... Auxiliary storage device, 114 ... Trace importance number, 1
15 ... Trace collection presence / absence flag.

Front page continuation (72) Inventor Akihiko Tobe 52-1 Omika-cho, Hitachi City, Ibaraki Prefecture Hitachi Ltd. Omika Plant, Ltd. (72) Toshio Ishihara 5-2-1 Omika-cho, Hitachi City, Ibaraki Hiritsu Process Computer Engineering Co., Ltd. (72) Inventor Shigeru Ogawa 5-2-1 Omika-cho, Hitachi City, Ibaraki Hitachi Ltd. Omika Factory, Hitachi Ltd. (72) Inventor Yuji Imaizumi 5-chome, Omika-cho, Hitachi City, Ibaraki Prefecture No. 2-1, Hiritsu Process Computer Engineering Co., Ltd. (72) Inventor Yoshinori Sato, 5-2-1, Omika-cho, Hitachi City, Hitachi, Ibaraki (72) Inventor Tadashi Kojima Hitachi City, Ibaraki Prefecture 5-2-1 Omika-cho, Hitachi Process Computer Engineering Co., Ltd. (72) Inventor Shinichi Funa 5-2-1, Omika-cho, Hitachi City, Ibaraki Prefecture Russia Seth Computer Engineering Co., Ltd. in

Claims (6)

[Claims] 1. A trace collection method for collecting and storing a trace which is a program operation history, while monitoring a CPU load amount or a transaction amount, and increasing or decreasing the trace amount to be collected stepwise in response to a change in the amount. As described above, the trace collection range is set in advance, and the trace collection process called from an arbitrary module of the program being executed refers to the CPU load amount or transaction amount at the time and the trace collection range, and collects the trace of the module. A trace collection method characterized by determining whether or not is necessary and collecting a trace when it is necessary. 2. The trace collection range according to claim 1, wherein the trace collection range is set to a wide range from a low importance level to a high importance level according to a plurality of importance levels according to purposes, and the trace collection range is set to the CPU load amount or A trace collection method characterized in that a trace collection range to be referred to is selected according to the degree of importance corresponding to the transaction amount in stages. 3. The trace collection method according to claim 1, wherein the trace collection range registers the name of a trace collection calling module for each of the importance levels. 4. A trace collection method for collecting and storing a trace, which is a program operation history in a computer system, in advance, according to a plurality of importance levels according to purposes, a large number of module names having a large trace amount at a low importance level. Then, set the trace collection range with a small number of module names that reduce the trace amount for high importance, and monitor the CPU load or transaction amount in parallel while processing an arbitrary executable program consisting of multiple modules. , When the trace processing program is called from the module, the trace collection range of the corresponding importance corresponding to the CPU load or the transaction amount is referred to, and the trace is collected only when the module name is registered. A trace collection method characterized by performing collection. 5. A computer system comprising a trace collection means for collecting and storing a trace of a program operation history by a call from a running program, and a CPU load monitoring means. A wide range in which the trace amount increases in a high degree, and a narrow range in which the trace amount decreases in a high degree of importance is stored in a stepwise manner, and a trace collection range storage unit is provided. A computer system comprising a switching means for changing the trace collection range for each importance level. 6. The trace collection range storage unit according to claim 5, wherein the trace collection unit storage unit is configured to be registered with a module name that calls the trace collection unit, and the trace collection unit is changed by the switching unit. A computer system configured to determine whether or not a calling module name of the program being executed is registered with reference to the trace collection range, and collect the trace if the name is registered.