JPH08147194A - Data logging method - Google Patents

Abstract

PURPOSE: To improve the utilization efficiency of an entire logging buffer area so as to remain even the data of low logging frequency by dynamically allocating the logging areas corresponding to the logging frequency. CONSTITUTION: Logging areas A-C are provided with counters Ac-Cc showing a logging amount per unit time and data areas Ax-Cx showing the number of logging blocks, and logging blocks Al-Ck are provided with pointers Ap-Cp showing data positions to store logging data and link terminals At-Ct showing the order of connection of blocks. The same logging block size length (z) is defined regardlessly of the kinds A-C of logging data and concerning the dynamic allocation of logging areas corresponding to the logging frequency, the logging amounts of the respective counters Ac-Cc are investigated. Then, when there is the logging area C the frequency of which is lower than the logging amount of the relevant logging area A and this logging block number Cx is more than a minimum block number, the fixed number of blocks from the logging area C are moved/connected to the relevant logging area A.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to improvement of a data logging method in a computer system.

[0002]

2. Description of the Related Art Logging means recording an operation record of a computer system. In order for an administrator to grasp the operation status of the system or to facilitate the maintenance work of the system, the usage status of system resources, The operating state of the system is recorded according to the time transition. Generally, the logging data records a user name, a used resource name, a used time, a processing content and a result thereof, a failure content of the device, an occurrence time and the like. As for the failure content, error information or trace information for finely tracing an unexpected error is recorded according to the importance of the error information and the error pattern.

FIG. 7 is an explanatory view for explaining allocation of a logging buffer area for storing logging data in the prior art. In FIG. 7, the buffer area 1 indicated by the logging buffer size length L is a logging area including logging portions a1 to ai for storing logging data for A with respect to different types of logging data A to C in the illustrated example. Logging unit b1 that stores logging data for A and B
a logging area B composed of bj, a logging area C composed of logging units c1 to ck for storing logging data for C,
When logging data, the data is cyclically logged in the logging areas A to C that are statically divided for each of the various logging data (A to C) described above.

[0004]

However, in the conventional method, since the data is cyclically logged, the logging data is overwritten based on the respective logging frequency after a certain period of time has elapsed. Cause logging data loss. Also, because each logging area is statically allocated,
If you want to retain more logging data,
When allocating a logging area in advance at the time of system design, etc., only the allocation area is increased to the logging area expected to have a high logging frequency. Therefore, the logging area could not be efficiently used for logging in which the logging load dynamically fluctuates.

As another method, it is possible to cyclically log the entire area without dividing the logging area for each logging data. According to this method, a large amount of data with a high logging frequency can be left, but the control of leaving a data with a low logging frequency cannot be achieved. This is because if the data with high logging frequency cyclically uses the entire area, other logging data will be overwritten and disappear when the data with high logging frequency makes a round.

The present invention has been made in view of the above points, and an object thereof is to solve the above-mentioned problems and to provide a system having means for cyclically logging a plurality of different logging data in one logging buffer area. Even if each logging frequency dynamically changes, the logging area is allocated according to the logging frequency to improve the utilization efficiency of the entire logging buffer area, and to save the data with low logging frequency to the logging buffer area to some extent. It is to provide a data logging method that can be left.

[0007]

In order to achieve the above object, the present invention provides a logging area comprising a plurality of logging blocks for storing a plurality of logging data,
In the data logging method, one logging buffer area is configured for each type of logging data, and the logging data is stored cyclically in the logging area each time the logging data occurs. And a data area indicating the number of logging blocks included in the logging area, and the logging block includes a pointer including a block number indicating the data position where the current logging data is stored and a block serial number. A link terminal that indicates the connection order of blocks in the logging area is provided, and the same logging block size length is configured regardless of the type of logging data.

[0008]

With the above-mentioned configuration means, when the logging data is generated, the data is logged in the last logging block of the logging area, and then the first logging block is returned,
Check the logging amount per unit time of the counter of each logging area, and there is a logging area whose frequency is lower than the logging amount of the relevant logging area, and the number of logging blocks of this logging area is less than the predetermined minimum number of logging blocks. When it is also large, it is judged that the logging block can be stripped from another logging area, and a certain number of logging blocks that are set in advance from the infrequent logging area are moved / connected to the logging area and the logging frequency is changed according to the logging frequency. The logging area can be dynamically allocated with a specific size.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a flow chart for dynamically allocating a logging area by a data logging method according to an embodiment of the present invention, and FIG. 2 is an explanatory view for explaining a logging area allocated for each logging data according to an embodiment. FIG. 3 is an explanatory diagram of logging data generated when the system is started up, and the data is stored in the corresponding logging block. FIG. 4 is an explanatory diagram of storing data in the same logging block.
Is an explanatory diagram of storing data in the next logging block, and FIG. 6 shows dynamic allocation processing from another logging area to the logging area when returning to the first logging block after logging data in the last logging block. It is an explanatory view explaining.

To simplify the explanation, the structure of a logging system according to the present data logging method will be described first with reference to FIGS. 2 and 3. In FIG. 2A, a logging buffer area 1 for storing logging data shows an overall image of a logging buffer having a size length L. In the illustrated example, there are n logging blocks (z1 to zn) having a size length z. To divide.
For simplification, let us say that the types of logging data stored in this logging buffer area 1 are A to C.
Logging block as logging area A for data A
Display (z1 to zi) as (A1 to Ai), and set logging block (zi + 1 to zi + j) as (B1 to B1 as logging area B for data B).
~ Bj), logging blocks (zi + j + 1 to zi + j + k = zn) are displayed as (C1 to Ck) as the logging area C for data C, and allocation is performed. FIG. 2B shows the top logging blocks A1 to C1 of the respective logging areas A to C. In the illustrated example, the logging blocks a1 to a3 for storing three logging data are included in the logging block A1. Logging unit b1 to b2 that stores two logging data in B1
However, the logging block C1 is configured with logging units c1 to c4 that store four logging data. The size length of each logging unit is as, bs, cs, and the size length z of the logging block is an integer multiple of as, bs, cs, z = 3a.
s = 2bs = 4cs.

FIG. 3 shows components when logging areas are assigned to the types A to C of logging data when the system is started up. The allocation element is a logging block unit of size length z shown in FIG. 2, and Ai, Bj, and Ck are allocated for logging data types A, B, and C, and in the illustrated example, size length z is assigned. A link terminal is added to the beginning of each logging block that it has, and a link is formed by connecting the link terminals At, Bt, and Ct. Logging is performed in the connection order of this link, and when the final logging block (link connection number is NULL) is reached,
Return to the top logging block and continue logging cyclically.

Further, each data link is provided with pointers Ap, Bp, Cp indicating the current logging block, and by moving the pointer prior to logging, the pointer is updated to the latest logging position. Pointers Ap, Bp, Cp
Strictly has a logging block number and a block serial number. When logging is performed, the block serial number is incremented, and when the block serial number is full, the serial number is set to 0 and the logging block number is updated.

Further, Ac, Bc, and Cc are counters indicating the logging amount per unit time for each logging area unit,
This logging counter is incremented each time one logging block is logged, and this count value is cleared every unit time. Further, Ax, Bx, and Cx indicate the number of logging blocks in the logging areas 2 to 4 for each type A to C of each logging data.

Next, the contents of the processing in this data logging method will be described with reference to FIGS. 4 to 6 and FIG. Hereinafter, the processing will be described with respect to logging, assuming that a logging request has occurred in the logging data A. In FIG.
First, the pointer Ap is updated to obtain the next logging position. In the illustrated example, the pointer Ap at the latest past logging position occurring in the logging data A is A2, a2-2. A2 indicates the logging block A2 for the logging data A, and a2-2 indicates the logging part of the logging block A2 (the block serial number is 2).
It means -2) in the second meaning. In updating the pointer Ap, first, the block serial number is incremented, so a2-2 is changed to a2-3, and data is logged in the a2-3 area indicated by the dotted arrow.

Next, in FIG. 5, the pointer Ap is A2, a2-3.
Then, the update of the pointer Ap at this time is because a2-3 is the last in the block, so follow the link of the logging block, set A2 to A3, set a2-3 to a3-1, and set the start position of the next logging block to Set as the logging position. In this case, since the logging block is updated, the logging amount counter Ac per unit time is incremented to 100 to 101.

Next, in FIG. 6, the pointer Ap is Ai, ai-3.
And Since the position of this logging block is the end of the logging area A for the logging data A, either return to the head logging block A1, a1-1 indicated by the terminal At, or log from another logging area (B or C). A block is stripped off, and it is determined whether or not to dynamically connect to this logging area A.

The determination in this case is performed as follows. That is, comparing the logging amount counter values Bc, Cc per unit time of the other logging data B and C with the logging amount counter value Ac per unit time of the main logging data A, there is a logging area smaller than the logging area A. Find out whether or not.
When there is a logging area (C in the illustrated example) with a small logging amount counter value per unit time, it is predetermined from the logging block (Ch + 1) next to the pointer Ch of the logging area C in the searched order. A certain number of logging blocks, for example, 10 logging blocks in the illustrated example, are stripped off, and the logging area A is moved and connected. If the number of logging blocks to be stripped is less than 10 by the final block, the logging blocks are returned to the first logging block and the remaining logging blocks are stripped. If the total number of blocks of the logging area C to be stripped is equal to or less than a predetermined constant value, this stripping process is not performed.

For example, assuming that Ac = 100, Bc = 120, Cc = 60, 10 blocks are stripped from the logging area C of Cc, and the last block of the logging area A is connected / connected. Then, the pointer Ap of the logging area A is updated to Ai + 1, a (i + 1) -1, and logging is performed in the newly connected logging block. The number of logging blocks Ax at this time is from i to i + 10
Then, Cx is changed from k to k-10.

By the above processing, the number of logging blocks allocated to the logging area A increases, the number of latest logging blocks increases as the logging frequency increases, and a large amount of logging data can be left in the logging buffer area 1. . On the other hand, on the side of the logging area C where the logging frequency is low, the number of logging blocks has decreased, but there are logging blocks of a certain value or more, so a certain amount of logging data can be left within that range.

If Ac = 100, Bc = 120, Cc = 140, the logging area A does not strip the logging block from the logging areas B and C, but returns to the head block indicated by At and performs logging. To do. FIG. 1 shows a flowchart of the entire logging process described above. In FIG. 1, when the logging data is generated in step S1,
Update the block serial number of the pointer of the logging area. At this time, in step S2, it is checked whether the serial number in the block of the pointer to be updated is the last in the block.If No, the serial number is not the final one, so the process jumps to step S12 to write the data to the updated logging block position Logging. In step S2, if the in-block serial number is the final Yes, the process proceeds to step S3 to check whether this block is the final, and if No, jumps to step S9 to update the block number of the pointer. In step S3, if this block is the final Yes, the process proceeds to step S4.

This step S4 and subsequent steps correspond to the dynamic logging area allocation method described above with reference to FIG. Step
In S4, the counter value of each logging area is checked to see if there is a logging area whose frequency is lower than the logging amount of the logging area. In step S4, if the infrequent logging area exists Yes, the process proceeds to step S5, and it is checked whether the number of logging blocks in this infrequent logging area is larger than a predetermined minimum number of logging blocks.

In step S5, if the number of logging blocks is larger than the minimum number of logging blocks, that is, Yes, the process moves to step S6, and the pointer indicating the logging block next to the current pointer of the infrequently logging area is extracted. , Go to step S7, take out a predetermined number of logging blocks from this infrequent logging area, go to step S8, and check that the relevant logging block is the last Yes in the logging area at step S3. Connect a certain number of logging blocks taken out,
Go to step S9, update the block number of the newly added logging block to update the pointer, update the counter value of the logging area in step S10,
In step S12, the relevant logging data is logged at the corresponding position of the newly added logging block.

Infrequent logging area exists in step S4
If No and the number of logging blocks is larger than the minimum number of logging blocks in step S5, the process proceeds to step S11, the pointer of the logging area is returned to the leading logging block, and the logging data is logged in step S12.

[0024]

As described above, according to the present invention, when logging different types of logging data in a limited logging buffer area, logging blocks are dynamically allocated to the corresponding logging area according to the logging amount. Since the usage of the logging buffer area per unit time increases and efficiency is improved, logging data with a small logging amount can be cyclically logged in the logging area according to the logging amount. Can be left within a certain amount.

[Brief description of drawings]

FIG. 1 is a flowchart for dynamically allocating a logging area by a data logging method according to an embodiment of the present invention.

FIG. 2 is an explanatory view illustrating a logging area allocated for each logging data according to an embodiment.

FIG. 3 is an explanatory diagram in which logging data is generated at system startup and the data is stored in the corresponding logging block.

FIG. 4 is an explanatory diagram for storing data in the same logging block.

FIG. 5 is an explanatory diagram for storing data in the next logging block.

FIG. 6 is an explanatory diagram illustrating a method of dynamically processing an update of a logging block when the block number of the logging counter points to the end.

FIG. 7 is an explanatory diagram illustrating allocation of a logging buffer area for storing logging data in the related art.

[Explanation of symbols]

1 Logging buffer area A to C Logging area A1 to Ai, B1 to Bj, C1 to Ck, z1 to zn, Ch to Ch + 10 Logging block a1 to a3, b1 to b2, c1 to c4, a2-1 to a3- 2, ai-3 Logging part L, z, as, bs, cs Size length Ac, Bc, Cc Counter value Ap, Bp, Cp Pointer At, Bt, Ct Terminal Ax, Bx, Cx Number of logging blocks

Claims (1)

[Claims] 1. A logging area comprising a plurality of logging blocks for storing a plurality of logging data, one logging buffer area being provided for each type of logging data, and each time the logging data is generated, the logging data is generated. In the data logging method of sequentially and cyclically storing data in the logging area, the logging area includes a counter indicating the amount of logging per unit time and a data area indicating the number of logging blocks in the logging area. The block is provided with a pointer consisting of a block number indicating the data position where the current logging data is stored and a block serial number, and a link terminal indicating the connection order of the blocks in the logging area, regardless of the type of logging data. Same logging block size length, When logging data occurs and the data is logged in the last logging block of the relevant logging area and then returns to the first logging block, the logging amount per unit time of the counter of each logging area is checked and the logging amount of the relevant logging area is checked. If there is a logging area of lower frequency and the number of logging blocks of this logging area is larger than a predetermined minimum number of logging blocks, a predetermined number of logging blocks from the low frequency of logging area The data logging method is characterized by migrating and connecting to the relevant logging area.