CN109542742A - Database server hardware health evaluating method based on expert model - Google Patents
Database server hardware health evaluating method based on expert model Download PDFInfo
- Publication number
- CN109542742A CN109542742A CN201811352658.3A CN201811352658A CN109542742A CN 109542742 A CN109542742 A CN 109542742A CN 201811352658 A CN201811352658 A CN 201811352658A CN 109542742 A CN109542742 A CN 109542742A
- Authority
- CN
- China
- Prior art keywords
- index
- hardware
- expert model
- database server
- database
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3447—Performance evaluation by modeling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3466—Performance evaluation by tracing or monitoring
- G06F11/3476—Data logging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2201/00—Indexing scheme relating to error detection, to error correction, and to monitoring
- G06F2201/80—Database-specific techniques
Abstract
The database server hardware health evaluating method based on expert model that the invention discloses a kind of, the appraisal procedure includes acquisition operating system hardware index, hardware index is divided into index in expert model, for Distribution Indexes fractional weight in expert model, and multilevel threshold and standard of deducting point are formulated, index value falls in different threshold ranges and deducts corresponding score；Automated periodic acquisition hardware index；Database server hardware index, the healthy score of output database server hardware are calculated according to expert model rule.Above-mentioned appraisal procedure has chosen the key index of DATABASE HARDWARE, solves selecting index integrity problem；Indicator combination is matched, DATABASE HARDWARE health model is constructed, by assessing unicity problem for Distribution Indexes weight in model, and setting dynamic threshold, solution；It is calculated by automatic collection and model, DATABASE HARDWARE health degree is assessed, solves the problems, such as that manual evaluation is at high cost.
Description
Technical field
The present invention relates to the O&M fields of database, hard more particularly to a kind of database server based on expert model
Part health evaluating method.
Background technique
Database operation problem is on the one hand from database software itself, and there are also greatly the reason is that being derived from data
The hardware problem in library.In analytical database hardware problem, usually by DBA by referring to manually or using tool acquisition hardware key
Mark, assesses hardware health degree in conjunction with experience, and according to phenomenon combination index analysis problem.
Although large database monitoring tools are very various at present, these usual tools only can periodically obtain some data
The key index of library hardware, the prefabricated threshold value of these indexs of tool generate some threshold values and accuse when the deterioration of database index
It is alert.By collecting day when operation maintenance personnel inspection, in week, the achievement data of the moon is analyzed after integration, manually to hardware health degree
It is assessed.
Since current operation and maintenance tools are collecting part key index, key index is shown in the value of sampled point or draws change
Change curve, generates alarm in index deterioration.It is more than that some is normal that the metrics-thresholds setting of common operation and maintenance tools, which is sometimes referred as scale value,
Amount generates alarm, and rule is too simple, can not cope with complicated dynamic threshold.
The variation of the health degree of DATABASE HARDWARE is as caused by multi objective collective effect, and the variation of single index can not
The size influenced on database is embodied, common operation and maintenance tools lack the energy that systemic assessment is carried out to DATABASE HARDWARE health degree
Power has only served the effect of index collection device, and DBA is finally still needed to carry out manual analysis.
Therefore, it is necessary to a kind of new technical solutions to solve the above technical problems.
Summary of the invention
To solve the deficiencies in the prior art, the present invention proposes a kind of database server hardware health based on expert model
Appraisal procedure systemic can must assess database server hardware, and assessment result is accurate, high reliablity.
In order to achieve the above objectives, the present invention adopts the following technical scheme that: a kind of database clothes based on expert model
Business device hardware health evaluating method, which comprises
Establish expert model: hardware index is divided into index in expert model by acquisition operating system hardware index, is expert
Distribution Indexes fractional weight in model, and multilevel threshold and standard of deducting point are formulated, index value falls in different threshold ranges and deducts
Corresponding score；
Expert model is applied to database server hardware objects, automated periodic acquisition hardware index；
Database server hardware index is calculated according to expert model rule, realizes automatic health assessment, output database
The healthy score of server hardware.
In a specific embodiment, operating system provides basic expert model, and user can be according to operating system
Feature carries out duplication to basic expert model and parameter regulation establishes expert model；
It in a specific embodiment, further include writing hardware index value after the periodical acquisition hardware index
Enter database, periodically acquires the index of typing in nearest a period of time in database, be calculated according to expert model rule
Healthy score.
In a specific embodiment, constant value and expression formula are supported in the threshold value setting of the index.
In a specific embodiment, the expert model appraisal procedure includes:
Step S11: achievement data in assessment cycle is obtained；
Step S12: pointer type in expert model is obtained；
Step S13: judging whether it is original value, if it is not, entering step S14 ': calculating average value or difference, enters back into step
Rapid S14: the index value that will be obtained, metrics-thresholds input rule engine in expert model；If so, being directly entered step S14；
Step S15: whether judgment threshold is constant, if it is not, entering step S16 ': expression solution obtains constant threshold,
It enters step S16: seeing if fall out threshold value；If the judging result of step S15 be it is yes, be directly entered step S16；If step
The judging result of S16 be it is yes, then enter step S17: deducting the corresponding score of threshold levels, enter step S18: expert model is total
Point subtracting deduction of points obtains hardware health score；If the judging result of step S16 be it is no, be directly entered step S18.
In a specific embodiment, the assessment cycle is greater than collection period.
It in a specific embodiment, include 2 samplings in the assessment cycle, assessment cycle is 2 minutes, sampling week
Phase is 30 seconds.
In a specific embodiment, the total score calculation of the expert model is the sum of 100- index deduction of points, is referred to
When the sum of mark deduction of points is more than 100, expert model total score is handled according to 0 point.
In a specific embodiment, the index includes CPU usage, physical memory surplus, magnetic disc i/o read-write
Delay, CPU iowait percentage, business and operating system kernel catalogue utilization rate, network interface card packet loss mistake packet number, network card status,
CPU core number, operating system process number.
In a specific embodiment, the pointer type includes:
Original value: the index value of the last sampled point in assessment cycle is taken；
Average value: the sum of index value of multiple sampled points/number of sampling points in assessment cycle；
Difference: the last sampled point index value-earliest sampled point index value in assessment cycle.
Above-mentioned appraisal procedure proposed by the present invention has chosen the key index of DATABASE HARDWARE, and it is reliable to solve selecting index
Property problem.Indicator combination is matched, DATABASE HARDWARE health model is constructed, by for Distribution Indexes weight, Yi Jishe in model
Dynamic threshold is set, assessment unicity problem is solved.It is calculated by automatic collection and model, DATABASE HARDWARE health degree is commented
Estimate, solves the problems, such as that manual evaluation is at high cost.Meanwhile load rise occurs in database, performance deteriorates under scene, passes through score
Form can intuitively show the health status of server hardware, provide adequately reference to operation maintenance personnel, avoid production system
Economic loss brought by failure.
Detailed description of the invention
Fig. 1 is the scheme main-process stream of the database server hardware health evaluating method of the invention based on expert model
Figure；
Fig. 2 is the model evaluation process of the database server hardware health evaluating method of the invention based on expert model
Figure.
Specific embodiment
The preferred embodiment of apparatus of the present invention and method is described in further detail with reference to the accompanying drawing.
One, overall plan
The invention proposes a kind of the database server hardware health evaluating method based on expert model, the appraisal procedure
Including acquisition operating system hardware index, including from CPU, memory, I/O, file system, network carries out selecting index, by hardware
Index is divided into index and assistant analysis index in expert model；
It formulates expert model rule: for Distribution Indexes fractional weight in expert module, and formulating multilevel threshold and deduction of points
Standard, index value fall in different threshold ranges and deduct corresponding score；
Expert model is applied to database server hardware objects, operating system refers to automated periodic acquisition hardware
Mark；
Database server hardware index is calculated according to expert model rule, realizes automatic health assessment, output database
The healthy score of server hardware.
In a specific embodiment, operating system provides basic expert model, and user can be according to operating system
Feature carries out duplication to basic expert model and parameter regulation establishes expert model.
As shown in Figure 1, the net assessment process of the appraisal procedure includes: periodical acquisition hardware index, by hardware index
Value write-in database, periodically acquires the index of typing in nearest a period of time in database, calculates according to expert model rule
Obtain healthy score.For example, hardware index is stored in analytical database, it is multiple out of in analytical database extraction assessment cycle to adopt
The hardware achievement data of sample time is calculated according to new expert model rule, obtains the health of database server hardware
The step of score.
As shown in Fig. 2, expert model appraisal procedure process specifically includes:
Step S11: achievement data in assessment cycle is obtained；
Step S12: pointer type in expert model is obtained；
Step S13: judging whether it is original value, if it is not, entering step S14 ': calculating average value or difference, enters back into step
Rapid S14: the index value that will be obtained, metrics-thresholds input rule engine in expert model；If so, being directly entered step S14；
Step S15: whether judgment threshold is constant, if it is not, entering step S16 ': expression solution obtains constant threshold,
It enters step S16: seeing if fall out threshold value；If the judging result of step S15 be it is yes, be directly entered step S16；If step
The judging result of S16 be it is yes, then enter step S17: deducting the corresponding score of threshold levels, enter step S18: expert model is total
Point subtracting deduction of points obtains hardware health score；If the judging result of step S16 be it is no, be directly entered step S18.
In a specific embodiment, assessment cycle is greater than collection period.For example, being adopted in assessment cycle including 2 times
Sample, assessment cycle are 2 minutes, and the sampling period is 30 seconds.
Two, initial data is acquired
In conjunction with database server O&M practice summary of the invention, analysis, which has obtained, will affect database operating index,
It is as follows to obtain index from hardware device where database at present: CPU usage, physical memory surplus, magnetic disc i/o read-write are prolonged
Late, CPU iowait percentage, business and operating system kernel catalogue utilization rate, network interface card packet loss mistake packet number, network card status, CPU
Nucleus number, operating system process number.
Three, health model
1. at least being needed in assessment cycle comprising double sampling, the assessment cycle of recommendation is 2 minutes, the sampling period 30
Second.
2. health model total score 100 is divided.
3. model index, type and index score are as follows:
A.CPU utilization rate, type: average value, score: 20
B. physical memory surplus, type: original value, score: 15
C. magnetic disc i/o read-write delay, type: average value, score: 15
D.CPU iowaitb percentage, type: original value, score: 15
E. business and operating system kernel catalogue utilization rate, type: original value, score: 15
F. network interface card packet loss mistake packet number, type: difference, score: 20
4. pointer type calculation specifications
Original value: the index value of the last sampled point in assessment cycle is taken
Average value: the sum of index value of multiple sampled points/number of sampling points in assessment cycle
Difference: the last sampled point index value-earliest sampled point index value in assessment cycle
5. metrics-thresholds and deduction of points
The setting of metrics-thresholds can support constant value, be also possible to complex rule, for example, regular expression.Threshold value point
At multiple ranks, the corresponding trigger condition of different stage is different with deduction of points, and the metrics-thresholds and deduction of points of this programme are provided that
A.CPU utilization rate
Level-one: > 80%, deduction of points: 10
Second level: > 90%, it sets up when operating system process number is more than 2 times of CPU core number, otherwise drops
For level-one, deduction of points: 20
Three-level: > 99%, it is more than 3 times of Shi Chengli of CPU core number and if only if operating system process number, is otherwise reduced to two
Grade, deduction of points: 100
B. physical memory surplus
Level-one: < 2G, deduction of points: 10
Second level: < 1G, deduction of points: 20
Three-level: < 500M, deduction of points: 100
C. magnetic disc i/o read-write delay
Level-one: > 15ms, deduction of points: 10
Second level: > 30ms, deduction of points: 15
D.iowait percentage
Level-one: > 30%, deduction of points 10
Second level: > 40%, deduction of points 15
E. with operating system kernel catalogue utilization rate
Level-one: > 80%, deduction of points 10
Second level: > 90%, deduction of points 15
Three-level :=100%, deduction of points 100
F. network interface card packet loss mistake packet number
Level-one: number of sampling points in >=10* assessment cycle, deduction of points 20
Second level: the non-UP of network card status, deduction of points 100
6. health model total score is that calculation is the sum of 100- index deduction of points, when the sum of index deduction of points is more than 100, model
Total score is handled according to 0 point.
7. model total score falls in [90,100], it is believed that database server hardware is more healthy, [75,90) be it is general,
[60,75) need to carry out carry out early warning, [0,60) such case typicallys represent DATABASE HARDWARE there are failures.
The basic principles, main features and advantages of the invention have been shown and described above.The technical staff of the industry should
Understand, the invention is not limited in any way for examples detailed above, all skills obtained by the way of equivalent substitution or equivalent transformation
Art scheme, all falls within protection scope of the present invention.
Claims (10)
1. a kind of database server hardware health evaluating method based on expert model, which is characterized in that the described method includes:
Establish expert model: hardware index is divided into index in expert model by acquisition operating system hardware index, is expert model
Interior Distribution Indexes fractional weight, and multilevel threshold and standard of deducting point are formulated, index value falls in different threshold ranges and deducts accordingly
The score of grade；
Expert model is applied to database server hardware objects, automated periodic acquisition hardware index；
Database server hardware index is calculated according to expert model rule, realizes automatic health assessment, output database service
The healthy score of device hardware.
2. database server hardware health evaluating method according to claim 1, which is characterized in that operating system provides
Basic expert model, user can carry out duplication to basic expert model and parameter regulation built according to operating system feature
Vertical expert model.
3. database server hardware health evaluating method according to claim 1, which is characterized in that the periodicity is adopted
It further include that database is written into hardware index value, is periodically acquired in database in nearest a period of time after collection hardware index
Healthy score is calculated according to expert model rule in the index of typing.
4. database server hardware health evaluating method according to claim 1, which is characterized in that the threshold of the index
Constant value and expression formula are supported in value setting.
5. database server hardware health evaluating method according to claim 1, which is characterized in that the expert model
Appraisal procedure includes:
Step S11: achievement data in assessment cycle is obtained；
Step S12: pointer type in expert model is obtained；
Step S13: judging whether it is original value, if it is not, entering step S14 ': calculating average value or difference, enters back into step
S14: the index value that will be obtained, metrics-thresholds input rule engine in expert model；If so, being directly entered step S14；
Step S15: whether judgment threshold is constant, if it is not, entering step S16 ': expression solution obtains constant threshold, enters
Step S16: threshold value is seen if fall out；If the judging result of step S15 be it is yes, be directly entered step S16；If step S16
Judging result be it is yes, then enter step S17: deducting the corresponding score of threshold levels, enter step S18: expert model total score
It subtracts deduction of points and obtains hardware health score；If the judging result of step S16 be it is no, be directly entered step S18.
6. database server hardware health evaluating method according to claim 5, which is characterized in that the assessment cycle
It is greater than collection period.
7. database server hardware health evaluating method according to claim 6, which is characterized in that the assessment cycle
It inside include 2 samplings, assessment cycle is 2 minutes, and the sampling period is 30 seconds.
8. database server hardware health evaluating method according to claim 1, which is characterized in that the expert model
Total score calculation be 100- index deduction of points the sum of, index deduction of points the sum of more than 100 when, expert model total score exists respectively according to 0
Reason.
9. database server hardware health evaluating method according to claim 1, which is characterized in that the index includes
CPU usage, physical memory surplus, magnetic disc i/o read-write delay, CPU iowait percentage, business and operating system kernel
Catalogue utilization rate, network interface card packet loss mistake packet number, network card status, CPU core number, operating system process number.
10. database server hardware health evaluating method according to claim 1, which is characterized in that the index class
Type includes:
Original value: the index value of the last sampled point in assessment cycle is taken；
Average value: the sum of index value of multiple sampled points/number of sampling points in assessment cycle；
Difference: the last sampled point index value-earliest sampled point index value in assessment cycle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352658.3A CN109542742A (en) | 2018-11-14 | 2018-11-14 | Database server hardware health evaluating method based on expert model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352658.3A CN109542742A (en) | 2018-11-14 | 2018-11-14 | Database server hardware health evaluating method based on expert model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109542742A true CN109542742A (en) | 2019-03-29 |
Family
ID=65847202
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811352658.3A Pending CN109542742A (en) | 2018-11-14 | 2018-11-14 | Database server hardware health evaluating method based on expert model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109542742A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111581045A (en) * | 2020-03-18 | 2020-08-25 | 平安科技（深圳）有限公司 | Database anomaly monitoring method and device, computer device and storage medium |
| CN112596991A (en) * | 2020-12-27 | 2021-04-02 | 卡斯柯信号有限公司 | Hot standby reverse cutting method based on machine health state |
| CN112890816A (en) * | 2020-12-11 | 2021-06-04 | 万达信息股份有限公司 | Health index scoring method and device for individual user |
| CN113094245A (en) * | 2021-03-26 | 2021-07-09 | 四川新网银行股份有限公司 | Method for measuring health of database cluster |
| CN115794590A (en) * | 2023-01-30 | 2023-03-14 | 麒麟软件有限公司 | Health assessment method and device for domestic Linux operating system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139007A (en) * | 2011-12-05 | 2013-06-05 | 阿里巴巴集团控股有限公司 | Method and system for detecting application server performance |
| CN104573984A (en) * | 2015-01-27 | 2015-04-29 | 国网四川省电力公司电力科学研究院 | B/S information system state partitioning method based on healthy tree |
| CN106411609A (en) * | 2016-11-08 | 2017-02-15 | 上海新炬网络信息技术有限公司 | IT software and hardware running state monitoring system |
| CN106776214A (en) * | 2016-12-12 | 2017-05-31 | 广州市申迪计算机系统有限公司 | A kind of server health degree appraisal procedure |
-
2018
- 2018-11-14 CN CN201811352658.3A patent/CN109542742A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139007A (en) * | 2011-12-05 | 2013-06-05 | 阿里巴巴集团控股有限公司 | Method and system for detecting application server performance |
| CN104573984A (en) * | 2015-01-27 | 2015-04-29 | 国网四川省电力公司电力科学研究院 | B/S information system state partitioning method based on healthy tree |
| CN106411609A (en) * | 2016-11-08 | 2017-02-15 | 上海新炬网络信息技术有限公司 | IT software and hardware running state monitoring system |
| CN106776214A (en) * | 2016-12-12 | 2017-05-31 | 广州市申迪计算机系统有限公司 | A kind of server health degree appraisal procedure |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111581045A (en) * | 2020-03-18 | 2020-08-25 | 平安科技（深圳）有限公司 | Database anomaly monitoring method and device, computer device and storage medium |
| CN112890816A (en) * | 2020-12-11 | 2021-06-04 | 万达信息股份有限公司 | Health index scoring method and device for individual user |
| CN112596991A (en) * | 2020-12-27 | 2021-04-02 | 卡斯柯信号有限公司 | Hot standby reverse cutting method based on machine health state |
| CN112596991B (en) * | 2020-12-27 | 2023-09-08 | 卡斯柯信号有限公司 | Hot standby reverse cutting method based on machine health state |
| CN113094245A (en) * | 2021-03-26 | 2021-07-09 | 四川新网银行股份有限公司 | Method for measuring health of database cluster |
| CN113094245B (en) * | 2021-03-26 | 2023-06-06 | 四川新网银行股份有限公司 | Method for measuring health of database cluster |
| CN115794590A (en) * | 2023-01-30 | 2023-03-14 | 麒麟软件有限公司 | Health assessment method and device for domestic Linux operating system |
| CN115794590B (en) * | 2023-01-30 | 2023-10-31 | 麒麟软件有限公司 | Health assessment method and device for domestic Linux operating system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109542742A (en) | Database server hardware health evaluating method based on expert model | |
| CN106154209B (en) | Electrical energy meter fault prediction technique based on decision Tree algorithms | |
| CN105588995B (en) | A kind of automatic power-measuring system line loss method for detecting abnormality | |
| CN102692615B (en) | System capable of automatically acquiring electric quantity data | |
| CN106570778A (en) | Big data-based data integration and line loss analysis and calculation method | |
| CN112946483B (en) | Comprehensive evaluation method for battery health of electric vehicle and storage medium | |
| CN110119948B (en) | Power consumer credit evaluation method and system based on time-varying weight dynamic combination | |
| CN109816031A (en) | A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data | |
| CN109472293A (en) | A kind of grid equipment file data error correction method based on machine learning | |
| US20140379303A1 (en) | Identifying utility resource diversion | |
| CN110927654A (en) | Batch running state evaluation method for intelligent electric energy meters | |
| CN110968703B (en) | Method and system for constructing abnormal metering point knowledge base based on LSTM end-to-end extraction algorithm | |
| CN113902241A (en) | Power grid equipment maintenance strategy system and method based on comprehensive state evaluation | |
| CN114290960A (en) | Method and device for acquiring battery health degree of power battery and vehicle | |
| CN106682383B (en) | To the accurate statistical processing methods of acquisition tables code value in a kind of metering system | |
| CN110738415A (en) | Electricity stealing user analysis method based on electricity utilization acquisition system and outlier algorithm | |
| CN109064211B (en) | Marketing business data analysis method and device and server | |
| CN112345972A (en) | Power failure event-based power distribution network line transformation relation abnormity diagnosis method, device and system | |
| CN114665610B (en) | Capacitor monitoring method, system and equipment based on reactive power acquisition | |
| CN108596450B (en) | Power grid risk early warning method and system | |
| CN110866696A (en) | Method and device for training shop falling risk assessment model | |
| CN111080364A (en) | Engineering cost data center management system and device | |
| CN205643673U (en) | Metering device scraps alarm device based on measurement instrument follow -up of quality evaluation system | |
| CN114389962A (en) | Broadband loss user determination method and device, electronic equipment and storage medium | |
| JP2010102462A (en) | Apparatus, method and program for estimating trouble |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190329 |