CN108121639A - A kind of data center's total management system design method based on cloud platform - Google Patents
A kind of data center's total management system design method based on cloud platform Download PDFInfo
- Publication number
- CN108121639A CN108121639A CN201711395908.7A CN201711395908A CN108121639A CN 108121639 A CN108121639 A CN 108121639A CN 201711395908 A CN201711395908 A CN 201711395908A CN 108121639 A CN108121639 A CN 108121639A
- Authority
- CN
- China
- Prior art keywords
- data
- monitoring
- management system
- data center
- total management
- 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
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3006—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/32—Monitoring with visual or acoustical indication of the functioning of the machine
- G06F11/323—Visualisation of programs or trace data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/182—Distributed file systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computer And Data Communications (AREA)
Abstract
Propose a kind of data center's total management system design method and data center's total management system based on cloud platform, data center's total management system has the function of automatic management, it is laid down a regulation according to historical statistical data, realize the operation such as being set automatic threshold, the data acquisition modes of monitoring agent are relatively abundanter, data acquisition modes are various including IT kind equipments and power ＆ environment supervision categorical data, so that data center's total management system is uniformly included under designed integrated service frame, construct one efficiently, stable data center's total management system.
Description
Technical field
The present invention relates to information technology field, more particularly to a kind of data center's total management system based on cloud platform is set
Meter method and data center's total management system.
Background technology
Modular data center (Module Data Center, MDC) is the New Generation of IDC portion based on cloud computing
Administration's form in order to tackle the trend of the servers such as cloud computing, virtualization, centralization, high densification development, uses modularized design
Theory reduces coupling of the infrastructure to building environment to the greatest extent.Be integrated with power supply and distribution, refrigeration, cabinet, air-flow containment,
The subsystems such as comprehensive wiring, power ＆ environment supervision, improve the whole efficiency of operation of data center, realize rapid deployment, resilient expansion and
Green energy conservation.
With the rapid development of big data information industry, the development of data center also enters a new stage.Management
System is the important component of configuration inside data center.Traditional management system mainly based on power ＆ environment supervision, possesses more
Kind of data-interface, can access UPS, power distribution cabinet, precision air conditioner, gate inhibition, Temperature Humidity Sensor, smoke detector, temperature detector,
A variety of monitored object such as leakage sensor, turning roof window and web camera.
The power of data center mainly include with environmental data center total management system UPS, power distribution cabinet, precision air conditioner,
Gate inhibition, Temperature Humidity Sensor, smoke detector, temperature detector, leakage sensor, turning roof window and web camera etc. are a variety of
Monitored object, abbreviation rotating ring data center total management system, core equipment are power ＆ environment supervision host.
Data center's total management system mainly includes information technoloy equipment management (server admin, virtual management), rotating ring number
According to the complete data center facility management such as center total management system management.
Currently, with the fast development of cloud computing, big data and internet, basic turn has occurred in information-based infrastructure
Become, the demand of monitoring management is converted into integral platform, the unified platform, the system being managed collectively from some individual system requirements
It is required that.Every application server is no longer individual computing module, but will be calculated, deposited by platforms such as cloud computing, big datas
Storage resource is united, and forms in large scale, unified monitoring and management resource pools across data center's scope, it is therefore desirable to energy
Enough uniform data center total management systems for monitoring extensive, the distributed, virtual resource of cross-region and physical resource.
Data center's infrastructure is the core of cloud computing framework, it is supplied to user to including CPU, memory, storage, net
The use of the computing resources such as network is effectively reduced the cost and complexity of IT O＆Ms.Cloud computing framework has distributed, inter-network
The characteristics of network, more resource category, brings unprecedented challenge, compared to traditional services for resource management aspect thereupon
Device aggregated structure, except the management to physical resources such as Web server, application servers, it is also necessary to CPU, memory, storage,
The unified management of the virtual resources such as network, virtual machine.
Effective management to resource and service is a core requirement in cloud computing delivery process.According to cloud computing framework
Distinguishing hierarchy can be divided into hardware platform management, virtual platform management, middleware management, application management etc., according to function pair
The difference of elephant can be divided into user management, storage management, network management, management of computing etc..
Currently, with the fast development of cloud computing technology, the quantity of cloud data center is also more and more.In cloud data center
There is large number of server, the very important composition portion of cloud platform is become to the monitoring of data center and operation condition of server
Point.Efficiently monitoring can ensure the stabilization of cloud platform in real time, improve the availability of cloud service.And traditional data center's synthesis
Management system is in the case of current cloud data center server is numerous, it is difficult to ensure that the real-time and high efficiency of monitoring.Therefore,
Study the server data center total management system of real-time, efficient, low occupancy still highly significant.
Not only include storage resource, computing resource and the Internet resources in server in cloud computing platform, be additionally included in cloud
Various software resources on platform and the more complicated resource under distributed environment.In addition, resource quantity is more, network environment
Complexity, monitoring software can not only generate substantial amounts of monitoring data, but also also need to be efficiently completed policer operation.This environment and
Under it is required that, traditional monitoring programme is difficult efficiently to carry out integrated management to data center server operating status.
The architecture of server data center total management system generally has centralization and two kinds of moulds of layer-stepping at present
Formula.
Centralized architecture is made of data center's total management system server and monitoring agent two parts, centralization
In data center's total management system, monitoring agent is distributed on the node that each needs monitors, and it is monitored to be responsible for acquisition
Then monitoring data is transmitted to data center's total management system server by certain mode, and connect by the monitoring data of resource
The control instruction from data center's total management system server is received, is operated accordingly；Integrated management system of data center
Server unite on a certain specific server, is responsible for obtaining the data that monitoring agent transmission comes, at the same it is responsible to data into
Row analysis, processing, storage and data display are responsible for carrying out dynamic configuration to monitoring agent.
There are problems that low-response when single point failure and more node in a centralized architecture, in order to make up this lack
, there is Layered Architecture in point, and in Layered Architecture, monitoring agent is divided into several groups with hierarchy,
There are several monitoring nodes in each group, there is a capability this group of things is handled.
Each group is equivalent to a centralized data center total management system environment, and the local node that monitors serves as data
The role of center total management system server, global monitoring node are responsible for monitoring each local monitoring node.This structure subtracts
Lack the situation that a large amount of monitoring datas are sent simultaneously to a server, reduce network load；Although the meeting inside a group
There is the problem of single point failure, but Fault Isolation is realized between group and group.
Centralized data center total management system relative hierarchical formula data center total management system is simple in structure, delay
It is smaller, it easily manages, is easy to dispose.But if server where data center's total management system server goes wrong,
Just it is present with the situation of the completely paralysed disease of data center's total management system；In addition when the node for needing to monitor is excessive, largely
Monitoring data is transmitted in a network, and system can be led to problems such as to respond, and slow, network occupancy is high, influences the fortune of distributed system
Line efficiency reduces the real-time of data center's total management system.
When node is more, although layered structure solves the problems, such as that the response speed of hubbed mode is slow, but is layered
Structure program is disposed and implements that difficulty is larger, and the mode of information is transferred in layering, and system delay is bigger, and real-time is bad.
In short, centralized monitoring architecture is deposited monitoring data due to intensive data center total management system server
Enter single node, it is possible that the problem of single point failure, and the transmission of substantial amounts of monitoring data may result in network and gather around
Plug；Layered Architecture solves the problems, such as single point failure, but due to layered structure, system needs the access for specifying node
Data are successively transferred, access efficiency is caused to reduce, and are disposed complex.
The present invention uses following technical term：
MDC modular data centers (Module Data Center, MDC)
SOA Enterprise SOAs (Service-Oriented Architecture)
API application programming interfaces
The content of the invention
The technical issues of in order to solve as above, the present invention propose a kind of integrated management system of data center based on cloud platform
System design method.Invention defines data center's total management system design methods based on cloud platform, define one point
Cloth service layer frame, defines monitoring agent module, defines data collector function module；Define data center's synthesis
Tomcat-AdminPortal functional method.
Wherein, the present invention proposes a kind of data center's total management system design method based on cloud platform, including such as
Lower step：
Step 1, the design of agreement connector：Agreement connector is divided into connector server end and connector client, often
A connector client all has the distal end view there are one connector server end, and connector server end receives from various companies
The connection request of device client is connect, and handles these requests；
Step 2, the design of monitoring agent, monitoring agent are the monitoring implementation sections of data center's total management system,
It is distributed in data center's monitoring host computer and server cluster, is responsible for acquisition, threshold test and the data-pushing of monitoring data, separately
It is outer that corresponding operating is carried out according to the instruction received；
Step 3, the design of data collector, data collector are responsible for collecting the information that a large amount of monitoring agents are transferred, and
Calling database purchase, data collector is distributed in the part of nodes of data center's total management system into cloud platform database
On, for gathering the monitoring information for the monitoring agent specified；
Step 4, the design of data center's total management system server, data center's total management system server is to be
Administrative center, monitoring center and the data processing centre of system, for performing following management function：Threshold value is crossed the border management, data
Processing management, system configuration management；
Step 5, Web server designs, the interface that Web server is user to be interacted with data center total management system,
Interface is provided for system configuration, it is necessary to complete Sign-On authentication, data display function.
Step 6, the design of data storage, data center's total management system carry out data storage using HBase databases,
For storing the monitoring parameter of data center's total management system, monitoring data and treated statistics.
Preferably, agreement connector is based on RMI protocol connector, and connector server end is responsible for establishing RMI
Connector servers, connector client are Web browser；
Preferably, the data that monitoring agent is gathered include but not limited to CPU real time loads, memory usage, current magnetic
Disk service condition, network utilization, disk read-write rate and system information；Monitoring agent is established not for different monitoring resources
With correspondent entity class preserve monitoring data, and data center's total management system server is allowed to obtain corresponding monitoring number
According to；
Preferably, when monitoring agent starts, two threads can be started：The data acquisition thread of server
The GatherThread and thread ServerThread for establishing JMX servers；After collecting thread starts, GatherThread is adopted
The related data of virtual machine or power ＆ environment supervision host where collection, and by the data got compared with threshold value, when there is threshold value
Cross the border happen or transmission mode for push when, monitoring agent sends JMX notification informations, informs the data being registered on itself
The current state status of collector oneself；
Preferably, data collector needs to collect the information of which monitoring agent node by data center's total management system
Server is specified, and when system starts, data center's total management system server can use JMX technologies by the monitoring agent of registration
Fifty-fifty distribute to data collector, data collector according to the list of the monitoring agent of acquisition, polling request monitoring agent
Data；
Preferably, data center's total management system server can with data collector timed communication, when there is data collection
When rock machine occurs in monitoring node where device, data center's total management system server, which can be withdrawn, is distributed to the data collector
Monitoring agent information, and these information are reassigned to active data collector；
Preferably, the threshold value performed by data center's total management system server cross the border management be divided into data center synthesis
Threshold value Preliminary detection is responsible at Tomcat-AdminPortal end and monitoring agent end two parts, monitoring agent end, and threshold value is crossed the border
Monitoring data is pushed to data center's total management system server；Data center's total management system server obtains monitoring letter
After breath, monitoring data is parsed, obtains the data item beyond threshold value, warning content of crossing the border is organized, and with web page notification or sends postal
Part mode notifies administrator and user in time；
Preferably, data center's total management system server needs to realize multiple corresponding functions using multiple threads：
CountThread threads are used for the timing statistical disposition of monitoring data；RegisterListener threads are used for monitoring agent
And data collector registration monitor, when there is notification message to send, the corresponding class that handles is handled；SeverThread lines
Journey is used to implement JMX server ends, and connecting interface is provided for Web server；
Preferably, data center's total management system reads data from database is accessed in units of monitoring agent
, monitoring data is divided into three kinds according to the difference of type：Real-time monitoring data, monitoring parameter and the prison after data processing
Control data.
Accordingly, the present invention also proposes a kind of data center's total management system based on cloud platform, realizes in cloud computing
In platform, which includes：
Agreement connector：Agreement connector includes connector server end and connector client, each connector client
End all tools are there are one the distal end view of connector server end, and the receiving of connector server end is from various connector clients
Connection request, and handle these requests；
At least one monitoring agent；Monitoring implementation section of the monitoring agent as data center's total management system, distribution
In data center's monitoring host computer and server cluster, it is responsible for acquisition, threshold test and the data-pushing of monitoring data, in addition root
Corresponding operating is carried out according to the instruction received；
At least one data collector is responsible for collecting the information that a large amount of monitoring agents are transferred, and calls database purchase
Into cloud platform database, at least one data collector is distributed in the part of nodes of data center's total management system
On, for gathering the monitoring information for the monitoring agent specified；
Data center's total management system server, as the administrative center of system, monitoring center and data processing centre
For performing following management function：Threshold value is crossed the border management, data processing management, system configuration management；
As the interface that user interacts with data center total management system, boundary is provided for system configuration for Web server
Sign-On authentication, data display function are completed in face.
HBase databases are responsible for carrying out data storage, for storing the monitoring parameter of data center's total management system,
Monitoring data and treated statistics.
By as above System and method for proposed by the invention, the unified management of data center is can contribute to, especially
The scene monitored is needed simultaneously being related to physical resource and virtual resource, is disposed by data center's total management system of design
It in cloud platform, solves the problems, such as data center's total management system server failure, improves the stability of management.The party
Method has reference significance for studying the systems such as similar high availability management, unified management.
Description of the drawings
Fig. 1 is the design method workflow schematic diagram in the embodiment of the present invention；
Fig. 2 is the Solution Architecture figure in the embodiment of the present invention；
Specific embodiment
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to needed in the embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some embodiments of the present invention, for ability
For the those of ordinary skill of domain, without creative efforts, it can also be obtained according to these attached drawings other attached
Figure.
The present invention proposes a kind of data center's total management system design method based on cloud platform, referring to such as Fig. 1 institutes
The method flow diagram shown, including：
Step 1, the design of agreement connector：Agreement connector is divided into connector server end and connector client, often
A connector client all has the distal end view there are one connector server end, and connector server end receives from various companies
The connection request of device client is connect, and handles these requests；
The agreement connector of the Distributed Services layer of data center's total management system based on JMX frames, management system
In data center's total management system server deposited with server, the monitoring agent of Agent layer and data collector and data
The design method of storage is described as follows：
Distributed Services layer is the important component that program distribution management is realized in JMX specifications, mainly including agreement
Connector part.Agreement connector is made of agent side, data collector end and server end, and being mainly responsible for makes part operation
Become transparent, and a unified view is provided for the service in different server.
Data center's total management system server is JMX management clients, runs monitoring generation on the remote server
Reason or data collector program then cannot be with the direct mutual operations of server, and need using connection protocol, connector service
Device end need to use identical transport protocol with connector client.
JMX Distributed Services layer includes the agreement connector based on Simple Network Management Protocol, based on right inside internet
Agreement connector as request agency agreement, the agreement connector based on hypertext markup language, based on hypertext transfer protocol
Agreement connector and agreement connector based on remote method call.The course of work of agreement connector is fully transparent
's.Agreement connector is responsible for providing point-to-point company between data collector and data center's total management system server
It connects, connector depends on specific connection protocol.
Data center's total management system supports browser access monitoring agent, while also needs to support data center's synthesis
The agreement that Tomcat-AdminPortal is acted on behalf of by remote method call management and monitoring.Realization to Distributed Services layer is mainly pair
The realization of agreement connector, the system use the agreement connector based on RMI.
Distributed Services layer is the important component that program distribution management is realized in JMX specifications, mainly including agreement
Connector and protocol adaptor two parts.
Agreement connector：It is made of agent side and server end two parts, being mainly responsible for makes part operation become transparent,
And provide a unified view for the service in different server or power ＆ environment supervision host.
Protocol adaptor：There is no client, be operated in server end, and can be with format transmission number that client is understood that
According to.Master monitor, monitoring agent and data collector in data center's total management system meet framework specification, number
According between center total management system server and monitoring agent, the Distributed Services layer between data collector and monitoring agent
Connection is created using connector, connector is divided into connector server end and connector client, server end and client
Between mutual call for user be transparent.Connector client is connected to agency using remote method call connector
Afterwards, an interface identical with connector server end can be obtained, Agent layer application is called by this interface.
Data collector can obtain the monitoring data of monitoring agent by modes such as remote procedure calls；Data center is comprehensive
The monitoring parameter of monitoring agent can be set by closing Tomcat-AdminPortal；Meanwhile monitoring agent is when needed, it can be actively
Notice, notice data center total management system server Managed Resource shape are sent to data center's total management system server
Change of state etc..
Agreement connector in the embodiment of the present invention is divided into connector server end and connector client, each connector
Client generally all tool there are one Connection Service device end distal end view.Connector server end receives from various connectors
The connection request of client, and pass through server and go to handle these requests.Connection between the two is once successfully established, distal end
Monitoring agent reforms into transparent for monitoring server.
It is responsible for establishing RMI connector servers based on RMI protocol connector server end, connector client is
Web browser.Therefore realize that the agreement connector based on HTTP mainly realizes connector server end, implementation step is such as
Under：
First, a MBean server is created, is used as the container of MBean.Secondly, an ObjectName is created
Object, to specify the title for needing to register MBean and port, it would be desirable to which MBean's is registered in MBean servers.3rd,
An agreement connector object is created, and specifies the administration interface of MBean as HTML types interface.4th, by agreement connector pair
As being also registered in MBean servers, and start agreement device, wait HTTP connections.
JMX creates MBean servers there are two types of method, a kind of to need, using MBeanServerFactory factory class, to be somebody's turn to do
Class all exists in different JMX is realized, it contains the method for creating MBean server.When creating MBean servers,
The createMBeanServer provided in factory class methods can be called to create.This method can return to a new MBean clothes
Business device, after using, it is necessary to which releaseMBeanServer methods is called to discharge newly-built MBean servers.
Step 2, the design of monitoring agent, monitoring agent are the monitoring implementation sections of data center's total management system,
It is distributed in data center's monitoring host computer and server cluster, is responsible for acquisition, threshold test and the data-pushing of monitoring data, separately
It is outer that corresponding operating is carried out according to the instruction received；
Monitoring agent is the monitoring implementation section of data center's total management system, is distributed in data center's monitoring host computer
In server cluster, it is responsible for acquisition, threshold test and the data-pushing of monitoring data, according further to from monitoring data center
Instruction with server carries out corresponding operating.Monitoring agent uses modularized processing, between function module independently of each other, under utilization
Layer API realizes corresponding business.
Data center's total management system is directed to the monitoring of the operating status of the virtual machine server in cloud computing platform.Cloud
Server in computing platform has different operating system, and system utilizes the cross-platform sampling instrument of third party, can realize to cloud
The data acquisition of majority operation system in computing platform.Monitoring agent is realized to the part identical data of more than operating system
Acquisition, such as CPU real time loads, memory usage, when front disk service condition, network utilization, disk read-write rate and system
Information etc..
Monitoring agent establishes different correspondent entity classes to preserve monitoring data for different monitoring resources, and is entity
In attribute addition get and set attributes, data center's total management system server is allowed to pass through above method and obtain and is corresponded to
Monitoring data.
When monitoring agent starts, two threads can be started：The data acquisition thread GatherThread of server and
Establish the thread ServerThread of JMX servers.After collecting thread starts, virtual machine or dynamic where GatherThread acquisitions
The related data of ring monitoring host computer, and by the data got compared with threshold value, happen or pass when there is threshold value to cross the border
When defeated pattern is pushes, monitoring agent sends JMX notification informations, informs the current state of the monitor oneself being registered on itself
Situation.
Step 3, the design of data collector, data collector are responsible for collecting the information that a large amount of monitoring agents are transferred, and
Calling database purchase, data collector is distributed in the part of nodes of data center's total management system into cloud platform database
On, for gathering the monitoring information for the monitoring agent specified；
Since cloud platform database is more suitable for the concurrently write-in of mass data, the frequent of many small datas is not suitable for
Write-in, therefore data collector is devised in data center's total management system, it is transferred to be responsible for collecting a large amount of monitoring agents
Information, and call database purchase into cloud platform database.
Data collector is distributed on some nodes of system, for gathering the monitoring information for the monitoring agent specified.Number
The information for collecting which monitoring agent node is needed to be specified by data center's total management system server according to collector.System
When system starts, data center's total management system server can make JMX that the monitoring agent of registration be tried one's best average divide with technology
Dispensing data collector, data collector is according to the list of the monitoring agent of acquisition, the data of polling request monitoring agent.
Data center's total management system server can be with data collector timed communication, the prison where having data collector
When rock machine occurs in control node, data center's total management system server can withdraw the monitoring agent for being distributed to the data collector
Information, and these information are reassigned to active data collector.
In JMX frames, data collector is acted on behalf of using connector (Connector) access monitoring, automatic regular polling request
Up-to-date information.There are two types of the modes of the method for data center's total management system server calls monitoring agent：One kind is long-range
The correlation technique of monitoring agent is called, is got parms；Another kind is to directly acquire relevant parameter after creating agency.
Data collector is responsible for collecting the monitoring data for the monitoring agent acquisition specified, and monitoring data is saved in data
In storehouse, therefore data collector needs to configure the address of database, and data collector is made to be able to access that database.Data collection
The realization of device equally based on two threads, establishes server thread and data collection thread.After data collection thread starts, need
Data center's total management system server is waited to need the address for the monitoring agent collected for its distribution, works as data collector
Monitoring agent list for it is empty when, data collector just by way of training in rotation request, is acted on behalf of using client access monitoring,
Obtain monitoring data.
Step 4, the design of data center's total management system server, data center's total management system server is to be
Administrative center, monitoring center and the data processing centre of system, for performing following management function：Threshold value is crossed the border management, data
Processing management, system configuration management；
Data center's total management system server is administrative center, monitoring center and the data processing centre of system.Number
It is the most important part of data center's total management system according to center total management system server, integrated management system of data center
System server needs completion work more, below will be to number of data center's total management system server based on Map/Reduce
According to processing and the design of threshold process function.
The management that threshold value is crossed the border is divided into data center's total management system server end and monitoring agent end two parts, monitoring
Agent side is responsible for threshold value Preliminary detection, and the monitoring data that threshold value is crossed the border is pushed to data center's total management system service
Device；After data center's total management system server obtains monitoring information, monitoring data is parsed, obtains the data beyond threshold value
, warning content of crossing the border is organized, and administrator and user are notified in time in a manner of web page notification, transmission mail etc..
After data center's total management system server obtains monitoring data, data collection module transmits monitoring information
To data processing module, data processing module is responsible for carrying out the operations such as threshold determination and data statistics.Data collection module gathers
To after a data set, each monitoring parameter therein is parsed, the threshold value of corresponding parameter is then taken out in threshold list, so
The size of both comparisons is made to determine whether to send threshold alarm afterwards, if necessary to alert, will check to see whether corresponding pretreatment
Strategy if there is then triggering dependent event according to strategy, and sends threshold alarm, if directly sending threshold alarm without if,
Comparison until completing all monitored item.
During data processing, input data is the set of the monitoring data of certain server, and every monitoring data is json numbers
According to key-value pair, form is as follows：
Timestamp,{“item1”:value1,“item2”:value2,…,“itemN”:valueN}
The operation object of Map/Reduce computation models is Map<Key,Value>Value pair.In data processing, first will
Incoming data carry out Map and are decomposed into prescribed form subdata, then summarize subdata progress Reduce.First time Map is grasped
When making, incoming key is null, and value is a monitoring record, and when Map is exported, monitoring record is decomposed into key for prison
Control project item is worth for the form of { current time stamp, Current data values }
Reduce operations are carried out after the completion of Map operations, Map results are subjected to Reduce operations according to monitoring data item, it is defeated
Go out the statistical result according to monitoring data item.After Map/Reduce is operated, you can complete to certain server or certain number
Statistic of classification is carried out according to the monitoring data at center, then carries out data processing according to time series.
Data center's total management system server is most important part in data center's total management system, be responsible for
Data collector distributes the address of its monitoring agent that needs gather, and receives at notification message and message from monitoring agent
Reason, while it is also responsible for the statistical disposition of monitoring data.It realizes the function of data center's total management system server, also needs
Use multiple threads：CountThread threads are used for the timing statistical disposition of monitoring data；RegisterListener threads are used
In registering monitor to monitoring agent and data collector, when there is notification message to send, the corresponding class that handles is handled；
SeverThread threads are used to implement JMX server ends, and connecting interface is provided for Web server.During startup of server, also need
Monitoring agent is distributed into data collector, data collector is enable to obtain the IP address of monitoring agent, carries out data receipts
Collection.
Step 5, Web server designs, the interface that Web server is user to be interacted with data center total management system,
Interface is provided for system configuration, it is necessary to complete Sign-On authentication, data display function.
The interface that Web server is user to be interacted with data center total management system is, it is necessary to complete Sign-On authentication, data
The functions such as displaying, it is also necessary to provide interface for system configuration.
Web page is expression layer in Web server structure, is used for and user interaction；Request processing and data transmission are industry
It is engaged in logical layer, for handling web-page requests, it is specified that data transmission format, and Basic API is called to be obtained from database and shows number
According to and update the data monitoring parameter in storehouse；Database manipulation management and the connection of database, and api interface is provided for upper strata tune
With.
Step 6, the design of data storage, data center's total management system carry out data storage using HBase databases,
For storing the monitoring parameter of data center's total management system, monitoring data and treated statistics.
Data center's total management system carries out data storage using HBase databases, for storing data center's synthesis
The monitoring parameter of management system, monitoring data and treated statistics etc..
It is mainly what is accessed in units of monitoring agent that data center's total management system reads data from database, because
This mainly designs the line unit of monitoring data part the design of line unit, and monitoring data is divided into three kinds according to the difference of type：
Real-time monitoring data, monitoring parameter and the monitoring data after data processing.For real time data, line unit is by each prison
The form for controlling both unique AgentID and monitoring data acquisition time of agency composition " AgentID_ times " carries out tissue,
Monitoring data is stored in the column family of monitoring data；For monitoring parameter, the information such as threshold value are included in parameter, only need to specify one
Line unit；For the monitoring data after data processing, line unit includes opening for docket number, measurement type and timing statistics section
Begin the time, distinguished with the form of " AgentID_ measurement types _ time ".
The design key point of the embodiment of the present invention is：Design reasonably data center's total management system based on cloud platform
Design method is deployed in cloud platform by data center's total management system of design, solves data center's integrated management
The problem of system server fails improves the stability of management, and data center's total management system has automatic management,
It is laid down a regulation according to historical statistical data, realizes the operation such as being set automatic threshold, the data acquisition modes ratio of monitoring agent
Relatively abundant, data acquisition modes are various including IT kind equipments and power ＆ environment supervision categorical data so that integrated management system of data center
It completely one includes under designed integrated service frame, constructs data center's total management system efficiently, stable.
Accordingly, the present invention also proposes a kind of data center's total management system based on cloud platform, realizes in cloud computing
In platform, referring to Fig. 2, which includes：
Agreement connector：Agreement connector includes connector server end and connector client, each connector client
End all tools are there are one the distal end view of connector server end, and the receiving of connector server end is from various connector clients
Connection request, and handle these requests；
At least one monitoring agent；Monitoring implementation section of the monitoring agent as data center's total management system, distribution
In data center's monitoring host computer and server cluster, it is responsible for acquisition, threshold test and the data-pushing of monitoring data, in addition root
Corresponding operating is carried out according to the instruction received；
At least one data collector is responsible for collecting the information that a large amount of monitoring agents are transferred, and calls database purchase
Into cloud platform database, at least one data collector is distributed in the part of nodes of data center's total management system
On, for gathering the monitoring information for the monitoring agent specified；
Data center's total management system server, as the administrative center of system, monitoring center and data processing centre
For performing following management function：Threshold value is crossed the border management, data processing management, system configuration management；
As the interface that user interacts with data center total management system, boundary is provided for system configuration for Web server
Sign-On authentication, data display function are completed in face.
HBase databases are responsible for carrying out data storage, for storing the monitoring parameter of data center's total management system,
Monitoring data and treated statistics.
Method proposed by the invention can contribute to the unified management of data center, especially be related to physical resource with
Virtual resource needs the scene monitored simultaneously, is deployed in by data center's total management system of design in cloud platform, solves
The problem of data center total management system server failure, improves the stability of management.This method is similar for studying
High availability management, the systems such as unified management there is reference significance.
The foregoing description of the disclosed embodiments enables those skilled in the art to realize or use the present invention.To this
A variety of modifications of a little embodiments will be apparent for a person skilled in the art, and the general principles defined herein can
Without departing from the spirit or scope of the present invention, to realize in other embodiments.Therefore, the present invention will not be limited
The embodiments shown herein is formed on, but meets the most wide model consistent with the principles and novel features disclosed herein
It encloses.
Claims (10)
1. a kind of data center's total management system design method based on cloud platform, which is characterized in that include the following steps：
Step 1, the design of agreement connector：Agreement connector is divided into connector server end and connector client, Mei Gelian
It connects device client and all has distal end view there are one connector server end, connector server end receives from various connectors
The connection request of client, and handle these requests；
Step 2, the design of monitoring agent, monitoring agent are the monitoring implementation sections of data center's total management system, distribution
In data center's monitoring host computer and server cluster, it is responsible for acquisition, threshold test and the data-pushing of monitoring data, in addition root
Corresponding operating is carried out according to the instruction received；
Step 3, the design of data collector, data collector is responsible for collecting the information that a large amount of monitoring agents are transferred, and calls
Into cloud platform database, data collector is distributed on the part of nodes of data center's total management system database purchase,
For gathering the monitoring information for the monitoring agent specified；
Step 4, the design of data center's total management system server, data center's total management system server is system
Administrative center, monitoring center and data processing centre, for performing following management function：Threshold value is crossed the border management, data processing
Management, system configuration management；
Step 5, Web server designs, the interface that Web server is user to be interacted with data center total management system, to be
Under unified central planning put provides interface, it is necessary to complete Sign-On authentication, data display function；
Step 6, the design of data storage, data center's total management system carry out data storage using HBase databases, are used for
Store the monitoring parameter of data center total management system, monitoring data and treated statistics.
2. design method as described in claim 1, which is characterized in that agreement connector is based on RMI protocol connector, connection
Device server end is responsible for establishing RMI connector servers, and connector client is Web browser.
3. design method as claimed in claim 2, which is characterized in that the data that monitoring agent is gathered include but not limited to
CPU real time loads, memory usage, when front disk service condition, network utilization, disk read-write rate and system information；Prison
Control agency establishes different correspondent entity classes to preserve monitoring data for different monitoring resources, and data center is allowed to integrate
Tomcat-AdminPortal obtains corresponding monitoring data.
4. design method as claimed in claim 3, which is characterized in that when monitoring agent starts, two threads can be started：Clothes
The data acquisition thread GatherThread of the business device and thread ServerThread for establishing JMX servers；Collecting thread opens
After dynamic, the related data of virtual machine or power ＆ environment supervision host where GatherThread acquisitions, and by the data got and threshold
Value is compared, when have threshold value cross the border happen or transmission mode for push when, monitoring agent sends JMX notification informations, accuse
Know the current state status of the data collector oneself being registered on itself.
5. design method as claimed in claim 4, which is characterized in that data collector needs which monitoring agent node collected
Information specified by data center's total management system server, system start when, data center's total management system server
It can use JMX technologies that the monitoring agent of registration fifty-fifty distributed to data collector, data collector is according to the monitoring of acquisition
The list of agency, the data of polling request monitoring agent.
6. design method as claimed in claim 5, which is characterized in that data center's total management system server meeting and data
Collector timed communication, when there is rock machine in monitoring node where having data collector, data center's total management system service
Device can withdraw the monitoring agent information for being distributed to the data collector, and these information are reassigned to active data collection
Device.
7. design method as claimed in claim 6, which is characterized in that performed by data center's total management system server
Threshold value management of crossing the border is divided into data center's total management system server end and monitoring agent end two parts, and monitoring agent end is responsible for
Threshold value Preliminary detection, and the monitoring data that threshold value is crossed the border is pushed to data center's total management system server；Data center
After total management system server obtains monitoring information, monitoring data is parsed, obtains the data item beyond threshold value, organizes the announcement that crosses the border
Alert content, and administrator and user are notified with web page notification or transmission lettergram mode in time.
8. design method as claimed in claim 7, which is characterized in that data center's total management system server needs to use
Multiple threads realize multiple corresponding functions：CountThread threads are used for the timing statistical disposition of monitoring data；
RegisterListener threads are used to register monitor to monitoring agent and data collector, when there is notification message to send,
Corresponding processing class is handled；SeverThread threads are used to implement JMX server ends, and providing connection for Web server connects
Mouthful.
9. design method as claimed in claim 8, which is characterized in that data center's total management system is read from database
Data are accessed in units of monitoring agent, and monitoring data is divided into three kinds according to the difference of type：Real-time monitoring data, monitoring
Parameter and the monitoring data after data processing.
10. a kind of data center's total management system based on cloud platform, is realized in cloud computing platform, which is characterized in that should
Data center's total management system includes：
Agreement connector：Agreement connector includes connector server end and connector client, each connector client
There are one the distal end view of connector server end, connection of the connector server end receiving from various connector clients for tool
Request, and handle these requests；
At least one monitoring agent；Monitoring implementation section of the monitoring agent as data center's total management system, is distributed in number
According in center monitoring host and server cluster, it is responsible for acquisition, threshold test and the data-pushing of monitoring data, according further to institute
The instruction of reception carries out corresponding operating；
At least one data collector is responsible for collecting the information that a large amount of monitoring agents are transferred, and calls database purchase to cloud
In platform database, at least one data collector is distributed on the part of nodes of data center's total management system,
For gathering the monitoring information for the monitoring agent specified；
Data center's total management system server, is used for as the administrative center of system, monitoring center and data processing centre
Perform following management function：Threshold value is crossed the border management, data processing management, system configuration management；
As the interface that user interacts with data center total management system, interface is provided for system configuration for Web server, complete
Into Sign-On authentication, data display function.
HBase databases are responsible for carrying out data storage, for storing the monitoring parameter of data center's total management system, monitoring
Data and treated statistics.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711395908.7A CN108121639A (en) | 2017-12-21 | 2017-12-21 | A kind of data center's total management system design method based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711395908.7A CN108121639A (en) | 2017-12-21 | 2017-12-21 | A kind of data center's total management system design method based on cloud platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108121639A true CN108121639A (en) | 2018-06-05 |
Family
ID=62230992
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711395908.7A Pending CN108121639A (en) | 2017-12-21 | 2017-12-21 | A kind of data center's total management system design method based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108121639A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109327335A (en) * | 2018-10-07 | 2019-02-12 | 杭州安恒信息技术股份有限公司 | A kind of cloud monitoring solution system and method |
| CN111800297A (en) * | 2020-07-07 | 2020-10-20 | 浪潮云信息技术股份公司 | Snmp-based intelligent monitoring method and system for cloud physical host |
| CN112104739A (en) * | 2020-09-18 | 2020-12-18 | 江苏工程职业技术学院 | Cloud computing remote data monitoring system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103559306A (en) * | 2013-11-18 | 2014-02-05 | 电子科技大学 | Query system and method for accessing data centers through cloud platform |
| CN107203454A (en) * | 2017-05-23 | 2017-09-26 | 郑州云海信息技术有限公司 | A kind of kernel internal memory monitoring method of power ＆ environment supervision main frame |
-
2017
- 2017-12-21 CN CN201711395908.7A patent/CN108121639A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103559306A (en) * | 2013-11-18 | 2014-02-05 | 电子科技大学 | Query system and method for accessing data centers through cloud platform |
| CN107203454A (en) * | 2017-05-23 | 2017-09-26 | 郑州云海信息技术有限公司 | A kind of kernel internal memory monitoring method of power ＆ environment supervision main frame |
Non-Patent Citations (1)
| Title |
|---|
| 吴夫丹: "基于云平台的服务器监控系统设计", 《 中国优秀硕士学位论文全文数据库》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109327335A (en) * | 2018-10-07 | 2019-02-12 | 杭州安恒信息技术股份有限公司 | A kind of cloud monitoring solution system and method |
| CN111800297A (en) * | 2020-07-07 | 2020-10-20 | 浪潮云信息技术股份公司 | Snmp-based intelligent monitoring method and system for cloud physical host |
| CN112104739A (en) * | 2020-09-18 | 2020-12-18 | 江苏工程职业技术学院 | Cloud computing remote data monitoring system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8230056B2 (en) | Enterprise management system | |
| CN101277304B (en) | Management system and management method for Web service operational environment based on rules | |
| CN100521628C (en) | Expandable dynamic network monitor system and its monitor method | |
| CN103973815A (en) | Method for unified monitoring of storage environment across data centers | |
| CN108092813A (en) | Data center's total management system server hardware Governance framework and implementation method | |
| CN110659109B (en) | System and method for monitoring openstack virtual machine | |
| CN106201754A (en) | Mission bit stream analyzes method and device | |
| CN101095307A (en) | Network management appliance | |
| CN101815013B (en) | Method for monitoring operation of satellite application system based on Ajax and Web service technology | |
| CN109194617A (en) | The automatically parsing of XML message, packaging method and device | |
| CN108121639A (en) | A kind of data center's total management system design method based on cloud platform | |
| Sun et al. | An architecture model of management and monitoring on cloud services resources | |
| CN103795575B (en) | A kind of system monitoring method towards multiple data centers | |
| CN109391516A (en) | Realize the cloud third party NMS system of more producer UTN equipment centralized maintenance management | |
| CN108923976A (en) | Space communication private network lightweight network operation management system | |
| CN111984505B (en) | Fortune dimension data acquisition device and acquisition method | |
| CN108182263A (en) | A kind of date storage method of data center's total management system | |
| CN104346168B (en) | A kind of monitoring management method for visualizing based on interexchange bus | |
| CN106713428A (en) | Business operation support system applied to Internet-of-things self-management platform | |
| CN110011984A (en) | A kind of distributed cluster system and method based on REST and RPC | |
| CN110929130B (en) | Public security level audit data query method based on distributed scheduling | |
| CN117370053A (en) | Information system service operation-oriented panoramic monitoring method and system | |
| CN102298648A (en) | Out-of-process access method of open real-time database | |
| Casola et al. | A reference architecture for sensor networks integration and management | |
| Bauer et al. | Services supporting management of distributed applications and systems |
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: 20180605 |