CN107797874A - A kind of resource management-control method based on embedded jetty and spark on yarn frameworks - Google Patents

A kind of resource management-control method based on embedded jetty and spark on yarn frameworks Download PDF

Info

Publication number
CN107797874A
CN107797874A CN201710947052.3A CN201710947052A CN107797874A CN 107797874 A CN107797874 A CN 107797874A CN 201710947052 A CN201710947052 A CN 201710947052A CN 107797874 A CN107797874 A CN 107797874A
Authority
CN
China
Prior art keywords
spark
jetty
tomcat
yarn
application
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.)
Granted
Application number
CN201710947052.3A
Other languages
Chinese (zh)
Other versions
CN107797874B (en
Inventor
邹涛
滕飞
刘冉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Sinovatio Technology LLC
Original Assignee
Nanjing Sinovatio Technology LLC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Sinovatio Technology LLC filed Critical Nanjing Sinovatio Technology LLC
Priority to CN201710947052.3A priority Critical patent/CN107797874B/en
Publication of CN107797874A publication Critical patent/CN107797874A/en
Application granted granted Critical
Publication of CN107797874B publication Critical patent/CN107797874B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/54Interprogram communication
    • G06F9/546Message passing systems or structures, e.g. queues

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Telephonic Communication Services (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

The invention discloses a kind of resource management-control method based on embedded jetty and spark on yarn frameworks, comprise the following steps:(1) hadoop cluster and spark clusters are built, disposes front and back tomcat servers;(2) user initiates spark application application resource requests to system;(3) backstage tomcat receives request, starts Jetty services by submit methods, loads servlet, generate Resident Process;Initiation message queue, initialization spark context generation spark applications, and most application is submitted in yarn clusters at last;(4) backstage tomcat will start returns to foreground tomcat using result;(5) foreground tomcat sends the Jetty Resident Process that follow-up spark tasks are submitted to background application;(6) background task Resident Process receives and performs task, returns to user's final result.Beneficial effects of the present invention are:User can apply for multiple spark application resources from a yarn resource queue, ensure that maximum resource utilization;Follow-up work under multiple different applications can perform simultaneously.

Description

A kind of resource management-control method based on embedded jetty and spark on yarn frameworks
Technical field
It is especially a kind of to be based on embedded jetty and spark the present invention relates to spark application resource application control fields The resource management-control method of on yarn frameworks.
Background technology
With the development of big data technology, the value of data is highlighted increasingly, and increasing enterprise is absorbed in from magnanimity number Business opportunity is found according to central analysis and excavation, and spark excavates most popular framework as current data, by more and more enterprises Favor.Spark provides a kind of resource management and control based on yarn, i.e., spark tasks is submitted into yarn clusters, yarn frameworks The concept of the resource queue of offer realizes that yarn cluster resources are relatively isolated.
As complete data mining platform, the foreground of user's operation and the backstage of processing request will possess.And user Whole process is asked to pay close attention to following problem to backstage application spark application resources by operating foreground:First, backstage Need uniformly to receive and return the framework of foreground message;Secondly, while there are multiple application resource applications to ask, while perform multiple Spark tasks, and resource isolation is realized in each spark applications;Again, it must have from the background and resided corresponding to spark applications Process, ensure that task dynamic is submitted;Finally, multiple spark applications can be submitted in same yarn resource queues, realize team Row resource makes full use of.
The content of the invention
The technical problems to be solved by the invention are, there is provided one kind is based on embedded jetty and spark on yarn frames It the resource management-control method of frame, can realize that backstage performs the spark application resources application request initiated on foreground, and realize resource pipe Control.
In order to solve the above technical problems, the present invention provide it is a kind of based on embedded jetty and spark on yarn frameworks Resource management-control method, comprises the following steps:
(1) hadoop cluster and spark clusters are built, disposes front and back tomcat servers;
(2) user initiates spark application application resource requests to system;
(3) backstage tomcat receives request, starts Jetty services by submit methods, loads servlet, generation is normal In process;Initiation message queue, initialization spark context generation spark applications, and most application is submitted to yarn clusters at last In;
(4) backstage tomcat will start returns to foreground tomcat using result;
(5) foreground tomcat sends the Jetty Resident Process that follow-up spark tasks are submitted to background application;
(6) background task Resident Process receives and performs task, returns to user's final result.
Preferably, in step (3), backstage tomcat starts jetty services, ensures each jetty Resident Process corresponding one Individual spark applications, the task requests that spark applications are persistently sent for subsequent user are prepared.
Preferably, in step (3), in Jetty service activation process, comprising loading servlet, front and back interaction is started Message queue, spark contexts are initialized, to yarn cluster application resources, controlled, protected by the realization of spring case mechanisms The succession of card event and the uniformity of environmental variance.
Preferably, in step (3), Jetty services are started by submit patterns, ensure that multiple spark application can be with It is submitted in a yarn resource queue.
Preferably, in step (5), foreground tomcat is follow-up spark tasks by different to row name and task mark Knowledge is sent in different message queues, and the backstage permanent processes of different Jetty receive the message in message queue, and then perform Different spark tasks.
Beneficial effects of the present invention are:User can by multiple spark application resources from a yarn resource queue Shen Please, it ensure that maximum resource utilization;Follow-up work under multiple different applications can perform simultaneously.
Brief description of the drawings
Fig. 1 is the structural framing schematic diagram of the present invention.
Fig. 2 is the method flow schematic diagram of the present invention.
Embodiment
Fig. 1 is the knot for the resource management and control based on embedded jetty and spark on yarn frameworks implemented according to the present invention Structure schematic diagram.Wherein include backstage tomcat modules, jetty engine modules, message queue module etc..
Backstage tomcat modules are mainly responsible for the http request for the spark application resource applications that receiving front-end comes, and start Jetty is serviced, record Jetty attribute informations to mysql databases, mistake rollback and return request results.
Jetty engine modules are mainly responsible for starting jetty Resident Process, load servlet, front and back interaction message team Row, spark contexts are initialized, submit task to yarn resource queues, and wrong rollback mechanism.
Message queue module is mainly responsible for follow-up front and back message communicating.User has applied for that the spark of resource should to some With new task requests are submitted, corresponding jetty resident service processes, corresponding jetty are transmitted directly to by message queue Spark tasks corresponding to response and processing.
Fig. 2 is the stream for the resource management and control based on embedded Jetty and spark on yarn frameworks implemented according to the present invention Journey schematic diagram.Specific step is described below:
(1) user initiates application resource request to backstage tomcat;
(2) backstage tomcat receives request, starts jetty services by submit methods,
(3) jetty calls start methods, loads servlet, generates Resident Process;
(4) spring frameworks built in jetty drive case mechanism initiation message queue successively, initialize spark contexts Spark applications are generated, and most application is submitted in yarn clusters at last.
(5) backstage tomcat will start returns to foreground using result, and errors excepted rollback simultaneously returns to error result;
By above-mentioned basic procedure, user can by multiple spark application resources from a yarn resource queue Shen Please, maximum resource utilization is ensure that, and the follow-up work under multiple different applications can simultaneously be performed and is independent of each other.
The present invention is based on spark on yarn Frame Designs, is realized by adding embedded Jetty services by foreground The request of spark application resources is converted to the resource allocation of yarn frameworks identification.
By using such scheme, user can easily carry out spark application resources point according to different business complexity Match somebody with somebody, and do not interfere with each other more using the operation of multitask notice, suitable for actual production requirement.
Although the present invention is illustrated and described with regard to preferred embodiment, it is understood by those skilled in the art that Without departing from scope defined by the claims of the present invention, variations and modifications can be carried out to the present invention.

Claims (5)

1. a kind of resource management-control method based on embedded jetty and spark on yarn frameworks, it is characterised in that including such as Lower step:
(1) hadoop cluster and spark clusters are built, disposes front and back tomcat servers;
(2) user initiates spark application application resource requests to system;
(3) backstage tomcat receive request, by submit methods start Jetty service, load servlet, generation reside into Journey;Initiation message queue, initialization spark context generation spark applications, and most application is submitted in yarn clusters at last;
(4) backstage tomcat will start returns to foreground tomcat using result;
(5) foreground tomcat sends the Jetty Resident Process that follow-up spark tasks are submitted to background application;
(6) background task Resident Process receives and performs task, returns to user's final result.
2. the resource management-control method as claimed in claim 1 based on embedded jetty and spark on yarn frameworks, it is special Sign is, in step (3), backstage tomcat starts jetty services, ensures that the corresponding spark of each jetty Resident Process should With the task requests that spark applications are persistently sent for subsequent user are prepared.
3. the resource management-control method as claimed in claim 1 based on embedded jetty and spark on yarn frameworks, it is special Sign is, in step (3), in Jetty service activation process, comprising loading servlet, starts the message team that front and back interacts Row, spark contexts are initialized, to yarn cluster application resources, controlled by the realization of spring case mechanisms, ensure event Succession and environmental variance uniformity.
4. the resource management-control method as claimed in claim 1 based on embedded jetty and spark on yarn frameworks, it is special Sign is, in step (3), starts Jetty services by submit patterns, ensure that multiple spark applications can be submitted to one In individual yarn resource queues.
5. the resource management-control method as claimed in claim 1 based on embedded jetty and spark on yarn frameworks, it is special Sign is, in step (5), foreground tomcat is sent to follow-up spark tasks to row name and task identification by different In different message queues, the backstage permanent processes of different Jetty receive the message in message queue, and then perform different Spark tasks.
CN201710947052.3A 2017-10-12 2017-10-12 Resource management and control method based on embedded jetty and spark on grow framework Active CN107797874B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710947052.3A CN107797874B (en) 2017-10-12 2017-10-12 Resource management and control method based on embedded jetty and spark on grow framework

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710947052.3A CN107797874B (en) 2017-10-12 2017-10-12 Resource management and control method based on embedded jetty and spark on grow framework

Publications (2)

Publication Number Publication Date
CN107797874A true CN107797874A (en) 2018-03-13
CN107797874B CN107797874B (en) 2021-04-27

Family

ID=61532986

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710947052.3A Active CN107797874B (en) 2017-10-12 2017-10-12 Resource management and control method based on embedded jetty and spark on grow framework

Country Status (1)

Country Link
CN (1) CN107797874B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108563508A (en) * 2018-04-27 2018-09-21 新华三大数据技术有限公司 YARN resource allocation methods and device
CN109408537A (en) * 2018-10-18 2019-03-01 网易(杭州)网络有限公司 Data processing method and device, storage medium and calculating equipment based on Spark SQL
CN110324188A (en) * 2019-07-05 2019-10-11 上海中通吉网络技术有限公司 Monitoring warning system and method
CN110471777A (en) * 2019-06-27 2019-11-19 中国科学院计算机网络信息中心 Multiple users share uses the method and system of Spark cluster in a kind of Python-Web environment

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5797005A (en) * 1994-12-30 1998-08-18 International Business Machines Corporation Shared queue structure for data integrity
CN102880475A (en) * 2012-10-23 2013-01-16 上海普元信息技术股份有限公司 Real-time event handling system and method based on cloud computing in computer software system
CN103501341A (en) * 2013-09-30 2014-01-08 乐视致新电子科技(天津)有限公司 Method and device for establishing Web service
CN105205169A (en) * 2015-10-12 2015-12-30 中国电子科技集团公司第二十八研究所 Distributed image index and retrieval method
CN105677382A (en) * 2015-12-28 2016-06-15 深圳市华讯方舟软件技术有限公司 Installation tool of big database system
CN106445705A (en) * 2016-10-08 2017-02-22 武汉烽火众智数字技术有限责任公司 Platform cascading method and system based on Rabbit MQ and Socket.io

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5797005A (en) * 1994-12-30 1998-08-18 International Business Machines Corporation Shared queue structure for data integrity
CN102880475A (en) * 2012-10-23 2013-01-16 上海普元信息技术股份有限公司 Real-time event handling system and method based on cloud computing in computer software system
CN103501341A (en) * 2013-09-30 2014-01-08 乐视致新电子科技(天津)有限公司 Method and device for establishing Web service
CN105205169A (en) * 2015-10-12 2015-12-30 中国电子科技集团公司第二十八研究所 Distributed image index and retrieval method
CN105677382A (en) * 2015-12-28 2016-06-15 深圳市华讯方舟软件技术有限公司 Installation tool of big database system
CN106445705A (en) * 2016-10-08 2017-02-22 武汉烽火众智数字技术有限责任公司 Platform cascading method and system based on Rabbit MQ and Socket.io

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
TROY BAER,PAUL PELTZ,JUNQI YIN: "Integrating Apache Spark Into PBS-Based HPC Environments", 《XSEDE "15: PROCEEDINGS OF THE 2015 XSEDE CONFERENCE》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108563508A (en) * 2018-04-27 2018-09-21 新华三大数据技术有限公司 YARN resource allocation methods and device
CN109408537A (en) * 2018-10-18 2019-03-01 网易(杭州)网络有限公司 Data processing method and device, storage medium and calculating equipment based on Spark SQL
CN110471777A (en) * 2019-06-27 2019-11-19 中国科学院计算机网络信息中心 Multiple users share uses the method and system of Spark cluster in a kind of Python-Web environment
CN110471777B (en) * 2019-06-27 2022-04-15 中国科学院计算机网络信息中心 Method and system for realizing multi-user sharing and using Spark cluster in Python-Web environment
CN110324188A (en) * 2019-07-05 2019-10-11 上海中通吉网络技术有限公司 Monitoring warning system and method

Also Published As

Publication number Publication date
CN107797874B (en) 2021-04-27

Similar Documents

Publication Publication Date Title
CN107493310B (en) Cloud resource processing method and cloud management platform
CN107797874A (en) A kind of resource management-control method based on embedded jetty and spark on yarn frameworks
WO2021073414A1 (en) Task scheduling method and related apparatus
US11716264B2 (en) In situ triggered function as a service within a service mesh
CN101262498B (en) A distributed message call method and device
CN101179622B (en) Multi-order processing arrangement in communication system and processing method thereof
CN103593236A (en) Computer equipment, method and device for scheduling service process
CN112395736B (en) Parallel simulation job scheduling method of distributed interactive simulation system
CN107919987B (en) Method for realizing micro-service cloud deployment
CN106027617A (en) Method for implementing dynamic scheduling of tasks and resources in private cloud environment
CN108519918A (en) Distributed transaction processing method based on transaction chain
CN114816730A (en) Robot process automation cloud service system and implementation method
CN105893320A (en) Remote task function calling method oriented at multi-core processor
CN111679911A (en) Management method, device, equipment and medium for GPU (graphics processing Unit) card in cloud environment
CN105740085A (en) Fault tolerance processing method and device
CN109858772A (en) A kind of processing method of work order, device and system
CN113032125A (en) Job scheduling method, device, computer system and computer-readable storage medium
CN112688915A (en) Cross-protocol communication method, device and server
CN114816694A (en) Multi-process cooperative RPA task scheduling method and device
CN107451211B (en) A kind of download system based on RabbitMQ and MongoDB
CN109634752A (en) A kind of client request processing method and system based on page gateway
CN106131020B (en) Firewall virtualization module and management method
CN102591714B (en) Process calling method, system and application server
CN109558254B (en) Asynchronous callback method, system, device and computer readable storage medium
CN116204307A (en) Federal learning method and federal learning system compatible with different computing frameworks

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
GR01 Patent grant
GR01 Patent grant