CN107786358A - The expansion method of distributed system and the distributed system - Google Patents
The expansion method of distributed system and the distributed system Download PDFInfo
- Publication number
- CN107786358A CN107786358A CN201610753112.3A CN201610753112A CN107786358A CN 107786358 A CN107786358 A CN 107786358A CN 201610753112 A CN201610753112 A CN 201610753112A CN 107786358 A CN107786358 A CN 107786358A
- Authority
- CN
- China
- Prior art keywords
- node
- calculate node
- dilatation
- new
- observer
- 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.)
- Withdrawn
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0896—Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer And Data Communications (AREA)
- Multi Processors (AREA)
Abstract
The present invention provides a kind of distributed system and the expansion method of the distributed system, the distributed system includes host node, multiple calculate nodes, cache node, observer and resource allocation system, the expansion method of the distributed system is first itself to carry out dilatation to each calculate node, when the quantity that dilatation is completed in itself of all calculate nodes is more than default quantity threshold values, the resource allocation system is increased the new calculate node by required hardware resource and load operating software image and carries out dilatation, so the expansion method optimizes the dilation process of the distributed system, realize the automation of dilatation, improve the efficiency of dilatation.
Description
Technical field
The present invention relates to IT (Internet technology, Internet Technology, abbreviation IT) and CT (telecommunications network technology,
Community Technology, abbreviation CT) field, the dilatation side of more particularly to a kind of distributed system and the distributed system
Method.
Background technology
Distributed system is built upon the software systems on network, has the cohesion and the transparency of height so that every
The distribution node high degree of autonomy of one database, and the distribution node of each database is transparent for the application of user
, do not see it is local or long-range, so the appearance of distributed system solves the problems, such as large-scale calculations and parallel computation,
Favored by users.
At present, the framework of common distributed system typically uses master-slave mode, i.e. host node is responsible for receiving and distributed and appoints
Business, it is multiple to perform specific task from node (being also calculate node).Host node, which is only done, to be distributed, and will not turn into system bottleneck；And
Execution task from node often as the increase of amount of calculation and the rising of resource utilization turn into bottleneck.At this moment with regard to needs pair
Distributed system carries out dilatation, and common expansion method is exactly to increase one or more calculate nodes using manual mode, such a
For expansion method although with certain effect, dilatation granularity is thicker, and less efficient.
The content of the invention
It is a kind of with efficient distributed system and the distributed decorum it is an object of the invention to propose based on this
Expansion method, the expansion method optimizes the dilation process of the distributed system, in a more accurate way to dilation process
Needed for resource distributed according to need, realize the automation of dilation process.
The present invention provides a kind of expansion method of distributed system, and the distributed system includes host node, multiple calculating
Node, the host node are responsible for receiving and distributing task, and the calculate node performs the corresponding task of the host node distribution,
Further comprise cache node and observer, the expansion method includes：
The cache node record distributes to the queue obstruction information of each calculate node, and sends the observation to
Person；
The host node receives the resource utilization information of itself that each calculate node periodically reports, and sends to
The observer；
The observer collects the queue obstruction information, the resource utilization information, and judges each calculating
Whether node exceedes itself default throughput threshold, if not exceeded, then being saved in a manner of increasing thread to each described calculate
The inside of point carries out first time dilatation.
The expansion method of the distributed system, wherein, the distributed system also includes resource management system, if each
The calculate node is above itself default throughput threshold, then judges whether the calculate node exceedes default resource profit
With rate threshold values, if exceeding, hardware resource and Software Mirroring needed for the resource management system distribution, to increase new calculating
The mode of node carries out second of dilatation.
The expansion method of the distributed system, wherein, the first time dilatation includes：
The observer receives the capacity expansion application that the host node is actively sent, and sends first time dilatation order；
The calculate node receives the first time dilatation order, and increased number of threads needed for calculating is simultaneously sent to described
Host node；
The observer receives the number of threads of the host node transmission, and judges whether the number of threads exceedes
Default higher limit, if not exceeded, then the calculate node increased number of threads needed for carries out dilatation, if super
Cross, then dilatation is carried out in a manner of increasing thread to the other calculate node.
The expansion method of the distributed system, wherein, second of dilatation includes：
Hardware needed for the quantity and the new calculate node of the increased new calculate node of observer's calculating need
Resource and load operating software image, and second of dilatation order is sent to the resource allocation system；
The resource allocation system receives second of dilatation order, distributes required hardware resource and load operating is soft
Part image, and the new calculate node is started by container mode；
The new calculate node is completed after starting, and is cached to the cache node solicitation message and is received the caching section
The allocation result that point is completed, while send addition request to the host node；
The host node issues new calculating task after receiving addition request, completes second of dilatation.
The expansion method of the distributed system, wherein,
When the resource allocation system includes infrastructure services system and container cluster management system, the observer
Need first to the hardware resource needed for the new calculate node of the infrastructure services system application, then to the container cluster management
Software Mirroring needed for the new calculate node of system application, the container cluster management system start new meter by container mode
Operator node；
When the resource allocation system is platform service system, the observer is new to the platform service system application
Calculate node needed for hardware resource and Software Mirroring, the platform service system starts new calculating section by container mode
Point.
Compare prior art, in expansion method of the present invention, itself dilatation first carried out to each calculate node,
When the quantity that dilatation is completed in itself of all calculate nodes is more than default quantity threshold values, the resource allocation system
Increase the new calculate node by required hardware resource and load operating software image and carry out dilatation, so the dilatation
Method optimizes the dilation process of the distributed system, realizes the automation of dilatation, improves the efficiency of dilatation.
A kind of distributed system, including host node, multiple calculate nodes, the host node are responsible for receiving and distributing task,
The calculate node performs the corresponding task of the host node distribution, further comprises cache node and observer；
The cache node between the host node and the calculate node, by record distribute to it is each it is described based on
The queue obstruction information of operator node, and send the observer to；
The host node is used to receive the resource utilization information of itself that each calculate node periodically reports, and passes
Give the observer；
The observer is used to collect the queue obstruction information, the resource utilization information, and judges each described
Whether calculate node exceedes itself default throughput threshold, if not exceeded, then described being counted to each in a manner of increasing thread
The inside of operator node carries out first time dilatation.
The distributed system, wherein, the distributed system also includes resource management system, the resource allocation system
For distributing required hardware resource and Software Mirroring, if each calculate node is above itself default handling capacity valve
Value, then judge whether the calculate node exceedes default resource utilization threshold values, if exceeding, to increase new calculate node
Mode carry out second of dilatation.
The distributed system, wherein, the first time dilatation includes：
The observer is used to receive the capacity expansion application that the host node is actively sent, and sends first time dilatation order；
The calculate node is used to receive the first time dilatation order, and increased number of threads needed for calculating is simultaneously sent to
The host node；
The observer is used for the number of threads for receiving the host node transmission, and whether judges the number of threads
More than default higher limit, if not exceeded, then the calculate node increased number of threads needed for carries out dilatation, if
Exceed, then dilatation is carried out in a manner of increasing thread to the other calculate node.
The distributed system, wherein, second of dilatation includes：
The observer is used to calculate needed for the quantity for needing increased new calculate node and the new calculate node
Hardware resource and load operating software image, and second of dilatation order is sent to the resource allocation system；
The resource allocation system is used to receive second of dilatation order, distributes required hardware resource and loading is transported
Row software image, and the new calculate node is started by container mode；
The new calculate node is completed after starting, described slow for caching and receiving to the cache node solicitation message
The allocation result of node completion is deposited, while addition request is sent to the host node；
What the host node was used to receiving the new calculate node adds the new calculating task of request Concurrency cloth, completes the
Secondary dilatation.
The distributed system, wherein,
When the resource allocation system includes infrastructure services system and container cluster management system, the observer
For the hardware resource needed for elder generation to the new calculate node of the infrastructure services system application, then to the container cluster pipe
Software Mirroring needed for the new calculate node of reason system application, the container cluster management system are started new by container mode
Calculate node；
When the resource allocation system is platform service system, the observer is used for the platform service system Shen
Hardware resource and Software Mirroring needed for calculate node that please be new, the platform service system start new meter by container mode
Operator node.
Brief description of the drawings
The above-mentioned and/or additional aspect and advantage of the present invention will become in the description from combination accompanying drawings below to embodiment
Substantially and it is readily appreciated that, wherein：
Fig. 1 is the structured flowchart of distributed system in the present invention；
Fig. 2 is the flow chart of the expansion method of distributed system in first embodiment of the invention；
Fig. 3 is the flow chart of the expansion method of distributed system in second embodiment of the invention；
Fig. 4 is the dilatation flow chart of first time dilatation in first embodiment of the invention；
Fig. 5 is the dilatation flow chart of second of dilatation in second embodiment of the invention；
Fig. 6 is the dilatation flow chart based on infrastructure services system in third embodiment of the invention；
Fig. 7 is the dilatation flow chart based on platform service system in fourth embodiment of the invention.
Main element symbol description：
| Host node | 10 | Cache node | 11 |
| Calculate node | 12 | Observer | 13 |
| Resource allocation system | 14 |
Following embodiment will combine above-mentioned accompanying drawing and further illustrate the present invention.
Embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein from beginning to end
Same or similar label represents same or similar element or the element with same or like function.Below with reference to attached
The embodiment of figure description is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.
In the description of the invention, term " longitudinal direction ", " transverse direction ", " on ", " under ", "front", "rear", "left", "right", " perpendicular
Directly ", the orientation of the instruction such as " level ", " top ", " bottom " or position relationship are based on orientation shown in the drawings or position relationship, are only
For the ease of the description present invention rather than require that the present invention must be therefore it is not intended that right with specific azimuth configuration and operation
The limitation of the present invention.
With reference to following description and accompanying drawing, it will be clear that these and other aspects of embodiments of the invention.In these descriptions
In accompanying drawing, some particular implementations in embodiments of the invention are specifically disclosed, to represent to implement the implementation of the present invention
Some modes of the principle of example, but it is to be understood that the scope of embodiments of the invention is not limited.On the contrary, the present invention
Embodiment includes all changes, modification and the equivalent fallen into the range of the spirit and intension of attached claims.
Referring to Fig. 1, the present invention provides a kind of distributed system, including host node 10, multiple calculate nodes 12, the master
Node 10 is responsible for receiving and distribution task, the calculate node 12 perform the corresponding task that the host node 10 distributes, enter one
Step includes cache node 11 and observer 13；
The cache node 11 is distributed to each between the host node 10 and the calculate node 12 for record
The queue obstruction information of the calculate node 12, and send the observer 13 to；
The host node 10 is used to receive the resource utilization information of itself that each calculate node 12 periodically reports,
And send the observer 13 to；
The observer 13 is used to collect the queue obstruction information, the resource utilization information, and judges each institute
State whether calculate node 12 exceedes itself default throughput threshold, if not exceeded, then to each institute in a manner of increasing thread
The inside for stating calculate node 12 carries out first time dilatation.
The distributed system also includes resource management system 14, and the resource allocation system 14 is used to distributing required hard
Part resource and Software Mirroring, if each calculate node 12 is above itself default throughput threshold, judge the meter
Whether operator node 12 exceedes default resource utilization threshold values, if exceeding, is carried out in a manner of increasing new calculate node 12
Second of dilatation.
When the resource allocation system 14 includes infrastructure services system and container cluster management system, the observation
Person 13 is used for first to the hardware resource needed for the new calculate node 12 of the infrastructure services system application, then to the container
Software Mirroring needed for the new calculate node 12 of cluster management system application, the container cluster management system pass through container mode
Start new calculate node 12；
When the resource allocation system 14 is platform service system, the observer 13 is used for the platform service system
System applies for the hardware resource and Software Mirroring needed for new calculate node 12, and the platform service system is started by container mode
New calculate node 12.
In above-mentioned distributed system, the queue obstruction information includes queue blocking time node, position, and queue resistance
Fill in the number of Information Repeater.The resource utilization includes CPU and uses percentage, free memory percentage when remanence disk percentage
Than.
Referring to Fig. 2, a kind of expansion method of the distributed system provided in first embodiment of the invention, the dilatation side
The step of method, includes：
Step S201, the cache node 11 record the queue obstruction information for distributing to each calculate node 12, and
Send the observer 13 to；
Step S202, the host node 10 receive the resource utilization of itself that each calculate node 12 periodically reports
Information, and send the observer 13 to；
Step S203, the observer 13 collects the queue obstruction information, the resource utilization information, and judges every
Whether the individual calculate node 12 exceedes itself default throughput threshold, if not exceeded, then to every in a manner of increasing thread
The inside of the individual calculate node 12 carries out first time dilatation.
Referring to Fig. 3, be a kind of expansion method of the distributed system provided in second embodiment of the invention, the dilatation
The step of method, includes：
Step S301, the cache node 11 record the queue obstruction information for distributing to each calculate node 12, and
Send the observer 13 to；
Step S302, the host node 10 receive the resource utilization of itself that each calculate node 12 periodically reports
Information, and send the observer 13 to；
Step S303, the observer 13 collects the queue obstruction information, the resource utilization information, and judges every
Whether the individual calculate node 12 exceedes itself default throughput threshold,
Step S304, if not exceeded, then carrying out the to the inside of each calculate node 12 in a manner of increasing thread
Dilatation.
Step S305, if exceeding, judge whether the calculate node 12 exceedes default resource utilization threshold values, if super
Cross, then the resource management system 14 distributes required hardware resource and Software Mirroring, to increase the side of new calculate node 12
Formula carries out second of dilatation.
Referring to Fig. 4, be the expansion method of first time dilatation in first embodiment of the invention, the step of the first time dilatation
Suddenly include：
Step S401, the observer 13 receive the capacity expansion application that the active of host node 10 is sent, and send for the first time
Dilatation order；
Step S402, the calculate node 12 receive the first time dilatation order, increased number of threads needed for calculating
And send the host node 10 to；
Step S403, the observer 13 receives the number of threads that the host node 10 transmits, and judges the line
Whether number of passes amount exceedes default higher limit, if not exceeded, then the calculate node 12 is according to the required increased Thread Count
Amount carries out dilatation, if exceeding, dilatation is carried out in a manner of increasing thread to the other calculate node 12.
Referring to Fig. 5, be the expansion method of second of dilatation in second embodiment of the invention, the step of second of dilatation
Suddenly include：
Step S501, the observer 13 calculate the quantity for needing increased new calculate node 12 and the new calculating section
Hardware resource and load operating software image needed for point 12, and second of dilatation order is sent to the resource allocation system
14；
Step S502, the resource allocation system 14 receive second of dilatation order, distribute required hardware resource
With load operating software image, and the new calculate node 12 is started by container mode；
Step S503, the new calculate node 12 are completed after starting, to the solicitation message of cache node 11 caching simultaneously
The allocation result that the cache node 11 is completed is received, while addition request is sent to the host node 10；
Step S504, the host node 10 issue new calculating task after receiving addition request, complete second of dilatation.
In above-mentioned dilation process, the distributed system combination periodic collection and actively triggering are realized dilatation monitoring and touched
Hair.Periodic collection includes the handling capacity that the observer 13 collects each calculate node 12, and each described calculate is saved
The resource utilization of itself of point 12, actively triggering referred between two collection cycles of the observer 13.When each institute
State the handling capacity of itself of calculate node 12 or when resource utilization reaches default threshold value, dilatation is actively initiated, when not up to
During the resource utilization threshold values, then do not process, when reaching the resource utilization threshold values, the observer 13 receives institute
The capacity expansion application that the active of host node 10 is sent is stated, and sends first time dilatation order.
In above-mentioned first time dilation process, each calculate node 12 persistently reports the use of the resource utilization of itself
Situation, first time dilatation is carried out to each calculate node 12 one by one, (can voluntarily be set by the observer 13 when having 80%
Put) more than the resource utilization of the calculate node 12 when being above default resource utilization threshold values, the observer 13
Send second of dilatation order.
Referring to Fig. 6, be the expansion method based on infrastructure services system provided in third embodiment of the invention, when
When the resource utilization of more than 80% calculate node 12 is above default resource utilization threshold values, the observer 13
Need first to the hardware resource needed for the new calculate node 12 of the infrastructure services system application, then to the sets of containers
Software Mirroring needed for the new calculate node 12 of cluster management system application, the container cluster management system pass through container side
Formula starts the new calculate node 12.Wherein, the step of expansion method includes：
Step S601, the observer 13 are needed first to the new calculate node 12 of the infrastructure services system application
Required hardware resource, then to the Software Mirroring needed for the new calculate node 12 of the container cluster management system application；
Step S602, the quantity for needing the increased calculate node 12 is calculated according to the observer 13, the basis is set
The hardware resource needed for service system distribution is applied, the container cluster management system distributes load operating software image；
Step S603, the container cluster management system start the calculate node 12 newly increased by container mode；
Step S604, the calculate node 12 newly increased are completed after starting, delayed to the solicitation message of cache node 11
The allocation result that the cache node 11 is completed is deposited and received, while addition request is sent to the host node 10；
Step S605, the host node 10 issue new calculating task after receiving addition request, complete dilatation.
Referring to Fig. 7, it is the expansion method based on platform service system in fourth embodiment of the invention, when more than 80%
When the resource utilization of the calculate node 12 is above default resource utilization threshold values, the observer 13 is to the platform
Hardware resource and Software Mirroring needed for the new calculate node 12 of service system application, the platform service system pass through appearance
Device mode starts the new calculate node 12.Wherein, the step of expansion method includes:
Step S701 is hard needed for the observer 13 to the new calculate node 12 of the platform service system application
Part resource and Software Mirroring；
Step S702, the platform service system is calculated by the observer 13 needs the increased calculate node 12
Quantity distributes required hardware resource and load operating software image；
Step S703, the platform service system start the calculate node 12 newly increased by container mode；
Step S704, the calculate node 12 newly increased are completed after starting, delayed to the solicitation message of cache node 11
The allocation result that the cache node 11 is completed is deposited and received, while addition request is sent to the host node 10；
Step S705, the host node 10 issue new calculating task after receiving addition request, complete dilatation.
To sum up, in expansion method of the present invention, itself dilatation first is carried out to each calculate node 12, when 80%
All calculate nodes 12 itself complete dilatation when, the resource allocation system 14 by required hardware resource and plus
Carrying row software image carries out dilatation to increase the new calculate node 12, so the expansion method optimizes the distribution
The dilation process of formula system, the automation of dilatation is realized, improve the efficiency of dilatation.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or the spy for combining the embodiment or example description
Point is contained at least one embodiment or example of the present invention.In this manual, to the schematic representation of above-mentioned term not
Necessarily refer to identical embodiment or example.Moreover, specific features, structure, material or the feature of description can be any
One or more embodiments or example in combine in an appropriate manner.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that：Not
In the case of departing from the principle and objective of the present invention a variety of change, modification, replacement and modification can be carried out to these embodiments, this
The scope of invention is limited by claim and its equivalent.
Claims (10)
1. a kind of expansion method of distributed system, the distributed system includes host node, multiple calculate nodes, the main section
Point is responsible for receiving and distribution task, and the calculate node performs the corresponding task of the host node distribution, it is characterised in that enters
One step includes cache node and observer, and the expansion method includes：
The cache node record distributes to the queue obstruction information of each calculate node, and sends the observer to；
The host node receives the resource utilization information of itself that each calculate node periodically reports, and sends to described
Observer；
The observer collects the queue obstruction information, the resource utilization information, and judges each calculate node
Whether itself default throughput threshold is exceeded, if not exceeded, then to each calculate node in a manner of increasing thread
Inside carries out first time dilatation.
2. expansion method according to claim 1, it is characterised in that the distributed system also includes resource management system
System, if each calculate node is above itself default throughput threshold, it is pre- to judge whether the calculate node exceedes
If resource utilization threshold values, if exceeding, hardware resource and Software Mirroring needed for resource management system distribution, to increase
The mode of new calculate node is added to carry out second of dilatation.
3. expansion method according to claim 1, it is characterised in that the first time dilatation includes：
The observer receives the capacity expansion application that the host node is actively sent, and sends first time dilatation order；
The calculate node receives the first time dilatation order, and increased number of threads needed for calculating simultaneously sends the main section to
Point；
The observer receives the number of threads of the host node transmission, and it is default to judge whether the number of threads exceedes
Higher limit, if not exceeded, then the calculate node increased number of threads needed for carries out dilatation, if exceeding,
Dilatation is carried out in a manner of increasing thread to the other calculate node.
4. expansion method according to claim 2, it is characterised in that second of dilatation includes：
Hardware resource needed for the quantity and the new calculate node of the increased new calculate node of observer's calculating need
With load operating software image, and send second of dilatation order to the resource allocation system；
The resource allocation system receives second of dilatation order, distributes required hardware resource and load operating software reflects
Picture, and the new calculate node is started by container mode；
The new calculate node is completed after starting, and is cached to the cache node solicitation message and to receive the cache node complete
Into allocation result, while to the host node send addition request；
The host node issues new calculating task after receiving addition request, completes second of dilatation.
5. expansion method according to claim 2, it is characterised in that
When the resource allocation system includes infrastructure services system and container cluster management system, the observer needs elder generation
Hardware resource needed for the calculate node new to the infrastructure services system application, then to the container cluster management system
Apply for the Software Mirroring needed for new calculate node, the container cluster management system starts new calculating section by container mode
Point；
When the resource allocation system is platform service system, the observer is to the new meter of the platform service system application
Hardware resource and Software Mirroring needed for operator node, the platform service system start new calculate node by container mode.
6. a kind of distributed system, including host node, multiple calculate nodes, the host node are responsible for receiving and distributing task, institute
State the corresponding task that calculate node performs the host node distribution, it is characterised in that further comprise cache node and observation
Person；
The cache node distributes to each calculating section between the host node and the calculate node, for recording
The queue obstruction information of point, and send the observer to；
The host node is used to receive the resource utilization information of itself that each calculate node periodically reports, and sends to
The observer；
The observer is used to collect the queue obstruction information, the resource utilization information, and judges each calculating
Whether node exceedes itself default throughput threshold, if not exceeded, then being saved in a manner of increasing thread to each described calculate
The inside of point carries out first time dilatation.
7. distributed system according to claim 6, it is characterised in that the distributed system also includes resource management system
System, the resource allocation system are used to distribute required hardware resource and Software Mirroring, if each calculate node is above
Itself default throughput threshold, then judge whether the calculate node exceedes default resource utilization threshold values, if exceeding,
Second of dilatation is carried out in a manner of increasing new calculate node.
8. distributed system according to claim 6, it is characterised in that the first time dilatation includes：
The observer is used to receive the capacity expansion application that the host node is actively sent, and sends first time dilatation order；
The calculate node is used to receive the first time dilatation order, and increased number of threads needed for calculating is simultaneously sent to described
Host node；
The observer is used for the number of threads for receiving the host node transmission, and judges whether the number of threads exceedes
Default higher limit, if not exceeded, then the calculate node increased number of threads needed for carries out dilatation, if super
Cross, then dilatation is carried out in a manner of increasing thread to the other calculate node.
9. distributed system according to claim 7, it is characterised in that second of dilatation includes：
Hardware of the observer needed for for the quantity of the increased new calculate node of calculating need and the new calculate node
Resource and load operating software image, and second of dilatation order is sent to the resource allocation system；
The resource allocation system is used to receive second of dilatation order, distributes required hardware resource and load operating is soft
Part image, and the new calculate node is started by container mode；
The new calculate node is completed after starting, for the caching section to be cached and received to the cache node solicitation message
The allocation result that point is completed, while send addition request to the host node；
What the host node was used to receiving the new calculate node adds the new calculating task of request Concurrency cloth, completes second
Dilatation.
10. expansion method according to claim 7, it is characterised in that
When the resource allocation system includes infrastructure services system and container cluster management system, the observer is used for
First to the hardware resource needed for the new calculate node of the infrastructure services system application, then to the container cluster management system
System applies for the Software Mirroring needed for new calculate node, and the container cluster management system starts new calculating by container mode
Node；
When the resource allocation system is platform service system, the observer is used for new to the platform service system application
Calculate node needed for hardware resource and Software Mirroring, the platform service system starts new calculating section by container mode
Point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610753112.3A CN107786358A (en) | 2016-08-29 | 2016-08-29 | The expansion method of distributed system and the distributed system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610753112.3A CN107786358A (en) | 2016-08-29 | 2016-08-29 | The expansion method of distributed system and the distributed system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107786358A true CN107786358A (en) | 2018-03-09 |
Family
ID=61441536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610753112.3A Withdrawn CN107786358A (en) | 2016-08-29 | 2016-08-29 | The expansion method of distributed system and the distributed system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107786358A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109062658A (en) * | 2018-06-29 | 2018-12-21 | 优刻得科技股份有限公司 | Realize dispatching method, device, medium, equipment and the system of computing resource serviceization |
| CN111290838A (en) * | 2020-05-09 | 2020-06-16 | 支付宝(杭州)信息技术有限公司 | Application access request processing method and device based on container cluster |
| CN111522624A (en) * | 2020-04-17 | 2020-08-11 | 成都安恒信息技术有限公司 | Message forwarding performance flexible extension system based on virtualization technology and extension method thereof |
| CN113497735A (en) * | 2020-04-08 | 2021-10-12 | 中兴通讯股份有限公司 | Open interface management method, electronic device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070022122A1 (en) * | 2005-07-25 | 2007-01-25 | Parascale, Inc. | Asynchronous file replication and migration in a storage network |
| CN103092698A (en) * | 2012-12-24 | 2013-05-08 | 中国科学院深圳先进技术研究院 | System and method of cloud computing application automatic deployment |
| CN104317658A (en) * | 2014-10-17 | 2015-01-28 | 华中科技大学 | MapReduce based load self-adaptive task scheduling method |
| CN104391930A (en) * | 2014-11-21 | 2015-03-04 | 用友软件股份有限公司 | Distributed file storage device and method |
| CN105357296A (en) * | 2015-10-30 | 2016-02-24 | 河海大学 | Elastic caching system based on Docker cloud platform |
| CN105786447A (en) * | 2014-12-26 | 2016-07-20 | 乐视网信息技术（北京）股份有限公司 | Method and apparatus for processing data by server and server |
-
2016
- 2016-08-29 CN CN201610753112.3A patent/CN107786358A/en not_active Withdrawn
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070022122A1 (en) * | 2005-07-25 | 2007-01-25 | Parascale, Inc. | Asynchronous file replication and migration in a storage network |
| CN103092698A (en) * | 2012-12-24 | 2013-05-08 | 中国科学院深圳先进技术研究院 | System and method of cloud computing application automatic deployment |
| CN104317658A (en) * | 2014-10-17 | 2015-01-28 | 华中科技大学 | MapReduce based load self-adaptive task scheduling method |
| CN104391930A (en) * | 2014-11-21 | 2015-03-04 | 用友软件股份有限公司 | Distributed file storage device and method |
| CN105786447A (en) * | 2014-12-26 | 2016-07-20 | 乐视网信息技术（北京）股份有限公司 | Method and apparatus for processing data by server and server |
| CN105357296A (en) * | 2015-10-30 | 2016-02-24 | 河海大学 | Elastic caching system based on Docker cloud platform |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109062658A (en) * | 2018-06-29 | 2018-12-21 | 优刻得科技股份有限公司 | Realize dispatching method, device, medium, equipment and the system of computing resource serviceization |
| CN109062658B (en) * | 2018-06-29 | 2021-06-04 | 优刻得科技股份有限公司 | Scheduling method, device, medium, equipment and system for realizing computing resource servitization |
| CN113497735A (en) * | 2020-04-08 | 2021-10-12 | 中兴通讯股份有限公司 | Open interface management method, electronic device and storage medium |
| CN111522624A (en) * | 2020-04-17 | 2020-08-11 | 成都安恒信息技术有限公司 | Message forwarding performance flexible extension system based on virtualization technology and extension method thereof |
| CN111522624B (en) * | 2020-04-17 | 2023-10-20 | 成都安恒信息技术有限公司 | Message forwarding performance elastic expansion system and expansion method based on virtualization technology |
| CN111290838A (en) * | 2020-05-09 | 2020-06-16 | 支付宝(杭州)信息技术有限公司 | Application access request processing method and device based on container cluster |
| WO2021227954A1 (en) * | 2020-05-09 | 2021-11-18 | 支付宝(杭州)信息技术有限公司 | Application access request processing based on container cluster |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9805140B2 (en) | Striping of directed graphs and nodes with improved functionality | |
| CN107786358A (en) | The expansion method of distributed system and the distributed system | |
| CN103098014B (en) | Storage system | |
| CN104067239B (en) | Systems and methods for data chunk deduplication | |
| US9489231B2 (en) | Selecting provisioning targets for new virtual machine instances | |
| US9686330B2 (en) | Topic-based messaging using consumer address and pool | |
| CN108734559B (en) | Order processing method and device | |
| US20120158858A1 (en) | Resource Optimization for Online Services | |
| CN106453146B (en) | Method, system, device and readable storage medium for allocating private cloud computing resources | |
| US8424010B2 (en) | Shared resource management | |
| CN106445676A (en) | Distributed data calculation-based task allocation method and task allocation apparatus | |
| CN109600798A (en) | Multi-domain resource allocation method and device in a kind of network slice | |
| KR20110060849A (en) | Method and system for data distribution in high performance computing cluster | |
| CN109408590B (en) | Method, device and equipment for expanding distributed database and storage medium | |
| CN110225074A (en) | A kind of communication packet dissemination system and distribution method based on device address domain | |
| CN108366082A (en) | Expansion method and flash chamber | |
| CN104639645A (en) | Method and device for balancing network load, and cluster service system | |
| CN105516325A (en) | Cloud load balancing method for carrying out elastic expansion and traffic distribution expansion according to application load | |
| CN105786447A (en) | Method and apparatus for processing data by server and server | |
| CN106570006A (en) | Multimedia file recommendation system and releasing information storage method | |
| KR101029416B1 (en) | Ranking data system, ranking query system and ranking computation method for computing large scale ranking in real time | |
| CN110178119B (en) | Method, device and storage system for processing service request | |
| CN105740278A (en) | Data processing method and apparatus | |
| CN112805684A (en) | Resource allocation using recovery borrowing | |
| CN109933433A (en) | A kind of GPU resource scheduling system and its dispatching method |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180309 |
|
| WW01 | Invention patent application withdrawn after publication |