CN105491152A - A method for configuring a scheduler for cloud storage resources - Google Patents

Abstract

The invention relates to the technical field of cloud platform storage virtualization, in particular to a method for a cloud storage resource configurable scheduler. Firstly, setting a filter name list; setting a name list of weighing machines; loading a filter; loading a weigher; starting to create a storage volume; acquiring a candidate node; filtering the candidate nodes; weighing the candidate nodes; weighing and sorting; and finally, selecting the optimal node to create the storage volume. The method solves the problem of unreasonable distribution of newly-added cloud storage resources, and can be applied to cloud platform storage virtualization.

Description

A method for configuring a scheduler for cloud storage resources

technical field

The invention relates to the technical field of cloud platform storage virtualization, in particular to a method for configuring a scheduler of cloud storage resources.

Background technique

On a newly built cloud computing platform, it is usually necessary to create a large number of storage volumes to virtual machine system disks or cloud disks at the beginning, and cloud computing platforms usually provide automatic allocation of storage volumes, but automatic allocation is usually only based on a simple strategy To complete, this will bring the following problems:

One is that the allocation method does not meet user expectations. Users don't want to configure manually, but automatic configuration is not user-friendly enough.

Second, the automatic allocation requirements of different customers purchasing cloud computing platforms are not exactly the same, and the cloud platform does not support the combination of several commonly used business scenarios in a configurable way.

The third is that programmers maintain multiple sets of codes to meet business needs, which leads to low development efficiency and high maintenance costs.

In order to solve the above problems, there is a need for a method for configuring a scheduler for cloud storage resources. Users configure the required filters and scales according to their business needs, and the system scheduler automatically loads the corresponding filters and scales according to the configuration files, and selects the best candidate node to create storage volumes when automatically allocating storage volumes.

Contents of the invention

The technical problem solved by the present invention is to provide a method for a configurable scheduler of cloud storage resources, which solves the problem that the automatic allocation method is not humanized enough, the cloud platform does not support configurable ways to combine several commonly used business scenarios, and programmers maintain multiple sets When the code is used to meet business needs, the development efficiency is low and the maintenance cost is high.

The technical scheme that the present invention solves the problems of the technologies described above is:

Described method comprises the steps:

Step 1: Set the list of filter names;

Step 2: Set the weighing device name list;

Step 3: Load the filter;

Step 4: Load the scale;

Step 5: Start creating storage volumes;

Step 6: Obtain candidate nodes;

Step 7: filter candidate nodes;

Step 8: Candidate nodes are weighed;

Step 9: Weighing and sorting;

Step 10: Select the best node to create a storage volume.

The above-mentioned setting filter name list, set the name list of each filter on the configuration file;

Said setting the name list of the weighing device, the name list of each weighing device is set on the configuration file;

The loading filter loads the corresponding filter according to the name list of each filter set on the configuration file, and the filter can be 0 or more, configured by the user;

The loading of the weighing device loads the corresponding weighing device according to the name list of each weighing device set in the configuration file. There can be zero or more weighing devices configured by the user.

When starting to create a storage volume, the user can specify the destination node or be automatically allocated by the system scheduler; before the automatic allocation, the user can add some filtering conditions such as storage type;

The system scheduler is composed of core components such as filters and weighing devices.

For the acquisition of candidate nodes, select nodes whose status is active on all storage nodes, and storage nodes that have been shut down cannot be used as candidate nodes;

The filter candidate nodes mentioned above, all the loaded filter filter candidate nodes, only the nodes that meet all the filter requirements can enter the next step, otherwise they will be removed from the candidate node set; the filtering methods include filtering through specified nodes and storing Filter by type, filter by remaining resources.

The candidate nodes are weighed, and each weigher weighs the candidate nodes in order; that is, each weigher calculates a weight score for the candidate node, so each candidate node weight total score=the score of weigher 1 + Score of Scale 2 + ... + Score of Scale N;

The weigher is specially used to calculate the weight of candidate nodes in certain aspects.

The weighing sorting is sorted according to the total weight score of the candidate nodes, sorted from high to low, and the highest score is at the front of the queue;

Select the best node to start creating storage volumes, select the candidate node with the highest weight score (i.e. the front of the queue) as the best node, send a message to this node to create a storage volume, and the storage volume can be used as a virtual machine system disk or cloud disk .

The present invention uses a configurable scheduler method for cloud storage resources. When the allocation method does not meet user expectations, the cloud platform does not support configurable ways to combine several commonly used business scenarios, and programmers maintain multiple sets of codes to meet business needs. Low development efficiency and high maintenance costs. The present invention supports self-defining combined filters and weighing devices, and makes the automatic allocation and creation of storage volumes more intelligent and closer to the needs of users by calculating the weight and weighing the total score. The present invention is specially aimed at the field of cloud storage, and is different from other resource schedulers. The present invention can manually configure the combination of the filter and the weighing device, and distinguishes the scheduler which cannot be manually configured generally. The present invention can freely combine different filters and weighing devices, and is different from other schedulers that need to maintain multiple versions. The invention is a technical framework that supports rapid development of new filters or weighing devices, which is different from the low efficiency of developing filters or weighing devices by other schedulers.

Description of drawings

Below in conjunction with accompanying drawing, the present invention is further described:

Fig. 1 is a flowchart;

Fig. 2 is a logical structure diagram of the present invention;

detailed description

There are multiple implementations of the present invention, and one of the implementation methods is described here by taking the cloud platform as an example. The flow chart is shown in Figure 1, and the specific implementation process is as follows:

1. Set the list of filter names. The user manually modifies the configuration file and configures the name list of the filter.

2. Set the weighing device name list. The user manually modifies the configuration file and configures the name list of the scale.

3. Load the filter, the code is as follows:

/**

* Load filter.

*Path where the parampath configuration file is located

*return boolean returns the result

*/

public booleanloadFieldters(Stringpath){

return DispatbService.loadFieldters (path);

}

4. Load the scale, the code is as follows:

/**

＊Load the scale.

*Path where the parampath configuration file is located

*return boolean returns the result

*/

public booleanloadWeights(Stringpath){

return DispathService.loadWeights (path);

}

5. Start to create a storage volume, the code is as follows:

/**

＊Start creating storage volume.

*paramname volume name

*paramsize volume size

*return boolean returns the result

*/

public boolean createVolume(Stringname, intsize) {

return VolumeService.createVolume (name, size);

}

6. Obtain candidate nodes, the code is as follows:

/**

＊Get candidate nodes.

＊returnList<candidateHost> returns the result object

*/

public boolean getCandidates(){

return DispathService.getCandidates ();

}

7. Filter candidate nodes, the code is as follows:

/**

* Filter candidate nodes.

* paramchs candidate node object

*returnResult returns the result object

*/

publicList< CandidateHost >getFilteredHosts(List< CandidateHost >chs){

return DispathService.getFilteredHosts (chs);

}

8. Candidate nodes weigh, the code is as follows:

/**

＊Candidate nodes are weighed.

* paramchs candidate node object

*returnResult returns the result object

*/

publicList< CandidateHost >getWeightHosts(List< CandidateHost >chs){

return DispathService.getWeightHosts (chs);

}

9. Weighing and sorting, the code is as follows:

/**

* Weighing order.

* paramchs candidate node object

＊returnList<candidateHost> returns the result object

*/

publicList< CandidateHost >choose_host_weighers(List< CandidateHost >chs){

return DispathService.choose_host_weighers (chs);

}

10. Select the best node to create a storage volume, the code is as follows:

/**

＊Choose the best node to create storage volume.

* paramchs candidate node object

＊returnCandidateHost returns the result object

*/

public CandidateHost getTop(List< CandidateHost >chs){

returns. first();

}

The whole process is over.

Claims (7)

1. A method for a cloud storage resource configurable scheduler, characterized in that: the method comprises the steps of: Step 1: Set the list of filter names; Step 2: Set the weighing device name list; Step 3: Load the filter; Step 4: Load the scale; Step 5: Start creating storage volumes; Step 6: Obtain candidate nodes; Step 7: filter candidate nodes; Step 8: Candidate nodes are weighed; Step 9: Weighing and sorting; Step 10: Select the best node to create a storage volume. 2. method according to claim 1, is characterized in that: described setting filter name list, the name list of each filter is set on configuration file; Said setting the name list of the weighing device, the name list of each weighing device is set on the configuration file; The loading filter loads the corresponding filter according to the name list of each filter set on the configuration file, and the filter can be 0 or more, configured by the user; The loading of the weighing device loads the corresponding weighing device according to the name list of each weighing device set in the configuration file. There can be zero or more weighing devices configured by the user. 3. The method according to claim 1, characterized in that: when starting to create a storage volume, the user can specify the destination node or be automatically allocated by the system scheduler when the user starts to create the storage volume; before the automatic allocation, the user can add some filter conditions such as storage type; The system scheduler is composed of core components such as filters and weighing devices. 4. The method according to claim 2, characterized in that: when starting to create a storage volume, the user can specify the destination node or be automatically allocated by the system scheduler when the user starts to create the storage volume; before the automatic allocation, the user can add some filter conditions such as storage type; The system scheduler is composed of core components such as filters and weighing devices. 5. The method according to any one of claims 1 to 4, characterized in that: the acquisition of candidate nodes selects nodes whose status is active on all storage nodes, and storage nodes that have been shut down cannot be used as candidate nodes; The filter candidate nodes mentioned above, all the loaded filter filter candidate nodes, only the nodes that meet all the filter requirements can enter the next step, otherwise they will be removed from the candidate node set; the filtering methods include filtering through specified nodes and storing Filter by type, filter by remaining resources. 6. The method according to any one of claims 1 to 4, characterized in that: the candidate nodes are weighed, and each weigher weighs the candidate nodes in order; that is, each weigher weighs the candidate nodes The node calculates the weight score, then the total weight score of each candidate node = score of weigher 1 + score of weigher 2 + ... + score of weigher N; The weigher is specially used to calculate the weight of a candidate node in a certain aspect. The weighing sorting is sorted according to the total weight score of the candidate nodes, sorted from high to low, and the highest score is at the front of the queue; Select the best node to start creating storage volumes, select the candidate node with the highest weight score (i.e. the front of the queue) as the best node, send a message to this node to create a storage volume, and the storage volume can be used as a virtual machine system disk or cloud disk . 7. The method according to claim 5, characterized in that: the candidate nodes are weighed, and each weigher weighs the candidate nodes in order; that is, each weigher calculates a weight score for the candidate nodes, Then the total weight score of each candidate node = score of weigher 1 + score of weigher 2 + ... + score of weigher N; The weigher is specially used to calculate the weight of a candidate node in a certain aspect. The weighing sorting is sorted according to the total weight score of the candidate nodes, sorted from high to low, and the highest score is at the front of the queue; Select the best node to start creating storage volumes, select the candidate node with the highest weight score (i.e. the front of the queue) as the best node, send a message to this node to create a storage volume, and the storage volume can be used as a virtual machine system disk or cloud disk .