CN104461375B - A kind of distribution method and device of disk task - Google Patents
A kind of distribution method and device of disk task Download PDFInfo
- Publication number
- CN104461375B CN104461375B CN201310459185.8A CN201310459185A CN104461375B CN 104461375 B CN104461375 B CN 104461375B CN 201310459185 A CN201310459185 A CN 201310459185A CN 104461375 B CN104461375 B CN 104461375B
- Authority
- CN
- China
- Prior art keywords
- disk
- store tasks
- recycling
- slow
- speed
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0604—Improving or facilitating administration, e.g. storage management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0614—Improving the reliability of storage systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0629—Configuration or reconfiguration of storage systems
- G06F3/0631—Configuration or reconfiguration of storage systems by allocating resources to storage systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention discloses a kind of distribution method and device of disk task, this method is：In system initial configuration phase, pending each store tasks are determined, wherein, each store tasks is used to record the total data that a recording arrangement obtains；It will be preserved on pending each store tasks equilibrium assignment to each disk.Due in initial configuration phase, a store tasks being assigned on each disk represent the total data that is obtained on a recording arrangement and independently of one another, so when one piece of disk failures, the automatically equilibrium of task in failed disk can be switched on normal disk by managing device, it ensure that the Future Data of each recording arrangement can continue to preserve, while also reduce the loss of data scope that failed disk is brought.
Description
Technical field
The present invention relates to disk array technology, the more particularly to a kind of distribution method and device of disk task.
Background technology
In protection and monitor field, video recording store function is one of most crucial function.In safety defense monitoring system, camera is every
When often carve and all producing audio, video data, therefore there is substantial amounts of Video data to need to store.Original enhancing Virtual File System
（Enhanced Virtual File System, ENVFS）Using Dynamic Load-Balancing Strategy, i.e., all recording arrangements
Data are dynamically distributed on each disk according to the file packing time, so as to fulfill the equilibrium of time data memory.For example,
Assuming that there is three recording arrangement A, B, C, three disks D, E, F, the file packing time are 30 minutes, each recording arrangement respectively with
Select a disk, such as recording arrangement A to have selected disk E, then recording arrangement A can record it for 30 minutes at first machine
Data be assigned on disk E, then by its second 30 minutes record data be assigned on disk F, by it the 3rd
The data recorded are assigned on disk D within a 30 minutes, and so on, the data on recording arrangement A every 30 minutes polls
It is allocated between these three disks, meanwhile, the data distribution of each recording arrangement is independent of each other.In this way, can be real
The equilibrium of the time data memory of existing disk.But due to being possible to be stored with the record of all recording arrangements on each disk
Be possible to storage in data processed, or even each disk is data recording of all recording arrangements in the same period, so
Once disk failures, can cause the loss of data in broad range, result even in the extreme circumstances in a period of time
The data recording of all recording arrangements is lost.
The content of the invention
The embodiment of the present invention provides a kind of distribution method and device of disk task, to solve disk event in the prior art
The problem of loss of data is in extensive range during barrier.
Concrete technical scheme provided in an embodiment of the present invention is as follows：
In a first aspect, a kind of distribution method of disk task, including：
In system initial configuration phase, pending each store tasks are determined, wherein, each store tasks is used to remember
The total data that one recording arrangement of record obtains；
It will be preserved on pending each store tasks equilibrium assignment to each disk.
Distribution method in this way, ensure that a task is relatively fixed on one disk, and the number between disk
According to keeping separate, in the case of disk failure, the minority being assigned to before affecting only failure in failed disk is recorded
The equipment of equipment, reduces the coverage that disk failure is brought.
With reference to first aspect, in the first possible implementation, further comprise：
In the system operation stage, according to the cycle is set, the recycling speed of each disk is counted, when the fastest magnetic of recycling
When disk and the recycling speed difference recycled between the most slow disk of speed exceed first threshold, it will recycle on fastest disk
A store tasks distribute to the most slow disk of recycling speed.
Mode in this way, recycles load of the speed to disk according to disk and carries out equilibrium, realize each disk
Equilibrium of the task data in storage time, improves the utilization rate and operational efficiency of system entirety.
With reference to first aspect, in second of possible implementation, further comprise：
Monitor the store tasks performed on each disk；
When determining any disk failures, by each store tasks equilibrium assignment performed on any disk to its
On his normal disk.
Implementation in this way, the task balance in failed disk is rapidly switched on normal disk, it is ensured that
The Future Data of each recording arrangement remains to preserve in time.
Second of possible implementation with reference to first aspect, in the third possible implementation, described will appoint
After on each store tasks equilibrium assignment performed on one disk to other normal disks, continue to being performed on each normal disk
Store tasks be monitored.
In this way, being always maintained at monitoring to disk, the security performance of system entirety is improved.
The possible implementation of with reference to first aspect the first, in the 4th kind of possible implementation, calculates recycling
During recycling speed difference between fastest disk and the most slow disk of recycling speed, using the following formula：
In formula, M1 is that the maximum in the recovered frequency increment of each disk, M2 are between this statistics and last time statistics
Minimum value between this statistics and last time statistics in the recovered frequency increment of each disk.
The possible implementation of with reference to first aspect the first, in the 5th kind of possible implementation, recycles speed
A store tasks on most fast disk are distributed to the most slow disk of recycling speed, including：
Whether the number for the store tasks for judging to carry on the most slow disk of the recycling speed reaches upper limit value, determines
Not up to the upper limit value when, the store tasks recycled on fastest disk are distributed to the most slow magnetic of recycling speed
On disk.
In this way, avoiding the erasable too fast situation of a certain piece of disk task data, the entirety of whole Array Disk is improved
Performance.
Any one above-mentioned implementation with reference to first aspect, in the 6th kind of possible implementation, including：
When determining the new disk of system configuration, a store tasks are distributed from other each disks respectively to described new
Disk on.
In this way, by dynamically increasing new building, management and dimension of the artificial intervention to RAIDC disk array groups instead of
Shield, improves the maintainability of system.
The 6th kind of possible implementation with reference to first aspect, in the 7th kind of possible implementation, from a magnetic
Distributed on disk on a store tasks to the new disk, including：
Judge whether one store tasks are last task on one disk, determine it is not last
During a task, one store tasks are distributed to the new disk.
In this way, avoid the situation that last task on disk is also allocated to new disk, balanced system is overall
Performance.
Second aspect, a kind of distributor of disk task, including：
Dispensing unit, in system initial configuration phase, determining pending each store tasks, wherein, each
Store tasks are used to record the total data that a recording arrangement obtains；By pending each store tasks equilibrium assignment to each
Preserved on a disk.
By distribution method as dispensing unit, it ensure that a task is relatively fixed on one disk, and magnetic
Data between disk keep separate, in the case of disk failure, are assigned to before affecting only failure in failed disk
The equipment of a small number of recording arrangements, reduces the coverage that disk failure is brought.
With reference to second aspect, in the first possible implementation, further comprise：
Balanced unit, in the system operation stage, according to the setting cycle, counts the recycling speed of each disk, when returning
When the recycling speed difference received fastest disk and recycled between the most slow disk of speed exceedes first threshold, by recycling speed
A store tasks on the most fast disk of degree are distributed to the most slow disk of recycling speed.
Load of the speed to disk is recycled according to disk by balanced unit and carries out equilibrium, realizes the task of each disk
Equilibrium of the data in storage time, improves the utilization rate and operational efficiency of system entirety.
With reference to second aspect, in second of possible implementation, further comprise：
Monitoring unit, for monitoring the store tasks performed on each disk, when determining any disk failures, by institute
State on each store tasks equilibrium assignment to other the normal disks performed on any disk.
The task balance in failed disk is rapidly switched on normal disk by monitoring unit, it is ensured that Ge Gelu
The Future Data of control equipment remains to preserve in time.
With reference to second of possible implementation of second aspect, in the third possible implementation, the monitoring
Unit after on each store tasks equilibrium assignment to other the normal disks performed on any disk, will continue to it is each just
The store tasks performed on normal disk are monitored.
Monitoring unit can be always maintained at monitoring to each disk, so improve the security of system entirety.
With reference to the first possible implementation of second aspect, in the 4th kind of possible implementation, the equilibrium
When unit calculates the recycling speed difference for recycling fastest disk and recycling between the most slow disk of speed, using following public affairs
Formula：
In formula, M1 is that the maximum in the recovered frequency increment of each disk, M2 are between this statistics and last time statistics
Minimum value between this statistics and last time statistics in the recovered frequency increment of each disk.
With reference to the first possible implementation of second aspect, in the 5th kind of possible implementation, the equilibrium
Unit is specifically used for distributing the store tasks recycled on fastest disk to the most slow disk of recycling speed
Before, whether the number for the store tasks for judging to carry on the most slow disk of the recycling speed reaches upper limit value, determines not
When reaching the upper limit value, the balanced unit distributes the store tasks recycled on fastest disk to recycling speed
On the most slow disk of degree.
In this way, avoiding the erasable too fast situation of a certain piece of disk task data, the entirety of whole Array Disk is improved
Performance.
With reference to any one above-mentioned implementation of second aspect, in the 6th kind of possible implementation, further wrap
Include：
Allocation unit, during for determining the new disk of system configuration, distributing one from other each disks respectively and depositing
On storage task to the new disk.
In this way, dynamically increasing new building by allocation unit, pipe of the artificial intervention to RAIDC disk array groups instead of
Reason and maintenance, improve the maintainability of system.
With reference to the 6th kind of possible implementation of second aspect, in the 7th kind of possible implementation, the distribution
Unit is specifically used for before being distributed from a disk on a store tasks to the new disk, judges one deposit
Whether storage task is last task on one disk, is determined when not being last task, the allocation unit
One store tasks are distributed to the new disk.
In this way, avoid the situation that last task on disk is also allocated to new disk, balanced system is overall
Performance.
Brief description of the drawings
Fig. 1 is the flow chart that disk task is distributed in the embodiment of the present invention；
Fig. 2 is that disk recycles schematic diagram in the embodiment of the present invention；
Fig. 3 is the distributor figure of disk task in the embodiment of the present invention.
Embodiment
In order to reduce the loss of data scope that disk failure is brought, in the embodiment of the present invention, there is provided one kind is based on circulation
Queue file's system（Circular Queue File System, CQFS）Redundant array of inexpensive disk technology（Redundant
Array of Independent Disks on CQFS, RAIDC）.
The preferred embodiment of the present invention is described in detail below in conjunction with the accompanying drawings.
As shown in fig.1, in the embodiment of the present invention, the idiographic flow of managing device distribution disk task is as follows：
Step 100：In system initial configuration phase, managing device determines pending each store tasks, wherein, it is each
A store tasks are used to record the total data that a recording arrangement obtains.
In the embodiment of the present invention, managing device can be user terminal, such as PC, notebook or
Server, this is no longer going to repeat them.
Step 110：Managing device will be preserved on pending each store tasks equilibrium assignment to each disk.
Specifically, in the embodiment of the present invention, each store tasks are polled distribution by managing device between disk, i.e.,
One store tasks is distributed into first disk, next store tasks are then distributed into second disk, and so on.
Further, the recycling speed of each disk is counted according to the cycle is set in system operation stage, managing device,
When the recycling speed difference for recycling fastest disk and recycling between the most slow disk of speed exceedes first threshold, will return
On fastest disk store tasks is received to distribute to the most slow disk of recycling speed.
In the present embodiment, managing device can monitor the store tasks performed on each disk；When definite any disk occurs
During failure, just by each store tasks equilibrium assignment to other the normal disks performed on any disk.Then, management dress
Put and continue to be monitored the store tasks performed on each normal disk.
Due to representing a recording arrangement in initial configuration phase, a store tasks being assigned on each disk
The total data of upper acquisition and independently of one another, so when one piece of disk failures, managing device can be by failed disk
Task automatically equilibrium is switched on normal disk, it is ensured that and the Future Data of each recording arrangement can continue to preserve,
Also reduce the loss of data scope that failed disk is brought at the same time.Further, since RAIDC technologies are using regularly balanced plan
Slightly, i.e., the recycling speed of each disk is periodically calculated, when the difference of most fast and most slow recycling speed exceedes first threshold
When then carry out task distribution adjustment, by a task from distribution is recycled on fastest disk to the most slow disk of recycling speed
On, after settling out, all data of the task are still saved on one disk, can't be jumped on other disks, real
The equilibrium of the time data memory of multiple disks is showed.
In the embodiment of the present invention, on each disk, managing device is each to perform in a manner of CQFS file system
A store tasks, with a disk（Hereinafter referred to as disk A）Exemplified by, e.g., corresponding disk A creates a CQFS file system,
This document system includes a series of CQFS data blocks of default sizes and quantity, and each CQFS data blocks are added into a circulation
In queue, wherein, height is from complete between the data that each CQFS data blocks are stored, i.e., between any two CQFS data blocks
There is no data dependence.The data of generation are write each above-mentioned CQFS data block by system operation stage, managing device one by one
It is interior, wherein, often corresponding data block recovery time stamp is generated using a CQFS data block, the corresponding CQFS data blocks used.
When clearing data, order of the managing device according to the corresponding data block recovery time stamp of each CQFS data blocks from small to large, according to
The secondary data removed in each CQFS data block, wherein, the data in a CQFS data block are often removed, after clearing data
CQFS data blocks in continue to write to the data being newly generated.Mode in this way, realizes the cycle applications of CQFS data blocks.
In the load balancing between realizing disk, managing device needs further to count the recycling speed of each disk, with
Just the number of the store tasks on each disk is adjusted according to recycling speed.
The recycling speed of disk is calculated, the recovered frequency of disk is first calculated, formula can be passed through（1）Obtain：
Formula（1）In, CQFS recycling queue tail recovery time stamps are used to indicate the recycling of last time CQFS data blocks
Time, CQFS block numbers represent the block number of data Division in CQFS data blocks.For example, CQFS data data field quilt in the block
10 data blocks are divided into, then CQFS block numbers are 10；In the system operation stage, data are written into this 10 CQFS data one by one
Block, the corresponding data block recovery time stamp of this 10 CQFS data blocks generations are followed successively by 0 to 9, wherein, recovery time stamp is 0
Data block correspond to the afterbody of round-robin queue, recovery time stamp corresponds to the head of round-robin queue, CQFS recovery forces for 9 data block
Row afterbody recovery time stamp is the maximum in the corresponding recovery time stamp of this 10 CQFS data blocks, i.e. round-robin queue head
The corresponding timestamp of data block（I.e. recovery time stamp is 9）, according to formula（1）It can obtain, the initial value of the recovered frequency of disk
For 0.When disk is once recycled, according to the order of recovery time stamp from small to large, the recovery time for removing minimum stabs（Return
It is 0 to receive timestamp）Corresponding CQFS data data in the block, and new data are write in the CQFS data blocks, write number
After the completion of, which is added to the head of round-robin queue, CQFS recycling of its recovery time stamp in last time statistics
Add 1 on the basis of queue tail recovery time stamp, i.e. the recovery time stamp of the CQFS data blocks is changed into 10；So disk every time
Receive once, CQFS recycling queue tail recovery time stamps are increased by 1.
For example, refer to Fig. 2, it is assumed that 10 CQFS data blocks are shared in the round-robin queue of a CQFS file system, initially
Stage, toward this 10 CQFS data blocks in when writing data one by one, this 10 CQFS data blocks are corresponding in turn to generation from 0 to 9
Data block recovery time stabs, wherein, the corresponding recovery time stamp of CQFS data blocks on round-robin queue head is 9, round-robin queue's tail
The corresponding recovery time stamp of CQFS data blocks in portion is 0, i.e. the corresponding recovery time of CQFS data blocks stabs bigger, the CQFS data
Data in block are newer；The corresponding recovery time stamp of CQFS data blocks is smaller, and the data in the CQFS data blocks are older.Work as disk
When once being recycled, managing device selects data oldest in this 10 CQFS data blocks to be purged, that is, selects corresponding
Recovery time stamp is 0 CQFS data blocks, removes data therein and writes new data, after the completion of, by the CQFS data blocks
The head of round-robin queue is added to, while CQFS recycling queue tail recovery time of its recovery time stamp in last time statistics is stabbed
Add 1 on the basis of 9, save as 10, this is this CQFS recycling queue tail recovery time stamps.And so on, disk is every
Once recycled, CQFS recycling queue tail recovery time stamps are increased by 1.
Pass through formula（1）Disk this disk recovered frequency statistical value and the disk recovered frequency of last time can be calculated
Statistical value, and determine disk recovered frequency increment according to the two values, pass through formula（2）Obtain：
This statistical value-last time statistical value formula (2)
According to each disk recovered frequency increment, pass through formula（3）It can determine in disk recovered frequency increment most
Big value, passes through formula（4）It can determine the minimum value in disk recovered frequency increment：
Max (Δ 1, Δ 2 ...) formula (3)
Min (Δ 1, Δ 2 ...) formula (4)
In formula (3) and formula (4), Δ 1 represents the disk recovered frequency increment of first piece of disk, and Δ 2 represents second piece
The disk recovered frequency increment ... of disk and so on.
When calculating the recycling speed difference for recycling fastest disk and recycling between the most slow disk of speed, it can use
Formula（5）Obtain：
Formula（5）In, M1 is the maximum in the recovered frequency increment of each disk between this statistics and last time statistics,
M2 is the minimum value in the recovered frequency increment of each disk between this statistics and last time statistics.
Managing device distributes the store tasks recycled on fastest disk to the most slow disk of recycling speed
, it is necessary to which whether the number for judging to recycle the store tasks that can be carried on the most slow disk of speed reaches upper limit value when upper, determined not
When reaching the upper limit value, the store tasks recycled on fastest disk are distributed to the most slow disk of recycling speed
On, however, it is determined that reach upper limit value, then cancel distribution.
On the other hand, when the new disk of system configuration, managing device distributes one from other each disks respectively
Store tasks are on the new disk.Specifically, when distributing from a disk on a store tasks to the new disk,
Managing device needs first to judge whether the store tasks are last task on the disk, determines it is not last task
When, just the store tasks are assigned on new disk, however, it is determined that be last task, then cancel distribution.
Based on above-described embodiment, as shown in fig.3, in the embodiment of the present invention, the distributor 30 of disk task includes：
Dispensing unit 301, in system initial configuration phase, determining pending each store tasks, wherein, it is each
A store tasks are used to record the total data that a recording arrangement obtains；Then by pending each store tasks equilibrium point
It is assigned on each disk and is preserved.
Further, can also include：
Balanced unit 302, in the system operation stage, according to the setting cycle, counts the recycling speed of each disk, when
When the recycling speed difference for recycling fastest disk and recycling between the most slow disk of speed exceedes first threshold, it will recycle
A store tasks on fastest disk are distributed to the most slow disk of recycling speed.
In the present embodiment, balanced unit 302, which calculates, to be recycled between fastest disk and the most slow disk of recycling speed
Recycling speed difference when, using the following formula：
In formula, M1 is that the maximum in the recovered frequency increment of each disk, M2 are between this statistics and last time statistics
Minimum value between this statistics and last time statistics in the recovered frequency increment of each disk.
Specifically, balanced unit 302 is distributed to recycling speed by the store tasks recycled on fastest disk
Before on the most slow disk of degree, whether the number for the store tasks for first judging to carry on the most slow disk of the recycling speed has reached
Upper limit value, when determining not up to upper limit value, 302 ability of balanced unit divides the store tasks recycled on fastest disk
It is assigned on the most slow disk of recycling speed.
Further, can also include：
Monitoring unit 303, for monitoring the store tasks performed on each disk, when determining any disk failures,
Just by each store tasks equilibrium assignment to other the normal disks performed on any disk.
In the present embodiment, monitoring unit 303 is balanced by each store tasks performed on any disk that this breaks down
After on distribution to other normal disks, it will continue to be monitored the store tasks performed on each normal disk.
Further, can also include：
Allocation unit 304, during for determining the new disk of system configuration, distributes one from other each disks respectively
On store tasks to the new disk.
Specifically, allocation unit 304 is first sentenced before being distributed from a disk on a store tasks to new disk
Whether the store tasks of breaking are last task on the disk, determine that when not being last task, allocation unit 304 is
The store tasks are assigned on new disk.
In conclusion using technical solution described in the embodiment of the present invention, the task data of each disk is realized
Equilibrium in storage time, at the same time, it is ensured that a task is relatively fixed on one disk；And work as disk failures
When, the task balance in failed disk can be rapidly switched on normal disk, the Future Data of all tasks will not be lost,
I.e. each recording arrangement caused task data after disk failures remains to preserve in time.Therefore in disk failure
In the case of, system only lost the stored data in failed disk, is assigned to before affecting only failure in failed disk
A small number of recording arrangements data, reduce the coverage that disk failure is brought.
Obviously, those skilled in the art can carry out the embodiment of the present invention various modification and variations without departing from this hair
The spirit and scope of bright embodiment.In this way, if these modifications and variations of the embodiment of the present invention belong to the claims in the present invention
And its within the scope of equivalent technologies, then the present invention is also intended to comprising including these modification and variations.
Claims (14)
- A kind of 1. distribution method of disk task, it is characterised in that including：In system initial configuration phase, pending each store tasks are determined, wherein, each store tasks is used to record one The total data that a recording arrangement obtains；It will be preserved on pending each store tasks equilibrium assignment to each disk；In the system operation stage, according to the setting cycle, count the recycling speed of each disk, when recycle fastest disk with When recycling speed difference between the most slow disk of recycling speed exceedes first threshold, one on fastest disk will be recycled A store tasks are distributed to the most slow disk of recycling speed.
- 2. the method as described in claim 1, it is characterised in that further comprise：Monitor the store tasks performed on each disk；When determining any disk failures, by each store tasks equilibrium assignment performed on any disk to other just On normal disk.
- 3. method as claimed in claim 2, it is characterised in that each store tasks performed on any disk are balanced After on distribution to other normal disks, continue to be monitored the store tasks performed on each normal disk.
- 4. the method as described in claim 1, it is characterised in that calculate the fastest disk of recycling and recycling speed is most slow During recycling speed difference between disk, using the following formula：In formula, M1 is the maximum in the recovered frequency increment of each disk between this statistics and last time statistics, and M2 is this Minimum value between statistics and last time statistics in the recovered frequency increment of each disk.
- 5. the method as described in claim 1, it is characterised in that by the store tasks recycled on fastest disk point It is assigned on the most slow disk of recycling speed, including：Whether the number for the store tasks for judging to carry on the most slow disk of the recycling speed reaches upper limit value, determines not reach During to the upper limit value, the store tasks recycled on fastest disk are distributed to the most slow disk of recycling speed On.
- 6. such as claim 1-5 any one of them methods, it is characterised in that including：When determining the new disk of system configuration, a store tasks are distributed from other each disks respectively to the new magnetic On disk.
- 7. method as claimed in claim 6 a, it is characterised in that store tasks are distributed from a disk to described new On disk, including：Judge whether one store tasks are last task on one disk, determine it is not last During business, one store tasks are distributed to the new disk.
- A kind of 8. distributor of disk task, it is characterised in that including：Dispensing unit, in system initial configuration phase, determining pending each store tasks, wherein, each storage Task is used to record the total data that a recording arrangement obtains；By pending each store tasks equilibrium assignment to each magnetic Preserved on disk；Balanced unit, in the system operation stage, according to the setting cycle, counts the recycling speed of each disk, when recycling speed When the most fast disk of degree and the recycling speed difference recycled between the most slow disk of speed exceed first threshold, speed will be recycled most A store tasks on fast disk are distributed to the most slow disk of recycling speed.
- 9. device as claimed in claim 8, it is characterised in that further comprise：Monitoring unit, for monitoring the store tasks performed on each disk, when determining any disk failures, described will appoint On each store tasks equilibrium assignment performed on one disk to other normal disks.
- 10. device as claimed in claim 9, it is characterised in that the monitoring unit will perform each on any disk After on a store tasks equilibrium assignment to other normal disks, continue to supervise the store tasks performed on each normal disk Survey.
- 11. device as claimed in claim 8, it is characterised in that the balanced unit calculate recycle fastest disk with When recycling the recycling speed difference between the most slow disk of speed, using the following formula：In formula, M1 is the maximum in the recovered frequency increment of each disk between this statistics and last time statistics, and M2 is this Minimum value between statistics and last time statistics in the recovered frequency increment of each disk.
- 12. device as claimed in claim 8, it is characterised in that the balanced unit is specifically used for, fastest that will recycle Disk on a store tasks distribute to the most slow disk of recycling speed before, judge described to recycle the most slow magnetic of speed Whether the number for the store tasks that can be carried on disk reaches upper limit value, when determining the not up to upper limit value, the balanced unit The store tasks recycled on fastest disk are distributed to the most slow disk of recycling speed.
- 13. such as claim 8-12 any one of them devices, it is characterised in that further comprise：Allocation unit, during for determining the new disk of system configuration, a storage is distributed from other each disks and is appointed respectively On business to the new disk.
- 14. device as claimed in claim 13, it is characterised in that the allocation unit is specifically used for, from a disk Before distributing on a store tasks to the new disk, judge whether one store tasks are on one disk Last task, determine when not being last task, the allocation unit distributes one store tasks to institute State on new disk.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310459185.8A CN104461375B (en) | 2013-09-24 | 2013-09-24 | A kind of distribution method and device of disk task |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310459185.8A CN104461375B (en) | 2013-09-24 | 2013-09-24 | A kind of distribution method and device of disk task |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104461375A CN104461375A (en) | 2015-03-25 |
| CN104461375B true CN104461375B (en) | 2018-05-01 |
Family
ID=52907506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310459185.8A Active CN104461375B (en) | 2013-09-24 | 2013-09-24 | A kind of distribution method and device of disk task |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104461375B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101198046A (en) * | 2007-12-21 | 2008-06-11 | 北京天天宽广网络科技有限公司 | Contents distribution and memory system and method in video network based on P2P technology |
| CN102012946A (en) * | 2010-12-22 | 2011-04-13 | 武汉上博科技有限公司 | High-efficiency safety monitoring video/image data storage method |
| CN102333225A (en) * | 2010-07-14 | 2012-01-25 | 北京同步科技有限公司 | System and method for transmitting and recording four channels of standard definition video signals |
| CN102760136A (en) * | 2011-04-29 | 2012-10-31 | 苏州科达科技有限公司 | File system control method based on rapid video data storage |
| CN103092528A (en) * | 2011-11-01 | 2013-05-08 | 新太科技股份有限公司 | Method of space storage conversion according to disk space and performance changes |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040254940A1 (en) * | 2003-01-31 | 2004-12-16 | Brush Hector Cesar | Digital media distribution method and system |
-
2013
- 2013-09-24 CN CN201310459185.8A patent/CN104461375B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101198046A (en) * | 2007-12-21 | 2008-06-11 | 北京天天宽广网络科技有限公司 | Contents distribution and memory system and method in video network based on P2P technology |
| CN102333225A (en) * | 2010-07-14 | 2012-01-25 | 北京同步科技有限公司 | System and method for transmitting and recording four channels of standard definition video signals |
| CN102012946A (en) * | 2010-12-22 | 2011-04-13 | 武汉上博科技有限公司 | High-efficiency safety monitoring video/image data storage method |
| CN102760136A (en) * | 2011-04-29 | 2012-10-31 | 苏州科达科技有限公司 | File system control method based on rapid video data storage |
| CN103092528A (en) * | 2011-11-01 | 2013-05-08 | 新太科技股份有限公司 | Method of space storage conversion according to disk space and performance changes |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104461375A (en) | 2015-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11137940B2 (en) | Storage system and control method thereof | |
| CN102511030B (en) | Computer system and method for controlling same | |
| KR101502896B1 (en) | Distributed memory cluster control apparatus and method using map reduce | |
| CN109684074B (en) | Physical machine resource allocation method and terminal equipment | |
| US10884667B2 (en) | Storage controller and IO request processing method | |
| CN103051691B (en) | Partition allocation method, device and distributed memory system | |
| CN106233269A (en) | Fine granulation bandwidth supply in Memory Controller | |
| CN103384550B (en) | The method of storage data and device | |
| US20140237192A1 (en) | Method and apparatus for constructing memory access model | |
| CN110661824B (en) | Flow control method of server in distributed cluster and storage medium | |
| CN105306860A (en) | Video recording data storage method and device | |
| JP2010198056A (en) | Allocation control program and allocation control device | |
| CN103399781A (en) | Cloud server and virtual machine management method thereof | |
| CN103109283A (en) | Allocation of memory buffers in computing system with multiple memory channels | |
| CN105446792A (en) | Deployment method, deployment device and management node of virtual machines | |
| CN106293953B (en) | A kind of method and system of the shared display data of access | |
| CN104702654A (en) | Video cloud storage system based storage and extraction performance balancing method and device | |
| CN104461375B (en) | A kind of distribution method and device of disk task | |
| CN103761194B (en) | A kind of EMS memory management process and device | |
| CN109189739A (en) | Spatial cache recovery method and device | |
| CN105577756A (en) | Distributed database log collection and load regulation system adopting cross backup and method thereof | |
| US11003360B2 (en) | IO request processing according to processing sorting indexes | |
| CN106406771A (en) | Log recording method and log recorder | |
| CN103973744A (en) | Distributed file progressive storage technology | |
| CN113821165B (en) | Distributed cluster fusion storage method, system and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |