CN104391764B - A kind of MATLAB software method and system - Google Patents
A kind of MATLAB software method and system Download PDFInfo
- Publication number
- CN104391764B CN104391764B CN201410632804.3A CN201410632804A CN104391764B CN 104391764 B CN104391764 B CN 104391764B CN 201410632804 A CN201410632804 A CN 201410632804A CN 104391764 B CN104391764 B CN 104391764B
- Authority
- CN
- China
- Prior art keywords
- host apparatus
- virtual machine
- equipment
- stand
- guest virtual
- 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
Abstract
The present invention provides a kind of MATLAB software method and system.For the MATLAB software method applied in the MATLAB software system for including host apparatus and stand-by equipment, the host apparatus and stand-by equipment all include the guest virtual machine established on virtualization kernel, and the MATLAB software method includes：The synchronous trigger condition of default virtual memory；When the trigger condition occurs, it is synchronous to complete once the virtual memory of the host apparatus and the stand-by equipment；When the host apparatus breaks down, stop the operation of the guest virtual machine of the host apparatus, and start the operation of the guest virtual machine of the stand-by equipment.Technical scheme can trouble saving shut down risk, when host apparatus breaks down, business will directly migrate to stand-by equipment, higher without any loss of data, reliability using continuous uninterrupted；And the technical scheme is more extensive independently of operating system layer, application.
Description
Technical field
The present invention relates to a kind of computer safety system, more particularly to a kind of MATLAB software method and system.
Background technology
With the fast development of Information technology, especially Internet of Things and the rise of cloud computing, computer system passes through to society
The influence on development of Ji is increasing.Computer system is not applied only to the management and maintenance of data message, and is also employed in work
Industry controls to be performed with production.Such as manufacturing industry Process Control System, factory's manufacturing execution system, enterprise network data center systems, meter
Calculation machine system becomes the core of industrial IT infrastructure.The part of computer system may break down, if computer
The component malfunction of system, the operation of computer system can be influenceed unavoidably, even result in computer system collapse.So, such as
What ensure computer system when breaking down, computer system can normal work or recover normal work, just become meter
One key issue of calculation machine System Development.
For the safety and stability problem of computer system caused by computer system failure, there has been proposed many solution party
Case, include the MSCS failover clustering schemes of Microsoft, RoseHA clustered software schemes, Vmware virtualization HA schemes
Deng.But the mentality of designing of these solutions is the failover using master-slave equipment, realize main frame break down after client
The fast quick-recovery of machine system.When i.e. main frame breaks down, slave is taken over and restarts crucial application service.But this design has obvious
The defects of, switching is not only restarted between principal and subordinate when application service to cause operation system to interrupt, and the system is often
It is complicated, there is very big limitation to software systems and application environment.
Can be when computer system break down in consideration of it, how to find one kind, the service operation of computer system
Continuous unbroken MATLAB software method and system are can guarantee that, just become those skilled in the art's urgent problem to be solved.
The content of the invention
In view of the above the shortcomings that prior art, it is an object of the invention to provide a kind of MATLAB software method and it is
System, for solving in the prior art when computer system breaks down, the problem of service operation can be interrupted.
In order to achieve the above objects and other related objects, the present invention provides a kind of MATLAB software method, applied to including
In the MATLAB software system of host apparatus and stand-by equipment, the host apparatus and stand-by equipment all include establishing virtualizing
Guest virtual machine on kernel, the fault-tolerance approach synchronous based on virtual memory include：The synchronous triggering of default virtual memory
Condition；When the trigger condition occurs, it is synchronous to complete once the virtual memory of the host apparatus and the stand-by equipment；
The fault-tolerance approach synchronous based on virtual memory also includes：When the host apparatus breaks down, stop institute
The operation of the guest virtual machine of host apparatus is stated, and starts the operation of the guest virtual machine of the stand-by equipment, is allowed described
Stand-by equipment takes over the work of the host apparatus.
Once the virtual memory of the host apparatus and the stand-by equipment synchronously includes for the completion：Suspend described primary
The operation of the guest virtual machine of equipment, realize the memory pages content of the guest virtual machine of the stand-by equipment and the master
Memory pages content with the guest virtual machine of equipment is completely the same；Restart the guest virtual machine of the host apparatus
Operation, the guest virtual machine of the host apparatus completes the I/O operation included in this virtual memory synchronization, and protects
The data in magnetic disk for demonstrate,proving the data in magnetic disk and the stand-by equipment of the host apparatus is completely the same.
The content of pages of the guest virtual machine for realizing the stand-by equipment and the guest virtual machine of the host apparatus
Content of pages it is completely the same including：It is determined that visitor when occurring from the last trigger condition generation to presently described trigger condition
The virtual memory page that content of pages changes in the virtual machine of family, and by the content transmission of all identified virtual memory pages
To the stand-by equipment so that guest virtual machine described in the stand-by equipment it is corresponding to the identified virtual memory page
Virtual memory page content it is consistent with the content of identified virtual memory page in the host apparatus.
The synchronous trigger condition of the virtual memory includes：I/O occurs on the virtual client of the host apparatus
State change.
The present invention also provides a kind of MATLAB software system, including host apparatus and stand-by equipment, the host apparatus with
The stand-by equipment all includes synchronizing software module, fault management module, guest virtual machine module；The MATLAB software system
During operation, the synchronizing software module of the host apparatus and the stand-by equipment, fault management module, guest virtual machine module are all
Operate on virtualization kernel, correspond to the synchronizing software process virtualized on kernel, fault management process, Ke Huxu respectively
Plan machine；Wherein：The guest virtual machine is used for the operation for realizing application program；The guest virtual machine of host apparatus is in fortune
Row state, the guest virtual machine of the stand-by equipment are in synchronous operation but can not be accessed controlled state；The synchronization
Software process is used to preset the synchronous trigger condition of virtual memory, when the trigger condition occurs, realizes host apparatus and institute
The virtual memory stated between stand-by equipment is synchronous；The synchronizing software process of host apparatus is in running status, described standby
The synchronizing software process of equipment is in synchronous operation but can not be accessed controlled state；The fault management process is used for real
Now to the management and fault recovery of the host apparatus hardware, the guest virtual machine and the synchronizing software process；It is primary
The fault management process of equipment is in running status, and the fault management process of the stand-by equipment is in synchronous operation
But controlled state can not be accessed.
The synchronous implementation method of the virtual memory includes：Suspend the guest virtual machine operation of the host apparatus,
It is determined that content of pages occurs from guest virtual machine when the last time, the trigger condition occurred to occur to presently described trigger condition
The virtual memory page of change, and the virtual memory page to be changed to the content of pages synchronizes so that described standby to set
The content of the corresponding virtual memory page of guest virtual machine changes with content of pages described in the host apparatus described in standby
The content of the virtual memory page of change is consistent；Meanwhile page content is written to respective logic by host apparatus with stand-by equipment
In disk volume, I/O buffering areas are then discharged；Again the guest virtual machine operation of the host apparatus is recovered, it is described primary
The guest virtual machine of equipment completes the I/O read-write operations included in this virtual memory synchronization, and ensures the master
It is consistent with the data in magnetic disk of the stand-by equipment with the data in magnetic disk of equipment.
The synchronous trigger condition of the virtual memory includes：I/O shapes occur for the guest virtual machine of the host apparatus
State changes.
Storage I/O operation maximum read or write speed on the virtual client is per second for 50MB.
Network I/O operation maximum read or write speed on the virtual client is per second for 5MB.
When the fault management process detection of the host apparatus breaks down to the host apparatus, described in stopping
The operation of the guest virtual machine of host apparatus, and start the guest virtual machine operation of the stand-by equipment, start described standby
With the operation of the synchronizing software process and fault management process of equipment, the stand-by equipment is allowed to take over the work of the host apparatus
Make.
As described above, a kind of MATLAB software method and system of the present invention, have the advantages that：Event can be prevented
Barrier shuts down risk, and when host apparatus breaks down, business will migrate into stand-by equipment, because the internal storage data of stand-by equipment exists
Checkpoint keeps synchronous with the internal storage data of host apparatus, therefore operating system and software program can be continuously run not by shadow
Ring, business zero second is interrupted, higher without any loss of data, reliability using continuous uninterrupted；And the technical scheme independently of
Operating system layer, application are more extensive.
Brief description of the drawings
Fig. 1 is shown as a kind of schematic flow sheet of an embodiment of MATLAB software method of the present invention.
The virtual memory that Fig. 2 is shown as an a kind of embodiment of MATLAB software method of the present invention synchronously performs signal
Figure.
Fig. 3 is shown as a kind of module diagram of an embodiment of MATLAB software system of the present invention.
Fig. 4 is shown as a kind of structural representation of an embodiment of MATLAB software system of the present invention.
Component label instructions
1 MATLAB software system
11 host apparatus
111 guest virtual machine modules
112 synchronizing software modules
113 fault management modules
12 stand-by equipments
121 guest virtual machine modules
122 synchronizing software modules
123 fault management modules
S1~S3 steps
Embodiment
Illustrate embodiments of the present invention below by way of specific instantiation, those skilled in the art can be by this specification
Disclosed content understands other advantages and effect of the present invention easily.The present invention can also pass through specific realities different in addition
The mode of applying is embodied or practiced, the various details in this specification can also be based on different viewpoints with application, without departing from
Various modifications or alterations are carried out under the spirit of the present invention.
It should be noted that the diagram provided in the present embodiment only illustrates the basic conception of the present invention in a schematic way,
Then the component relevant with the present invention is only shown in schema rather than is painted according to component count, shape and the size during actual implement
System, kenel, quantity and the ratio of each component can be a kind of random change during its actual implementation, and its assembly layout kenel also may be used
Can be increasingly complex.
The present invention provides a kind of fault-tolerance approach synchronous based on virtual memory, applied to including host apparatus and stand-by equipment
MATLAB software system in, the host apparatus can be master server or main frame etc., and the stand-by equipment is and the master
With the identical equipment of the software and hardware setting of equipment, including from server, slave etc..The host apparatus and stand-by equipment are all
Including establishing the guest virtual machine on virtualization kernel.Between the host apparatus (main frame) and stand-by equipment (slave) generally
Based on the cascade of synchronous ethernet network, without transparent bridging NTB (None Transparent Bridge) cascades or infiniband
Cascade mode, the communication that subordinate computer node must also be used for test point (memory pages data syn-chronization) according to enough bandwidth need
Will.In one embodiment, as shown in figure 1, the fault-tolerance approach synchronous based on virtual memory includes：
Step S1, preset the synchronous trigger condition of virtual memory.The synchronous trigger condition of the virtual memory can include：
I/O state changes occur on virtual client on the host device.I/O state changes not only include storage I/O states or
Network I/O state changes, and this change is not limited to data variation, including various I/O request changes, all kinds of states become
Change.In one embodiment, the synchronous trigger condition of virtual memory is the I/O states of virtual client system on host apparatus
Change, including the change of disk, network data, resource occupation state, time state and linking status etc..Specifically, in system
The central processing unit of host apparatus starts synchronizing software process, and the moment monitors the virtual client I/O state changes, each time
Change can all generate the synchronous trigger condition of virtual memory and determine checkpoint (checkpoint), and perform the virtual memory page
It is synchronous.
Step S2, when the trigger condition occurs, complete once the virtual of the host apparatus and the stand-by equipment
Internal memory is synchronous.When the trigger condition occurs, it is necessary to complete once the host apparatus and the stand-by equipment it is virtual in
Deposit synchronization.Specifically, when the trigger condition occurs, the guest virtual machine for suspending the host apparatus is run, i.e.,
The state that (Held) main frame is " preparation " is kept, this " preparation " state is referred to as checkpoint (Checkpoint) or test point.
The time interval occurred to next trigger condition occurs for trigger condition, i.e., a checkpoint to another checkpoint when
Between, it can be described as the checkpoint cycle.In one embodiment, it is determined that occurring from the last trigger condition to presently described triggering
The virtual memory page that changes of content of pages in guest virtual machine when condition occurs, and by all identified virtual memorys
The content transmission of page is to the stand-by equipment so that the content of the virtual memory page of guest virtual machine described in the stand-by equipment
It is consistent with the content of identified virtual memory page in the host apparatus.Specifically, host apparatus will be recorded in an inspection
The virtual memory page modification situation occurred in the point cycle, when carrying out synchronous with the virtual memory of stand-by equipment, only to described
Send modification the virtual memory page synchronize, with realize the content of pages of the guest virtual machine of the stand-by equipment with it is described
The content of pages of the guest virtual machine of host apparatus is completely the same.This method is the side generally used in virtual memory synchronization
Method.In another embodiment, when carrying out synchronous with the virtual memory of stand-by equipment, by client virtual described in host apparatus
The content transmission of all virtual memory pages of machine is into the stand-by equipment so that client virtual described in the stand-by equipment
All virtual memory pages are consistent with the corresponding virtual memory page of guest virtual machine described in the host apparatus in machine.I.e. pair
All virtual memory pages in the host apparatus synchronize, to realize the page of the guest virtual machine of the stand-by equipment
Content and the content of pages of the guest virtual machine of the host apparatus are completely the same.Meanwhile host apparatus and stand-by equipment are by
Page content is deposited to be written in the form of Block (data block) in respective logic magnetic disc volume, it is real based on disk sector bitmap indexes
Existing increment synchronization, now, the central processing unit of the stand-by equipment feed back centre of the ack signal to the host apparatus
Device (CPU) is managed, disk synchronously completes and confirms that data are consistent.Then central processing unit release I/O buffering areas.Again described in recovering
Guest virtual machine described in host apparatus is run, and the guest virtual machine completes the I/O included in this virtual memory synchronization
Read-write operation, and ensure that the data in magnetic disk of the host apparatus is consistent with the data in magnetic disk of the stand-by equipment.And described in ensureing
The data in magnetic disk of host apparatus is consistent with the data in magnetic disk of the stand-by equipment.
A checkpoint is generated around an I/O state change, and carries out virtual memory page synchronization, the void on host apparatus
Intend management of the client machine system to disk and network and rely on " I/O buffering areas ", in each checkpoint cycle with the end of the step after, I/
O newly asks just to be released, and is ready to carry out next checkpoint cycle.The number that checkpoint occurs in unit interval is referred to as
Checkpoint rate, unit can include " cycle per second ".The characteristic of application load determines the cycle of checkpoint in computer system
Frequency.Collect that how many memory pages is changed, depends primarily on host apparatus virtual client system in the cycle of checkpoint
The frequency of I/O state changes.The cycle per second for reducing checkpoint, client operating system have bigger potentiality to be transported in large quantities to do
Calculate.Higher checkpoint rate, represents measurable resource occupation.Typically smaller than 200 cycles are per second to mean that system is not busy.
The cycle of checkpoint can take physical machine memory source and synchronizing network bandwidth.Higher checkpoint rate, it will cause to buffer in I/O
The deenergized period in area and I/O network delays all can be smaller.With the increase of I/O state changes (such as network activity), frequent
The checkpoint cycle reduces with the produced delay arrived.Different types of service generally has different internal memory synchronizing frequencies, such as：Meter
It is 1-10 times per second to calculate comprehensive sounding；Probe the memory business is 10-50 times per second；File duplication service is 5-20 times per second；SQL
Inquiry business is 10-30 times per second；Web document transfer business is 50-200 times per second；SQL transaction moderate business 50- per second
500 times；SQL merchandises frequent business as 500-1500 times per second.To ensure the actual effect of Client application, virtual i/o request response
Flow control measure are used including disk and network.In one embodiment, it is per second for network (each Microsoft Loopback Adapter)
Maximum 5MB files transmission bandwidth, for disk, maximum 50MB file read-writes per second.
In one embodiment, as shown in Fig. 2 the I/O states of host apparatus include storage and network I/O states are all closed
It is coupled in internal memory synchronized process, network state is kept (Hold) in active host node I/O buffering areas, storage read-write shape
State performs on the primary node, but stores write state and (Hold) is being kept from node in I/O buffering areas.Each I/O shapes
State change can be all monitored, and determined its type and judged whether to need to trigger checkpoint, if need memory pages synchronous.If
Need, operating guest virtual machine system is stopped just in host node, and is examined after last checkpoint to this
The modification of the memory pages in cycle content is made an inventory of to be collected and be sent in the internal memory synchronized process on secondary node.Once from
Host node captures memory pages modification content, and guest virtual machine system VM will again continue to run ... transports from host node
Capable internal memory synchronized process, the content of pages changed can be mapped in local memory, and trigger one group and perform request, respectively
Discharge network transmission requests in host node I/O buffering areas and from the disk write request in host node I/O buffering areas.Especially
" held from host node, therefore from the representative of host disk content it should be noted that disk write request is only kept (Hold)
Before row checkpoint " mirror image data, if host node is delayed machine before checkpoint is completed, subordinate computer node preservation is upper one
The data that individual checkpoint is completed, and regenerate I/O transmission requests.In recovery process, it can be ensured by disk mirroring mode
The uniformity of two side datas.This way, data in magnetic disk is repeatedly written between can avoiding failover period.
In one embodiment, the fault-tolerance approach synchronous based on virtual memory also includes：
Step S3, when the host apparatus breaks down, stop the fortune of the guest virtual machine of the host apparatus
OK, and start the operation of the guest virtual machine of the stand-by equipment, allow the stand-by equipment to take over the work of the host apparatus.
Specifically, when the fault management module detects that the host apparatus breaks down, the described of the host apparatus is stopped
The operation of guest virtual machine, and start operation and the I/O correspondence with foreign countries of the guest virtual machine of the stand-by equipment, receive client
Access Management Access is run；Complete the once transfer to the virtual client access control.In one embodiment, virtual machine client
I/O buffering areas will not be used during machine internal memory migration, after once successful internal memory migration terminates, having an of short duration network please
Pause is asked, guest virtual machine runs on former secondary node and confirms the checkpoint cycle, but this network request stalled cycles
Less than 1 millisecond, for business network transmission for it is negligible, Ethernet link state and TCP data transmission not by
Any influence.When therefore having cashed any hostdown generation, service switching is interrupted for zero second.Now, because host apparatus is (main
Machine) nodes break down, virtual memory synchronized process is out of service, and guest virtual machine no longer fault-tolerant operation, it is referred to as " drop
Level " state, working condition is simplex mode.Magnetic disc i/o (write-in), which replicates, to be stopped.
In the virtual memory synchronizing process of guest virtual machine, the active state of guest virtual machine is including following several：1,
Start/starting state：A period of time after startup guest virtual machine system, once network service is clear and coherent, state may switch to " fortune
OK ".Process before this state, referred to as " start ".2, stop/halted state：Turn-off request has been sent to client's void
Plan machine operating system.Until virtual machine disconnects internal bus, it is represented in " a stopping " state.Afterwards, define
For halted state.3, it is currently running state：, should when the network service that can confirm that guest virtual machine and the establishment of disk read-write state
State is represented as " being currently running ".4, transition state：When a guest virtual machine access rights in host apparatus node to standby
With the state for migrating (running from node until it is successfully transferred to) between device node, it is defined as " migrating ".
The present invention also provides a kind of MATLAB software system.In one embodiment, as shown in figure 3, MATLAB software system
System 1 includes host apparatus 11 and stand-by equipment 12.The host apparatus 11 can be master server or main frame etc., described standby to set
Standby 12 be the identical equipment of software and hardware setting with the host apparatus 11, including from server, slave etc..It is described primary
Equipment 11 and stand-by equipment 12 all include the guest virtual machine established on virtualization kernel.Host apparatus (main frame, the main clothes
Business device, host node etc.) it is typically based on synchronous ether between 11 and stand-by equipment (slave, from server, subordinate computer node etc.) 12
Cascade, without transparent bridging NTB (None Transparent Bridge) cascade or infiniband cascade sides
Formula, the communication that subordinate computer node must also be used for test point (memory pages data syn-chronization) according to enough bandwidth need.
The host apparatus 11 includes guest virtual machine module 111, synchronizing software module 112 and fault management module
113；The stand-by equipment 12 includes guest virtual machine module 121, synchronizing software module 122 and fault management module 123.Institute
When stating MATLAB software system 1 and running, the synchronizing software module of the host apparatus 11 and the stand-by equipment 12 (112 and
122), fault management module (113 and 123), guest virtual machine module (111 and 112) are all operated on virtualization kernel, respectively
Synchronizing software process, fault management process on the corresponding virtualization kernel, guest virtual machine.I.e. described host apparatus 11
Synchronizing software module (112), fault management module (113), guest virtual machine module (111) all operate in the host apparatus 11
Virtualization kernel on, correspond to synchronizing software process (112) on the virtualization kernel of host apparatus 11, failure pipe respectively
Reason process (113), guest virtual machine (111).Synchronizing software module (122), the fault management module of the stand-by equipment 12
(123), guest virtual machine module (112) is all operated on the virtualization kernel of the stand-by equipment 12, corresponds to standby set respectively
Synchronizing software process (122), fault management process (123) on the standby 12 virtualization kernel, guest virtual machine (121).
The guest virtual machine (111 and 121) is used for the operation for realizing application program.Wherein, host apparatus 11 is described
Guest virtual machine 111 is in running status, and the guest virtual machine 121 of the stand-by equipment 12 is in halted state.
The synchronizing software module or synchronizing software process (112 and 122) are used to preset the synchronous triggering bar of virtual memory
Part, when the trigger condition occurs, realize that the virtual memory between host apparatus 11 and the stand-by equipment 12 is synchronous.It is described
The synchronous trigger condition of virtual memory can include：The I/O state changes of virtual client system on host apparatus.Specifically,
I/O state changes not only include storage I/O states or network I/O state changes, and this change is not limited to data variation,
Ask to change including various I/O, all kinds of state changes.Each I/O state changes of host apparatus 11 can be all monitored in system,
Determine its type and judge whether to need to trigger checkpoint, if need memory pages synchronous.In one embodiment, Zhu Congshe
The synchronous trigger condition of virtual memory is the I/O state changes of virtual client system on host apparatus between standby, including magnetic
Disk, network data change, resource occupation state, time state and linking status etc..
The implementation method that virtual memory is synchronous between the host apparatus 11 and the stand-by equipment 12 includes：Described in stopping
The guest virtual machine 111 of host apparatus 11, it is determined that occurring from the last trigger condition to presently described trigger condition
The virtual memory page that content of pages changes in guest virtual machine 111 during generation, and the content of pages is changed
Virtual memory page synchronizes so that the corresponding virtual memory page of guest virtual machine 121 described in the stand-by equipment 12
Content is consistent with the content of the virtual memory page that content of pages changes described in the host apparatus 11；Described in restarting
Guest virtual machine 111 is run, and the guest virtual machine 111 completes the I/O operation included in this virtual memory synchronization, and
Ensure that the data in magnetic disk of the host apparatus 11 is consistent with the data in magnetic disk of the stand-by equipment 12 in a manner of disk mirroring.
When the trigger condition occurs, the guest virtual machine 111 of the host apparatus 11 will be stopped, that is, kept
(Held) main frame is the state of " preparation ", and this " preparation " state is referred to as checkpoint (Checkpoint) or test point.Once
The time of the time interval, i.e. a checkpoint to another checkpoint occurred to next trigger condition occurs for trigger condition,
It can be described as the checkpoint cycle.In one embodiment, it is determined that occurring from the last trigger condition to presently described triggering bar
The virtual memory page that changes of content of pages in guest virtual machine 111 when part occurs, and will be all determined by it is virtual in
The content transmission of page is deposited to the stand-by equipment 12 so that guest virtual machine 121 described in the stand-by equipment 12 with it is described
The content of the identified corresponding virtual memory page of virtual memory page and identified virtual memory page in the host apparatus 11
Content it is consistent.Specifically, host apparatus 11 will be recorded in the virtual memory page modification feelings occurred in a checkpoint cycle
Condition, when carrying out synchronous with the virtual memory of stand-by equipment 12, only the virtual memory page for sending modification is synchronized,
To realize the guest virtual machine 111 of the content of pages of the guest virtual machine 121 of the stand-by equipment 12 and the host apparatus 11
Content of pages it is completely the same.This method is the method generally used in virtual memory synchronization.In another embodiment, exist
When carrying out synchronous with the virtual memory of stand-by equipment 12, by all virtual of guest virtual machine 111 described in host apparatus 11
The content transmission of page is into the stand-by equipment 12 so that institute in guest virtual machine 121 described in the stand-by equipment 12
Some virtual memory pages are consistent with the corresponding virtual memory page of guest virtual machine 111 described in the host apparatus 11.I.e. pair
All virtual memory pages in the host apparatus 11 synchronize, to realize the guest virtual machine of the stand-by equipment 12
The content of pages of 121 content of pages and the guest virtual machine 111 of the host apparatus 11 is completely the same.Then, will open again
The guest virtual machine 111 for moving the host apparatus 11 is run, and the guest virtual machine 111 of the host apparatus 11 is completed
The I/O operation included in this virtual memory synchronization, and ensure the data in magnetic disk of the host apparatus 11 with it is described standby
The data in magnetic disk of equipment 12 is consistent.
Business network transmission have to be around a checkpoint and synchronize, and network transmission relies on " I/O buffering areas ", each
In the cycle of checkpoint with the end of the step after, I/O newly asks just to be released, and is ready to carry out next checkpoint cycle.During unit
The number that interior checkpoint occurs is referred to as checkpoint rate, and unit can include " cycle per second ".Using negative in computer system
The characteristic of load determines the period frequency of checkpoint.Collect that how many memory pages is changed, is mainly taken in the cycle of checkpoint
Certainly in network I/O (transmission) speed.The cycle per second for reducing checkpoint, client operating system have bigger potentiality to do in large quantities
Computing.Higher checkpoint rate, represents measurable resource occupation.Typically smaller than 200 cycles are per second to mean that system is not numerous
It is busy.The cycle of checkpoint can take physical machine memory source and synchronizing network bandwidth.Higher checkpoint rate, will cause in I/O
The deenergized period of buffering area and I/O network delays all can be smaller.With the increase of network activity, the frequent checkpoint cycle with
The produced delay arrived reduces.To ensure the actual effect of Client application, virtual i/o request response includes disk and network uses
Flow control measure.In one embodiment, it is maximum 5MB files transmission bandwidth per second for network (each Microsoft Loopback Adapter),
For disk, maximum 50MB file read-writes per second.
In one embodiment, host apparatus 11 (including main frame or master server etc.) is host node.Host node
Storage and network I/O operation are all associated with internal memory synchronized process, and network state is kept (Hold) in active main frame section
In point I/O buffering areas, storage read-write state performs on the primary node, but stores write state and be kept from node (Hold)
In I/O buffering areas.Each network transmission requests can be all monitored, and determined its type and judged whether to need to trigger checkpoint,
Whether need memory pages synchronous.If desired, operating guest virtual machine system 111 is stopped just in host node,
And it is collected and is sent to standby to the memory pages modification content in this checkpoint cycle after last checkpoint
In internal memory synchronized process on node.Once capture memory pages modification content, guest virtual machine system VM from host node
The internal memory synchronized process run from host node will be again continued to run ..., the content of pages changed can be mapped to local
In internal memory, and trigger one group and perform request, respectively network transmission requests in release host node I/O buffering areas and from main frame section
Disk write request in point I/O buffering areas.It is particularly to be noted that disk write request is only kept (Hold) from main frame section
On point, therefore from the mirror image data of host disk content representative " before performing checkpoint ", if main frame section before the completion of checkpoint
Point is delayed machine, and what subordinate computer node preserved is the data that a upper checkpoint is completed, and regenerates I/O transmission requests.Extensive
During multiple, the uniformity of two side datas can be ensured by disk mirroring mode.
The fault management module or fault management process (113 and 123) are used to realize to the host apparatus hardware, institute
State the management and fault recovery of guest virtual machine (111 and 121) and the synchronizing software module (112 and 122).Specifically,
When the fault management module detects that the host apparatus breaks down, stop the client virtual of the host apparatus
The operation of machine, and start the operation of the guest virtual machine of the stand-by equipment.Further, when the fault management module detects
When being broken down to the host apparatus, stop the operation of the guest virtual machine of the host apparatus, and start described standby
Operation and I/O correspondence with foreign countries with the guest virtual machine of equipment, receive client Access Management Access operation；Complete once to the void
Intend the transfer of client computer access control.
In one embodiment, as shown in figure 4, the fault detect (i.e. fault management module 113 and 123) in figure passes through private
There is network to be connected, network control module is responsible for internal memory synchronized process (realization of internal memory synchronization module 112 and 122) and performs host apparatus
The data transfer of virtual memory synchronization between 11 and stand-by equipment 12.It is virtual by realizing when host apparatus 11 breaks down
The migration of client computer, i.e., the virtual client 121 for moving to stand-by equipment 12 is run from the virtual client 111 of host apparatus 11
Upper operation.(moving to virtual client 121 from virtual client 111 to run) during virtual client internal memory migration to make
With I/O buffering areas.Now, stand-by equipment 12 will become host apparatus 11 and run virtual client (121 become 111), originally
Host apparatus will be stopped using.After once successful internal memory migration terminates, have an of short duration network request and pause, client virtual
Machine 111 runs on former secondary node and confirms the checkpoint cycle, but this network request stalled cycles is less than 1 millisecond, right
Negligible for business network transmission, Ethernet link state and TCP data transmission are not affected.Therefore
When having cashed any hostdown generation, service switching is interrupted for zero second.Now, the machine if host apparatus (main frame) node is delayed
Failure, virtual memory synchronized process is out of service, and the no longer fault-tolerant operation of guest virtual machine 111, it is referred to as the state that " degrades ",
Working condition is simplex mode.It can only operate on single host apparatus, magnetic disc i/o duplicating process is out of service.It is prior
It is, if the original node failure of host apparatus (main frame) 11, but non-stopped status, such as network interruption, fan failure etc., virtual visitor
Although family machine migrates, i.e., the virtual client for moving to stand-by equipment 12 is run from the virtual client 111 of host apparatus 11
Run on machine 121.The normal operation but virtual memory synchronized process remains unchanged, disk are synchronously also normally carried out.It is simply former standby to set
Standby 12 will become host apparatus 11, and virtual client performs migration (121 become 111).Now virtual client 111 is in portion
The fault-tolerant state of subdivision, it also referred to as " degrades " state, and working condition is non-fully fault-tolerant mode.Exactly because this is special
Sign advantage, it is possible to achieve the failure cross occurrence of the host apparatus and the stand-by equipment, or to intersect cooperating, than
Such as, the host apparatus network node sends failure, the stand-by equipment disk nodes break down, the virtual client according to
Old normal operation.
In one embodiment, MATLAB software system configures No. 1 network interfaces (NIC 1) of identical server to even by two
(one is main frame, and one is slave), using super Category-5 twisted pair.The double 10,000,000,000 network interface distiches that will be integrated, using multimode fibre
Wire jumper LC joints.Using KVM virtualization kernel, host system Cent OS Linux6.5 or more version of increasing income is installed,
Installation KVM virtual machines component simultaneously enables.Application APP or database are run on virtual client (121 become 111),
Synchronizing software module (112 and 122) includes providing that internal memory is synchronous and selection mediation service, system hair when avoiding the synchronization links from interrupting
Raw " fissure " phenomenon.Fault management module (113 and 123) creates a Domain0 fictitious host computer, can select Cent OS
Or the linux system of other main brands；And employ Apache tomcat Server and carry out self-defined developing user interface
UI(User interface).Fault management module (113 and 123) imports server master board IPMI packets, can realize pair
The browser mode of whole tolerant system is safeguarded, safeguards that object includes host hardware, guest virtual machine and synchronous operation state, tool
Standby resource distribution and trouble handling function.
In summary, a kind of MATLAB software method and system of the present invention, have the advantages that：Being capable of trouble saving
Risk is shut down, when host apparatus breaks down, business will migrate into stand-by equipment, and the internal storage data due to stand-by equipment is being examined
Make an inventory of keep it is synchronous with the internal storage data of host apparatus, therefore operating system and software program can continuously run it is unaffected,
Business zero second migrates, higher without any loss of data, reliability using continuous uninterrupted；And the technical scheme is independently of behaviour
Make system layer, application is more extensive.So the present invention effectively overcomes various shortcoming of the prior art and has high industrial profit
With value.
The above-described embodiments merely illustrate the principles and effects of the present invention, not for the limitation present invention.It is any ripe
Know the personage of this technology all can carry out modifications and changes under the spirit and scope without prejudice to the present invention to above-described embodiment.Cause
This, those of ordinary skill in the art is complete without departing from disclosed spirit and institute under technological thought such as
Into all equivalent modifications or change, should by the present invention claim be covered.
Claims (8)
1. a kind of MATLAB software method, applied in the MATLAB software system including host apparatus and stand-by equipment, it is special
Sign is that the host apparatus and stand-by equipment all include the guest virtual machine established on virtualization kernel, the computer
Fault-tolerance approach includes：
The synchronous trigger condition of default virtual memory；
When the trigger condition occurs, suspend the operation of the guest virtual machine of the host apparatus, it is determined that from the last time
The trigger condition occur to presently described trigger condition occurs when guest virtual machine in content of pages change it is virtual interior
Deposit page, and by the content transmission of all identified virtual memory pages to the stand-by equipment so that in the stand-by equipment
The content of the virtual memory page corresponding with the identified virtual memory page of the guest virtual machine and the host apparatus
In identified virtual memory page content it is consistent；
The guest virtual machine operation of the host apparatus is restarted, the guest virtual machine of the host apparatus is completed
The I/O operation included in this virtual memory synchronization, and ensure that the data in magnetic disk of the host apparatus standby is set with described
Standby data in magnetic disk is completely the same.
2. MATLAB software method according to claim 1, it is characterised in that：The MATLAB software method also includes：
When the host apparatus breaks down, stop the operation of the guest virtual machine of the host apparatus, and start described standby
With the operation of the guest virtual machine of equipment, the stand-by equipment is allowed to take over the work of the host apparatus.
3. MATLAB software method according to claim 1, it is characterised in that：The synchronous trigger condition of the virtual memory
Including：I/O state changes occur on the guest virtual machine of the host apparatus.
4. a kind of MATLAB software system, including host apparatus and stand-by equipment, it is characterised in that：The host apparatus with it is described
Stand-by equipment all includes synchronizing software module, fault management module, guest virtual machine module；The MATLAB software system operation
When, synchronizing software module, fault management module, the guest virtual machine module of the host apparatus and the stand-by equipment are all run
On virtualization kernel, the synchronizing software process virtualized on kernel, fault management process, client virtual are corresponded to respectively
Machine；Wherein：
The guest virtual machine is used for the operation for realizing application program；The guest virtual machine of host apparatus is in operation shape
State, the guest virtual machine of the stand-by equipment are in synchronous operation but can not be accessed controlled state；
The synchronizing software process is used to preset the synchronous trigger condition of virtual memory, when the trigger condition occurs, pause
The guest virtual machine operation of the host apparatus, it is determined that occurring from the last trigger condition to presently described triggering bar
The virtual memory page that changes of content of pages in guest virtual machine when part occurs, and to void that the content of pages changes
Intend page to synchronize so that the content of the corresponding virtual memory page of guest virtual machine described in the stand-by equipment and institute
The content for stating the virtual memory page that content of pages described in host apparatus changes is consistent；Meanwhile host apparatus is set with standby
It is standby that page content is written in respective logic magnetic disc volume, then discharge I/O buffering areas；Restart the host apparatus
Guest virtual machine operation, the guest virtual machine of the host apparatus completes being wrapped in this virtual memory synchronization
The I/O read-write operations contained, and ensure that the data in magnetic disk of the host apparatus is consistent with the data in magnetic disk of the stand-by equipment；It is primary
The synchronizing software process of equipment is in running status, and the synchronizing software process of the stand-by equipment is in synchronous operation
But controlled state can not be accessed；
The fault management process is used to realize to the hardware of the primary and backup equipment, the guest virtual machine and described
The management and fault recovery of synchronizing software process；The fault management process of host apparatus is in running status, described standby
The fault management process of equipment is in synchronous operation but can not be accessed controlled state.
5. MATLAB software system according to claim 4, it is characterised in that：The synchronous trigger condition of the virtual memory
Including：I/O state changes occur for the guest virtual machine of the host apparatus.
6. MATLAB software system according to claim 4, it is characterised in that：Storage I/O on the guest virtual machine
It is per second for 50MB to operate maximum read or write speed.
7. MATLAB software system according to claim 4, it is characterised in that：Network I/O on the guest virtual machine
It is per second for 5MB to operate maximum read or write speed.
8. MATLAB software system according to claim 4, it is characterised in that：When the failure pipe of the host apparatus
Reason process detection to the host apparatus break down when, stop the operation of the guest virtual machine of the host apparatus, and
Start the operation of the guest virtual machine of the stand-by equipment, start the synchronizing software process and fault management of the stand-by equipment
The operation of process, the stand-by equipment is allowed to take over the work of the host apparatus.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410632804.3A CN104391764B (en) | 2014-10-22 | 2014-11-11 | A kind of MATLAB software method and system |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2014105679235 | 2014-10-22 | ||
| CN201410567923 | 2014-10-22 | ||
| CN201410632804.3A CN104391764B (en) | 2014-10-22 | 2014-11-11 | A kind of MATLAB software method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104391764A CN104391764A (en) | 2015-03-04 |
| CN104391764B true CN104391764B (en) | 2018-02-16 |
Family
ID=52609672
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410632804.3A Active CN104391764B (en) | 2014-10-22 | 2014-11-11 | A kind of MATLAB software method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104391764B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105373418A (en) * | 2015-11-25 | 2016-03-02 | 北京汉柏科技有限公司 | Implementation method and device of virtual machine FT (Fault Tolerance) |
| CN105471994B (en) * | 2015-12-01 | 2019-01-15 | 华为技术有限公司 | A kind of control method and device |
| CN108885575B (en) * | 2016-04-01 | 2022-03-11 | 三菱电机株式会社 | Control device and restoration processing method for control device |
| CN106254236A (en) * | 2016-08-05 | 2016-12-21 | 成都广达新网科技股份有限公司 | A kind of multiserver slave method of work based on TCP event |
| CN106970861A (en) * | 2017-03-30 | 2017-07-21 | 山东超越数控电子有限公司 | A kind of virtual machine fault-tolerance approach and system |
| CN107315624B (en) * | 2017-06-30 | 2020-11-20 | 联想(北京)有限公司 | Information processing method and virtualization manager |
| CN109150596B (en) * | 2018-08-08 | 2021-08-03 | 新智能源系统控制有限责任公司 | SCADA system real-time data dump method and device |
| CN112131088B (en) * | 2020-09-29 | 2024-04-09 | 北京计算机技术及应用研究所 | High availability method based on health examination and container |
| CN112256477A (en) * | 2020-10-09 | 2021-01-22 | 上海云轴信息科技有限公司 | Virtualization fault-tolerant method and device |
| CN113741248B (en) * | 2021-08-13 | 2023-07-14 | 北京和利时系统工程有限公司 | Edge calculation controller and control system |
| CN114217905A (en) * | 2021-12-17 | 2022-03-22 | 北京志凌海纳科技有限公司 | High-availability recovery processing method and system for virtual machine |
| CN114501057A (en) * | 2021-12-17 | 2022-05-13 | 阿里巴巴（中国）有限公司 | Data processing method, storage medium, processor and system |
| CN115858222B (en) * | 2022-12-19 | 2024-01-02 | 安超云软件有限公司 | Virtual machine fault processing method, system and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101034364A (en) * | 2007-04-02 | 2007-09-12 | 华为技术有限公司 | Method, device and system for implementing RAM date backup |
| JP2011180871A (en) * | 2010-03-02 | 2011-09-15 | Nec Corp | Fault tolerant system and virtual machine construction method |
| CN102262558A (en) * | 2011-08-04 | 2011-11-30 | 中兴通讯股份有限公司 | Synchronizing method and system of virtual machine |
| CN103412800A (en) * | 2013-08-05 | 2013-11-27 | 华为技术有限公司 | Virtual machine warm backup method and equipment |
| US8826283B2 (en) * | 2008-10-28 | 2014-09-02 | Vmware, Inc. | Low overhead fault tolerance through hybrid checkpointing and replay |
-
2014
- 2014-11-11 CN CN201410632804.3A patent/CN104391764B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101034364A (en) * | 2007-04-02 | 2007-09-12 | 华为技术有限公司 | Method, device and system for implementing RAM date backup |
| US8826283B2 (en) * | 2008-10-28 | 2014-09-02 | Vmware, Inc. | Low overhead fault tolerance through hybrid checkpointing and replay |
| JP2011180871A (en) * | 2010-03-02 | 2011-09-15 | Nec Corp | Fault tolerant system and virtual machine construction method |
| CN102262558A (en) * | 2011-08-04 | 2011-11-30 | 中兴通讯股份有限公司 | Synchronizing method and system of virtual machine |
| CN103412800A (en) * | 2013-08-05 | 2013-11-27 | 华为技术有限公司 | Virtual machine warm backup method and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104391764A (en) | 2015-03-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104391764B (en) | A kind of MATLAB software method and system | |
| US8875134B1 (en) | Active/active storage and virtual machine mobility over asynchronous distances | |
| CN108984107B (en) | Increasing availability of storage systems | |
| US9423956B2 (en) | Emulating a stretched storage device using a shared storage device | |
| US8689054B1 (en) | Increased distance of virtual machine mobility over asynchronous distances | |
| CN106357787A (en) | Storage disaster tolerant control system | |
| US20150149813A1 (en) | Failure recovery system and method of creating the failure recovery system | |
| US20120303594A1 (en) | Multiple Node/Virtual Input/Output (I/O) Server (VIOS) Failure Recovery in Clustered Partition Mobility | |
| US20090157766A1 (en) | Method, System, and Computer Program Product for Ensuring Data Consistency of Asynchronously Replicated Data Following a Master Transaction Server Failover Event | |
| US9483352B2 (en) | Process control systems and methods | |
| US8726083B1 (en) | Synchronized taking of snapshot memory images of virtual machines and storage snapshots | |
| CN103793258A (en) | Hypervisor-based server duplication system and method | |
| US9442811B2 (en) | Emulating a stretched storage device using a shared replicated storage device | |
| CN205792734U (en) | The disaster recovery drilling system that a kind of facing cloud calculates | |
| US8682852B1 (en) | Asymmetric asynchronous mirroring for high availability | |
| Lu et al. | Fast memory state synchronization for virtualization-based fault tolerance | |
| CN108469996A (en) | A kind of system high availability method based on auto snapshot | |
| CN103885857A (en) | Virtual machine fault-tolerant method and device | |
| US20210294497A1 (en) | Storage system and method for analyzing storage system | |
| CN107111530A (en) | A kind of disaster recovery method, system and device | |
| CN111597079B (en) | Method and system for detecting and recovering MySQL Galera cluster faults | |
| CN111400086B (en) | Method and system for realizing fault tolerance of virtual machine | |
| CN105824571A (en) | Data seamless migration method and device | |
| CN110377487A (en) | A kind of method and device handling high-availability cluster fissure | |
| US20200387575A1 (en) | Migrating virtual machines using asynchronous transfer and synchronous acceleration |
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 |