CN107515772A - A kind of detection KVM virtual machines hang dead method and device - Google Patents
A kind of detection KVM virtual machines hang dead method and device Download PDFInfo
- Publication number
- CN107515772A CN107515772A CN201610423949.1A CN201610423949A CN107515772A CN 107515772 A CN107515772 A CN 107515772A CN 201610423949 A CN201610423949 A CN 201610423949A CN 107515772 A CN107515772 A CN 107515772A
- Authority
- CN
- China
- Prior art keywords
- virtual machine
- thread
- pointers
- virtual
- machine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/301—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is a virtual computing platform, e.g. logically partitioned systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45591—Monitoring or debugging support
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Debugging And Monitoring (AREA)
Abstract
Dead method and device is hung the invention discloses a kind of detection KVM virtual machines, is related to the technical field that virtual machine state is detected in cloud computing system, methods described includes：By inquiring about virtual machine, the PC pointers of its respective thread of virtual machine are obtained；Analyzed by the PC pointers of its respective thread to resulting virtual machine, determine the virtual machine whether in extension death situation state；If it is determined that the virtual machine, which is in, hangs death situation state, then virtual machine is reported to hang dead alarm information.By inquiring about PC pointers and analyzing, it can accurately determine that virtual machine hangs death situation state.
Description
Technical field
The present invention relates to cloud computing system detection in information technology (InformationTechnology, IT) field is virtual
The application of machine state, more particularly to a kind of detection KVM virtual machines hang dead method and device.
Background technology
Cloud computing is current very popular cutting edge technology, and a very important feature of cloud computing is virtualization.Make
After virtualization technology, user need not reuse physical computer, and all operations can be realized using virtual machine.Make
It is it will be apparent that on the one hand can easily carry out central processing unit (Central Processing with the advantages of virtual machine
Unit, CPU), the extension of internal memory, hard disk, the resource such as peripheral hardware, on the other hand can be with rapid deployment.
In general, virtual environment is made up of three parts：Hardware, monitor of virtual machine (Virtual Machine
Monitor, VMM) and virtual machine.Fig. 1 (a) is the physical machine model schematic that prior art provides, and as shown in Fig. 1 (a), is not being had
In the case of having virtualization, operating system is run directly on hardware, manages bottom physical hardware, forms one completely
Computer system, that is, " physical machine ".Fig. 1 (b) is the Virtual Machine Model schematic diagram that prior art provides, such as Fig. 1 (b) institutes
Show, in virtual machine environment, VMM seizes the position of operating system, becomes the manager of actual physical hardware, while to upper strata
Software shows virtual hardware platform, the operating system on " deception " upper strata.And now the operating system on upper strata is virtual
On platform, still manage it considers that be the virtual hardware of " physical hardware ", do not know just like below there occurs what, this is just
It is " virtual machine ".Therefore VMM major function is to create corresponding virtual resource based on physical resource, forms virtual machine,
Virtual platform is provided for VME operating system, so it needs to manage virtual environment and physical resource.
Realized according to the virtualization technology of current main-stream, VMM can be divided into three classes, and be that Fig. 2 (a) is provided respectively
Hypervisor models, the host model that the mixed model and Fig. 2 (c) that Fig. 2 (b) is provided provide.Shown in such as figure Fig. 2 (a),
In Hypervisor models, VMM can regard a complete operating system as, but unlike legacy operating system, VMM
Being designed for virtualization, all physical resources such as processor, internal memory and I/O equipment etc. all return VMM to own, therefore, VMM
Carry the responsibility of management physical resource；Secondly, VMM needs to provide virtual machine upwards for running VME operating system, because
This, VMM is also responsible for the establishment and management of virtual environment, and VMWare is exactly typical Hypervisor types.As shown in Fig. 2 (b), mix
Matched moulds type is the mutation of Hypervisor models, and VMM still grasps all physical resources, but can concede most of I/O and set
Standby control, a privileged operating system operated in privileged virtual machine is transferred to control them, in Fig. 2 (b)
VM0.Accordingly, the responsibility of VMM virtualizations is also shared, and the virtualization of processor and internal memory is still completed by VMM, and I/O
Virtual machineization then cooperated by VMM and privileged operating system to complete, typical mixed model is XEN.Such as Fig. 2 (c) institutes
Show, in host model, physical resource is by host (Host) operating system management.Host operating system is traditional operation system
System, such as Windows, Linux etc., these legacy operating systems are not to be designed for virtualization, therefore are not had in itself
Virtualization, actual virtualization are provided by VMM.VMM is typically the kernel module of host operating system independence.
VMM obtains resource by calling the service of host operating system, realizes the virtualization of processor, internal memory and I/O equipment.
VMM is created that after virtual machine, generally participates in scheduling using virtual machine as a process of host operating system.Based on kernel
Virtual machine (Kernel-based Virtual Machine, KVM) be exactly a kind of VMM of typical host model, KVM is created
Virtual machine participate in normal dispatch as a process on Host, and the virtual center processor in each virtual machine
(Virtial CPU, VCPU) exists as a thread of the process.
VMM needs to monitor the state of virtual machine as virtual management platform, collects information during virtual machine operation.Host
The VMM of model is at present typically by the way of two kinds conventional.Fig. 3 (a) is that the main frame that prior art provides and virtual machine use void
The standby interaction schematic diagram being monitored is proposed, as shown in Fig. 3 (a), virtual machine passes through virtual peripheral, such as virtual network port, and master
Machine establishes contact, and main frame sends number order to virtual machine, and virtual machine is parsed after getting order, then carries out corresponding
Relevant information, is returned to main frame by operation by this virtual peripheral.Fig. 3 (b) is the main frame and virtual machine that prior art provides
The interaction schematic diagram being monitored using agency, as shown in Fig. 3 (b), increase an Agent in virtual machine internal, it is simultaneously main
Machine and virtual machine directly establish a passage, such as virtual serial port etc., and main frame sends commands to virtual machine internal generation by passage
Reason, Agent is operated accordingly, and corresponding information is returned into main frame.
In above two method, the communication between virtual machine and main frame is realized dependent on the passage established, therefore can only
It can just be come into force in the case of virtual machine normal operation.But if exception occurs in virtual machine, passage is destroyed, then main frame
It will be unable to obtain any information of virtual machine.Especially when death situation state is hung in virtual machine processing, VMM can not detect that virtual machine is hung
Extremely, so as to being operated accordingly.
The content of the invention
The technical problem that the technical scheme provided according to embodiments of the present invention solves is how the VMM of host model detects void
The extension death situation state of plan machine.
A kind of detection KVM virtual machines provided according to embodiments of the present invention hang dead method, including：
By inquiring about virtual machine, the PC pointers of its respective thread of virtual machine are obtained；
Analyzed by the PC pointers of its respective thread to resulting virtual machine, determine whether the virtual machine is in
Hang death situation state；
If it is determined that the virtual machine, which is in, hangs death situation state, then virtual machine is reported to hang dead alarm information.
Preferably, it is described by inquiring about virtual machine, include the step of the PC pointers for its respective thread for obtaining virtual machine：
The relevant information of virtual machine is read from shared drive, the relevant information includes the identity ID number of virtual machine
With process number, virtual center processor VCPU thread numbers；
According to the ID number and process number of the virtual machine read, inquire about successively to each virtual machine；
According to the VCPU thread numbers of each virtual machine, the PC pointers of thread corresponding to the VPU thread numbers are inquired about.
Preferably, the PC pointers of its respective thread by resulting virtual machine are analyzed, and determine the void
Whether plan machine includes in extension death situation state：
The situation of change of the PC pointers of its respective thread of virtual machine is analyzed, and according to analysis result, it is determined that described
Whether virtual machine is in extension death situation state.
Preferably, if the PC pointers of its respective thread of virtual machine remain constant, it is determined that the virtual machine, which is in, to be hung
Death situation state.
Preferably, if the PC pointers of its respective thread of virtual machine keep constant in adjacent inquiry twice and analysis,
Determine that the virtual machine is in and hang death situation state.
The storage medium provided according to embodiments of the present invention, it is stored for realizing that above-mentioned detection KVM virtual machines are hung extremely
The program of method.
A kind of detection KVM virtual machines provided according to embodiments of the present invention hang dead device, including：
Enquiry module, for by inquiring about virtual machine, obtaining the PC pointers of its respective thread of virtual machine；
Analysis module, for being analyzed by the PC pointers of its respective thread to resulting virtual machine, it is determined that described
Whether virtual machine is in extension death situation state；
Notification module, however, it is determined that the virtual machine, which is in, hangs death situation state, then reports virtual machine to hang dead alarm information.
Preferably, the enquiry module read from shared drive virtual machine identity ID number including virtual machine and
The relevant information of process number, virtual center processor VCPU thread numbers, and according to the ID number and process number of the virtual machine read,
Inquire about successively to each virtual machine, according to the VCPU thread numbers of each virtual machine, inquire about thread corresponding to the VPU thread numbers
PC pointers.
Preferably, the analysis module is analyzed the situation of change of the PC pointers of its respective thread of virtual machine, and root
According to analysis result, determine the virtual machine whether in extension death situation state.
Preferably, the analysis module is when the PC pointers for analyzing its respective thread of virtual machine remain constant, really
The fixed virtual machine, which is in, hangs death situation state.
Preferably, the analysis module is equal when the PC pointers of its respective thread of virtual machine are in adjacent inquiry twice and analysis
When keeping constant, determine that the virtual machine is in and hang death situation state.
Technical scheme provided in an embodiment of the present invention has the advantages that：
The VMM of host model can accurately determine that virtual machine hangs death situation state by the inquiry and analysis to PC pointers.
Brief description of the drawings
Fig. 1 (a) is the physical machine model schematic that prior art provides；
Fig. 1 (b) is the Virtual Machine Model schematic diagram that prior art provides；
Fig. 2 (a) is the VMM schematic diagrames for the Hypervisor models that prior art provides；
Fig. 2 (b) is the VMM schematic diagrames for the mixed model that prior art provides；
Fig. 2 (c) is the VMM schematic diagrames for the host model that prior art provides；
Fig. 3 (a) is the interaction schematic diagram that the main frame that prior art provides and virtual machine are monitored using virtual unit；
Fig. 3 (b) is the interaction schematic diagram that the main frame that prior art provides and virtual machine are monitored using agency；
Fig. 4 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead method flow diagram；
Fig. 5 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead structure drawing of device；
Fig. 6 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead block diagram；
Fig. 7 is the state machine diagram of VCPU thread states provided in an embodiment of the present invention；
Fig. 8 is the flow chart of detection thread provided in an embodiment of the present invention.
Embodiment
Below in conjunction with accompanying drawing to a preferred embodiment of the present invention will be described in detail, it will be appreciated that described below is excellent
Select embodiment to be merely to illustrate and explain the present invention, be not intended to limit the present invention.
Fig. 4 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead method flow diagram, as shown in figure 1, step bag
Include：
Step S101：By inquiring about virtual machine, the PC pointers of its respective thread of virtual machine are obtained.
Before step S101 is performed, the relevant information for the virtual machine having been turned on is preserved to shared content, the correlation
Information includes the identity ID number and process number, virtual center processor VCPU thread numbers of virtual machine.
When performing step S101, the identity ID number including virtual machine of virtual machine is read first from shared drive
With process number, the relevant information of VCPU thread numbers, then according to the ID number and process number of the virtual machine read, inquiry pair successively
Each virtual machine, and according to the VCPU thread numbers of each virtual machine, inquire about the PC pointers of thread corresponding to the VPU thread numbers.
Step S102：Analyzed, determined described virtual by the PC pointers of its respective thread to resulting virtual machine
Whether machine is in extension death situation state.
When performing the step S102, the situation of change of the PC pointers of its respective thread of virtual machine is analyzed, and
According to analysis result, determine the virtual machine whether in extension death situation state.Furtherly, if the PC of its respective thread of virtual machine refers to
Pin remains constant, it is determined that the virtual machine, which is in, hangs death situation state.Furtherly, if the PC of its respective thread of virtual machine
Pointer keeps constant in adjacent inquiry twice and analysis, it is determined that the virtual machine, which is in, hangs death situation state.
Step S103：If it is determined that the virtual machine, which is in, hangs death situation state, then virtual machine is reported to hang dead alarm information.
After step S103 is performed, if having reported the virtual machine for hanging death situation state, in subsequent query and analysis, its is corresponding
The PC pointers of thread change, then report virtual machine to recover normal notification message.
The present embodiment utilizes PC pointer situations of change corresponding to each VCPU threads of virtual machine run in operating system,
Determine the virtual machine whether in extension death situation state.Specifically, KVM virtual machine is exactly a process, and each VCPU is exactly
One thread, as a Linux thread, each thread has the thread descriptor of oneself, has one in the thread descriptor
The variable of individual PC pointers, when thread is in operation, PC pointers can not stop to change, and run the thread function of correlation.If should
Thread is hung dead or thread and is scheduled out run mode, and the PC pointers will stop changing.Difference between the two is, if simply line
Journey is scheduled out run mode, then PC pointers can only keep PC pointers constant in a short period of time, scheduler program can quickly by
Thread puts into operation.And if the thread is hung extremely, then PC pointer president's time-preservings.
Can be with it will appreciated by the skilled person that realizing that all or part of step in above-described embodiment method is
The hardware of correlation is instructed to complete by program, described program can be stored in computer read/write memory medium, should
Program upon execution, including step S101 to step S103.Wherein, described storage medium can be ROM/RAM, magnetic disc, light
Disk etc..
Fig. 5 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead structure drawing of device, as shown in figure 5, including：
Enquiry module 10, for by inquiring about virtual machine, obtaining the PC pointers of its respective thread of virtual machine.Specifically,
Enquiry module 10 reads the identity ID number and process number including virtual machine of virtual machine, at virtual center from shared drive
The relevant information of device VCPU thread numbers is managed, and according to the ID number and process number of the virtual machine read, is inquired about successively to each void
Plan machine, according to the VCPU thread numbers of each virtual machine, inquire about the PC pointers of thread corresponding to the VPU thread numbers.
Analysis module 20, for being analyzed by the PC pointers of its respective thread to resulting virtual machine, determine institute
Virtual machine is stated whether in extension death situation state.Specifically, change of the analysis module 20 to the PC pointers of its respective thread of virtual machine
Whether situation is analyzed, and according to analysis result, determine the virtual machine in extension death situation state.Furtherly, analysis module
20 when the PC pointers for analyzing its respective thread of virtual machine remain constant, determine that the virtual machine is in and hang death situation state.
Furtherly, analysis module 20 is kept not when the PC pointers of its respective thread of virtual machine are in adjacent inquiry twice and analysis
During change, determine that the virtual machine is in and hang death situation state.
Notification module 30, however, it is determined that the virtual machine, which is in, hangs death situation state, then reports virtual machine to hang dead alarm information.
After reporting the dead alarm information of virtual machine extension, if having reported the virtual machine for hanging death situation state in subsequent query and analysis, its phase
The PC pointers of thread are answered to change, then notification module 30 reports virtual machine to recover normal notification message.
Fig. 6 is that detection KVM virtual machines provided in an embodiment of the present invention hang dead block diagram, as shown in fig. 6, monitoring process includes
Poll thread and alarm thread, wherein poll thread realize the function of enquiry module 10 and analysis module 20 in Fig. 5, alert thread
Realize the function of notification module 30 in Fig. 5.The virtual machine process and VCPU threads registered in poll thread pool shared drive, tool
Body is that the PC pointers of each VCPU threads of virtual machine process are inquired about, if some VCPU thread in virtual machine process be present
PC pointers remain constant, then illustrate to exist in virtual machine process VCPU threads and hang dead, i.e., virtual machine, which is in, hangs death situation state,
Now poll thread sends the message for including virtual machine information to alarm thread, informs that the alarm thread virtual machine hangs dead, alarm
After thread receives message, after being packed to the message received, VMM managing process is sent to, reports the extension of the virtual machine dead
Message.If poll thread in subsequent query, finds to have reported the PC pointers for each VCPU threads for hanging dead virtual machine to send out
Changing, then illustrate that the virtual machine is in normal condition, now poll thread is sent comprising virtual machine information to alarm thread
Message, inform that the alarm thread virtual machine recovers normal, after alarm thread receives message, the message received is packed
Afterwards, VMM managing process is sent to, reports the virtual machine to recover normal condition.
Specifically, the present embodiment starts when electric in host operating system, wrapped as a service (service)
Include a monitoring process in VMM layer.Because the monitoring process needs to monitor all virtual machines, so its startup must be
All virtual machines start before starting, and first, it can create a shared drive and be carried out for the monitoring process and virtual machine process
Data communication, the relevant information of each virtual machine is mainly deposited in the shared drive, include the ID number of virtual machine, virtual machine enters
Cheng Hao, VCPU number of virtual machine, and thread number corresponding to each VCPU (i.e. VCPU thread numbers) etc..These information are all by void
Plan machine process and VCPU threads are registered when starting.Meanwhile monitoring process can also start an alarm thread, the alarm thread
Wait for the message that monitoring process sends over.The information that monitoring process is registered by checking in shared drive is initiated to each void
The polling operation of plan machine, poll is actually and each VCPU threads is polled in each virtual machine, checks PC pointers
Situation.If it is judged that virtual machine has VCPU to hang extremely, then sends message (virtual machine information is included in message) and informs alarm line
Journey, after alarm thread receives message, message packing is carried out, being sent to VMM Virtual Machine Manager module, (module can be by VMM
Managing process is realized), report and hang dead message.
During monitoring process is polled, the letter for each virtual machine that monitoring process is recorded by inquiring about in shared drive
Cease to inquire about each virtual machine.Because virtual machine is as existing for a process in KVM, therefore virtual machine be present
The situation that process can be killed directly, so in queries firstly the need of judging that current virtual machine process whether there is, if not
In the presence of, then needs empty all information corresponding to the virtual machine, by the TID of each VCPU threads inquired in poll
Number (i.e. thread number) obtains the thread descriptor of each thread, so as to obtain the PC pointers of each VCPU threads.
Fig. 7 is the state machine diagram of VCPU thread states provided in an embodiment of the present invention, as shown in fig. 7, poll flow
In be that each VCPU threads set a state machine, each polling interval is defaulted as 10s (can pass through when whole service starts
Parameter sets this interval, minimum 10s).When carrying out first time poll, all state machines are defaulted as normal (the i.e. VCPU of CPU
State is normal), when carrying out second of poll, if PC pointers do not change (i.e. PC pointers are abnormal for the first time), then VCPU
State machine be just changed into abnormal (may be now that thread be scheduled out run mode, it is also possible to which thread is hung dead), but will not accuse
It is alert.When carrying out third time poll, if PC pointers change, VCPU state machine can revert to normally, if PC pointers do not have
Change (i.e. second of exception of PC pointers), then VCPU state machine is changed into hanging extremely.
If any one VCPU of a virtual machine be in hang death situation state, then be considered as the virtual machine be it is abnormal,
The dead message of virtual machine extension can be sent and give alarm thread.If in poll next time, VCPU state changes, then VCPU
State machine revert to normally, while send virtual machine recover message to inform alarm thread, VCPU state machines are by extension death situation state
Revert to normal condition.Otherwise, VCPU state machine continues hang death situation state, but be not sent further on virtual machine hang dead message to
Alert thread.
Alarm scanning thread (thread of alarming), the message of monitoring process is waited for after startup, receive the message of monitoring process
The Virtual Machine Manager module to VMM is transmitted afterwards, and VMM Virtual Machine Manager module is further processed for abnormal.
Fig. 8 is the flow chart of detection thread provided in an embodiment of the present invention, as shown in figure 8, step includes：
Step S201：Monitoring process creates shared content.
The shared drive created carries out data communication for monitoring process and virtual machine process, and it deposits each virtual machine
Relevant information, includes the ID number of virtual machine, the process number of virtual machine, VCPU number of virtual machine, and corresponding to each VCPU
Thread number etc..
Need to share what the information registerings such as the ID of virtual machine, process number to monitoring process created during virtual machine process initiation
In internal memory, while each VCPU threads of virtual machine process are also required to register relevant information on startup.
Step S202：Alarm thread is created in monitoring process.
The message that the alarm thread created waits monitoring process to send, if receiving message, sends out after the message is packed
Give VMM managing process.
Step S203：Monitoring process reads the virtual-machine data (i.e. the relevant information of virtual machine) in shared drive.
Step S204：Judge that current virtual machine whether there is, if in the presence of, perform step S205, otherwise perform step
S208。
Step S205：Each VCPU states in virtual machine are inquired about, state machine is set.
The information that monitoring process is registered by checking in shared drive initiates the polling operation to each virtual machine, every
Poll is actually and each VCPU threads is polled in individual virtual machine, checks the situation of PC pointers.
Step S206：Judge whether exception, step S207 is performed if abnormal, otherwise perform step S203.
Step S207：Monitoring process sends a warning message to alarm thread.If judging that it is dead that virtual machine has VCPU to hang,
So send message (virtual machine information is included in message) and inform alarm thread
After alarm thread receives message, message packing is carried out, is sent to VMM Virtual Machine Manager module, reports extension extremely to disappear
Breath.
Step S208：Virtual-machine data corresponding to removing.
The present embodiment realizes dead detection whether is hung to virtual machine under KVM modes, and the present embodiment is independent of virtual machine net
Network or the agency in virtual machine internal, but pass through the PC of thread corresponding to each VCPU in host's in-flight monitoring virtual machine
Whether pointer is not stopping to change, to judge the state of virtual machine.If do not changed for a long time, that is, think that the virtual machine has been hung
Extremely.
Although the present invention is described in detail above, the invention is not restricted to this, those skilled in the art of the present technique
Various modifications can be carried out according to the principle of the present invention.Therefore, all modifications made according to the principle of the invention, all should be understood to
Fall into protection scope of the present invention.
Claims (10)
1. a kind of detection KVM virtual machines hang dead method, including：
By inquiring about virtual machine, the PC pointers of its respective thread of virtual machine are obtained；
Analyzed by the PC pointers of its respective thread to resulting virtual machine, determine whether the virtual machine is dead in hanging
State；
If it is determined that the virtual machine, which is in, hangs death situation state, then virtual machine is reported to hang dead alarm information.
2. it is according to the method for claim 1, described by inquiring about virtual machine, obtain the PC pointers of its respective thread of virtual machine
The step of include：
The relevant information of virtual machine is read from shared drive, the relevant information includes the identity ID number of virtual machine and entered
Cheng Hao, virtual center processor VCPU thread numbers；
According to the ID number and process number of the virtual machine read, inquire about successively to each virtual machine；
According to the VCPU thread numbers of each virtual machine, the PC pointers of thread corresponding to the VPU thread numbers are inquired about.
3. method according to claim 1 or 2, the PC pointers of its respective thread by resulting virtual machine enter
Row analysis, determines whether the virtual machine includes in extension death situation state：
The situation of change of the PC pointers of its respective thread of virtual machine is analyzed, and according to analysis result, determined described virtual
Whether machine is in extension death situation state.
4. according to the method for claim 3, if the PC pointers of its respective thread of virtual machine remain constant, it is determined that institute
State virtual machine and be in extension death situation state.
5. according to the method for claim 4, if the PC pointers of its respective thread of virtual machine are inquired about and analyzed twice adjacent
When keep constant, it is determined that the virtual machine be in hang death situation state.
6. a kind of detection KVM virtual machines hang dead device, including：
Enquiry module, for by inquiring about virtual machine, obtaining the PC pointers of its respective thread of virtual machine；
Analysis module, for being analyzed by the PC pointers of its respective thread to resulting virtual machine, determine described virtual
Whether machine is in extension death situation state；
Notification module, however, it is determined that the virtual machine, which is in, hangs death situation state, then reports virtual machine to hang dead alarm information.
7. device according to claim 6, what the enquiry module read virtual machine from shared drive includes virtual machine
Identity ID number and process number, the relevant information of virtual center processor VCPU thread numbers, it is and virtual according to what is read
The ID number and process number of machine, inquire about successively to each virtual machine, according to the VCPU thread numbers of each virtual machine, inquire about the VPU
The PC pointers of thread corresponding to thread number.
8. the device according to claim 6 or 7, change of the analysis module to the PC pointers of its respective thread of virtual machine
Whether situation is analyzed, and according to analysis result, determine the virtual machine in extension death situation state.
9. device according to claim 8, the analysis module begins in the PC pointers for analyzing its respective thread of virtual machine
When keeping constant eventually, determine that the virtual machine is in and hang death situation state.
10. device according to claim 9, the analysis module its respective thread of virtual machine PC pointers adjacent two
When keeping constant when secondary inquiry and analysis, determine that the virtual machine is in and hang death situation state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610423949.1A CN107515772A (en) | 2016-06-15 | 2016-06-15 | A kind of detection KVM virtual machines hang dead method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610423949.1A CN107515772A (en) | 2016-06-15 | 2016-06-15 | A kind of detection KVM virtual machines hang dead method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107515772A true CN107515772A (en) | 2017-12-26 |
Family
ID=60720853
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610423949.1A Withdrawn CN107515772A (en) | 2016-06-15 | 2016-06-15 | A kind of detection KVM virtual machines hang dead method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107515772A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109634721A (en) * | 2018-12-17 | 2019-04-16 | 广东浪潮大数据研究有限公司 | A kind of the starting communication means and relevant apparatus of virtual machine and host |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101465770A (en) * | 2009-01-06 | 2009-06-24 | 北京航空航天大学 | Method for disposing inbreak detection system |
| CN102262537A (en) * | 2011-07-21 | 2011-11-30 | 复旦大学 | Method for processing abnormities of execution engine working in mixed model |
| CN104391754A (en) * | 2014-10-13 | 2015-03-04 | 北京星网锐捷网络技术有限公司 | Method and device for processing task exception |
| US9128743B1 (en) * | 2012-06-19 | 2015-09-08 | Bromium, Inc. | Graphic visualization of a suspended virtual machine |
| US20160164911A1 (en) * | 2014-12-09 | 2016-06-09 | At&T Intellectual Property I, L.P. | System and Method to Diffuse Denial-of-Service Attacks Using Virtual Machines |
-
2016
- 2016-06-15 CN CN201610423949.1A patent/CN107515772A/en not_active Withdrawn
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101465770A (en) * | 2009-01-06 | 2009-06-24 | 北京航空航天大学 | Method for disposing inbreak detection system |
| CN102262537A (en) * | 2011-07-21 | 2011-11-30 | 复旦大学 | Method for processing abnormities of execution engine working in mixed model |
| US9128743B1 (en) * | 2012-06-19 | 2015-09-08 | Bromium, Inc. | Graphic visualization of a suspended virtual machine |
| CN104391754A (en) * | 2014-10-13 | 2015-03-04 | 北京星网锐捷网络技术有限公司 | Method and device for processing task exception |
| US20160164911A1 (en) * | 2014-12-09 | 2016-06-09 | At&T Intellectual Property I, L.P. | System and Method to Diffuse Denial-of-Service Attacks Using Virtual Machines |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109634721A (en) * | 2018-12-17 | 2019-04-16 | 广东浪潮大数据研究有限公司 | A kind of the starting communication means and relevant apparatus of virtual machine and host |
| CN109634721B (en) * | 2018-12-17 | 2023-10-10 | 广东浪潮大数据研究有限公司 | Method and related device for starting communication between virtual machine and host |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104360878B (en) | A kind of method and device of application software deployment | |
| US20210111957A1 (en) | Methods, systems and apparatus to propagate node configuration changes to services in a distributed environment | |
| US11405464B2 (en) | Policy controlled semi-autonomous infrastructure management | |
| US9952891B2 (en) | Anomalous usage of resources by a process in a software defined data center | |
| TWI544328B (en) | Method and system for probe insertion via background virtual machine | |
| CN104321748B (en) | For catching the mthods, systems and devices of the error condition in light weight virtual machine manager | |
| US10521279B2 (en) | System and method for dynamic log level control | |
| US20140201357A1 (en) | Framework and method for monitoring performance of virtualized systems based on hardware base tool | |
| EP3140734A1 (en) | Mechanism for providing external access to a secured networked virtualization environment | |
| US8191062B2 (en) | System for processor frequency governors to govern a processor frequency by deriving CPU utilization information based on the state of virtual machine monitor | |
| EP3265910A1 (en) | Methods and apparatus to select virtualization environments for migration | |
| CN102790716A (en) | Techniques for securing a virtualized computing environment using a physical network switch | |
| US11258655B2 (en) | Holo-entropy based alarm scoring approach | |
| US9747119B2 (en) | Methods and apparatus to monitor virtual computing environments | |
| US9268591B2 (en) | Systems and methods for detecting system exceptions in guest operating systems | |
| US10476742B1 (en) | Classification of auto scaling events impacting computing resources | |
| CN105678164B (en) | Detect the method and device of Malware | |
| US10140141B2 (en) | Measuring accumulated load values of first level and second level virtual machines for modifying resource allocation | |
| CN111813497A (en) | Container environment anomaly detection method, device, medium and computer equipment | |
| EP3330855A1 (en) | Managing hardware resources | |
| CN107450962B (en) | Exception handling method, device and system in virtualized operation environment | |
| WO2017190083A1 (en) | Searching for information relating to virtualization environments | |
| CN110221816A (en) | Algorithm operation method, device, medium and algorithm platform based on algorithm platform | |
| CN115454636A (en) | Container cloud platform GPU resource scheduling method, device and application | |
| CN112313627A (en) | Mapping mechanism of events to serverless function workflow instances |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20171226 |
|
| WW01 | Invention patent application withdrawn after publication |