JP6582785B2 - Virtual machine management system, virtual machine management method and program - Google Patents

Description

  The present invention relates to a virtual machine management system, a virtual machine management method, and a program.

  Various techniques have been developed to increase the availability and reliability of the system. In a virtual machine, a technique called a fault tolerant (hereinafter, sometimes referred to as “FT”) function is being used as one of the purposes for improving availability. By the fault tolerant function, for example, when a failure occurs in a host that executes a virtual machine, the process of the virtual machine is taken over by another host prepared in advance, and the process of the virtual machine is continued. An example of the fault tolerant function is described in Non-Patent Document 1.

  Patent Document 1 describes a low-power consumption fault-tolerant computer system that can easily and quickly switch a system without using special hardware.

  Patent Document 2 describes a method for determining a migration destination of a virtual machine. With the technique described in Patent Document 2, a user can define not only a migration policy for optimizing the entire system but also a migration policy depending on an application. In the technique described in Patent Document 2, it is controlled to move to another server while the virtual machine on which the application is operating is operating while considering both policies.

JP 2012-213321 A JP 2009-116852 A

Brendan Cully, Geoffrey Lefebvre, Dutch Meyer, Mike Feeley, Norm Hutchinson, and Andrew Warfield. 2008.Remus: high availability via asynchronous virtual machine replication.In Proceedings of the 5th USENIX Symposium on Networked Systems Design and Implementation (NSDI'08), Jon Crowcroft and Mike Dahlin (Eds.). USENIX Association, Berkeley, CA, USA, 161-174.

  In a virtual machine having a fault tolerant function, processing with a large load may be performed in relation to the fault tolerant function. Therefore, the fault-tolerant function may affect the operation of the virtual machine, such as providing the service executed by the virtual machine to the outside. However, with the techniques described in each patent document or the like, it is difficult to reduce the influence on the operation of the virtual machine due to the fault tolerant function.

  The present invention has been made to solve the above-described problems, and has as its main object to provide a virtual machine management system and the like that have a small influence on the operation of the virtual machine due to the fault tolerant function.

The virtual machine management system according to an aspect of the present invention includes a service management table that holds conditions regarding the state of the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine;
A virtual machine management system comprising: control means for controlling the state of the fault tolerant function of the virtual machine based on the condition.


In the virtual machine management method according to one aspect of the present invention, at least one of the services is a condition based on a condition for enabling the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine. Whether the condition is satisfied is determined, and control is performed to change the state related to the fault tolerant function of the virtual machine based on the determination result.

  Further, the program according to one aspect of the present invention is based on the condition that the fault tolerant function for each of at least one service that is the target of the fault tolerant function of the virtual machine is enabled in the computer. A process for determining whether or not the condition is satisfied and a process for controlling to change the state related to the fault tolerant function of the virtual machine based on the determination result are executed.

  ADVANTAGE OF THE INVENTION According to this invention, the virtual machine management system etc. with a small influence on operation | movement of the virtual machine resulting from a fault tolerant function can be provided.

It is a figure which shows the structure of the virtual machine etc. which are the premise of each embodiment of this invention. It is a figure which shows the structure of the virtual machine management system in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the virtual machine management system in the 1st Embodiment of this invention. It is a flowchart which shows the modification of operation | movement of the virtual machine management system in the 1st Embodiment of this invention. It is a figure which shows the structure of the virtual machine management system in the 2nd Embodiment of this invention. It is a figure which shows one structural example of the information processing apparatus which implement | achieves the virtual machine management system etc. in each embodiment of this invention.

  Embodiments of the present invention will be described with reference to the accompanying drawings. In each embodiment of the present invention, each component of each system represents a functional unit block. Each component of each system is realized by, for example, an arbitrary combination of an information processing apparatus 500 and a program as shown in FIG. The information processing apparatus 500 includes the following configuration as an example.

CPU (Central Processing Unit) 501
ROM (Read Only Memory) 502
-RAM (Random Access Memory) 503
A program 504 loaded into the RAM 503
A storage device 505 for storing the program 504
A drive device 507 for reading / writing the recording medium 506
Communication interface 508 connected to the communication network 509
An input / output interface 510 for inputting / outputting data
-Bus 511 connecting each component
Each component of each system in each embodiment is realized by the CPU 501 acquiring and executing a program 504 that realizes these functions. The program 504 that realizes the function of each component of each device is stored in advance in the storage device 505 or the RAM 503, for example, and is read by the CPU 501 as necessary. Note that the program 504 may be supplied to the CPU 501 via the communication network 509 or may be stored in the recording medium 506 in advance, and the drive device 507 may read the program and supply it to the CPU 501. In addition, there are various modifications in the method of realizing each device. For example, each system in each embodiment can be realized by any combination of various hardware including a circuit different from the information processing apparatus 500 and a program.

  Prior to the description of each embodiment of the present invention, a fault tolerant function and its operation in a virtual machine which is a premise of each embodiment will be described. FIG. 1 is a diagram illustrating a configuration of a virtual machine or the like that is a premise of each embodiment.

  In the fault tolerant function, when a failure occurs in the hardware that executes the virtual machine, other virtual machines that are executed by other hardware continue to perform the processing that the virtual machine is performing. It is a function. In each embodiment of the present invention, the following operation is assumed as one of the operations for realizing the fault tolerant function.

  First, the state of a virtual machine that is running in a standby virtual machine is stored in another virtual machine that is executed by other hardware or the like at every predetermined interval called a checkpoint. When a failure occurs in the hardware that executes the operating virtual machine, the standby virtual machine takes over the processing performed by the operating virtual machine using the stored information. Run with. In a general fault-tolerant function, execution of a virtual machine is continued as it is when processing is taken over. By using such a fault tolerant function, the availability of the virtual machine can be increased.

  Hereinafter, a state in which the state of the virtual machine is stored at the checkpoint and processing can be taken over by another virtual machine when a failure occurs is referred to as a fault tolerant function being effective. In addition, a state in which the state of the virtual machine is not saved and a process is not taken over by another virtual machine when a failure occurs is referred to as an invalid fault tolerant function.

  An example of the operation for realizing the fault tolerant function is specifically as follows. In the example shown in FIG. 1, hosts 10-1 and 10-2 are connected via a communication network 50. The hosts 10-1 and 10-2 are, for example, the information processing apparatus 500 illustrated in FIG. 6 described above, but are not limited thereto. The communication network 50 is a network having an arbitrary configuration that enables data transfer between the hosts 10-1 and 10-2.

  In the hosts 10-1 and 10-2, virtual machine management systems 100-1 and 100-2 are executed, respectively. The virtual machine management systems 100-1 and 100-2 are general systems that realize a virtual machine in each of the hosts 10-1 and 10-2. The virtual machine management systems 100-1 and 100-2 are also referred to as a virtual machine management (VMM) or a hypervisor. The virtual machine management systems 100-1 and 100-2 have a configuration necessary for executing a general fault-tolerant function.

  Each of the virtual machines 20 and 40 is a virtualized computer system. The virtual machine 20 is executed by the virtual machine management system 100-1. The virtual machine 40 is executed by the virtual machine management system 100-2. The virtual machine 20 or 40 includes a guest physical memory 21 or 41, respectively. The guest physical memory 21 or 41 functions as a physical memory of the virtual machine 20 or 40, respectively. Each of the guest physical memories 21 or 41 is realized using a physical memory or the like included in the host 10-1 or 10-2.

  In the example illustrated in FIG. 1, the virtual machine 20 is operating, and a guest OS (Operating System) 30 operates in the virtual machine 20. Further, the virtual machine 40 is on standby and can take over the operation of the guest OS 30. That is, the virtual machine 40 does not perform any special operation when the guest OS 30 is operating normally in the virtual machine 20. When a problem occurs in at least one of the host 10-1 or the virtual machine management system 100-1, the virtual machine 40 starts operating. In this case, the guest OS 30 takes over the operation of the virtual machine 40 by the fault tolerant function provided in the virtual machine management system 100-2. The guest OS 30 continues to operate on the virtual machine 40.

  In order to enable the fault tolerant function, the virtual machine management systems 100-1 and 100-2 store the state of the virtual machine 20. The virtual machine management system 100-1 stores the state of the virtual machine 20 in the virtual machine 40 via the virtual machine management system 100-2 at a timing called the check point described above. Checkpoints are set at regular intervals, for example. The checkpoint interval described above is appropriately determined according to the type of service executed in the virtual machine 20. When saving the state of the virtual machine 20, the virtual machine management system 100-1 temporarily stops the operation of the virtual machine 20 and saves the state.

  The virtual machine management system 100-1 stores, for example, the contents of the guest physical memory 21 included in the virtual machine 20 in the guest physical memory 41 of the virtual machine 40 as the state of the virtual machine 20. As the state of the virtual machine 20, the state of the CPU that the virtual machine 20 includes virtually, the contents of the storage device that the virtual machine 20 includes virtually, and the like may be stored. When the contents of the guest physical memory 21 are stored as the state of the virtual machine 20, the virtual machine management system 100-1 updates the guest physical memory 21 updated after the state of the virtual machine 20 is stored at the latest checkpoint. Save the contents of. That is, the area of the memory 21 in which writing has been performed after the state of the virtual machine 20 has been saved at the most recent checkpoint becomes a storage target at the next checkpoint.

  In general, in the virtual machine 20 or 40, the contents of the guest physical memory 21 or 41 are managed in a certain unit called a page. Therefore, in the guest physical memory 21, it is determined on a page basis whether or not writing has been performed. Further, the virtual machine management system 100-1 stores the contents of the guest physical memory 21 in units of pages. A page that is written and is to be saved at the next checkpoint is called a dirty page.

  In the example shown in FIG. 1, a plurality of quadrangular figures included in an area enclosed as the guest physical memory 21 represent each page in the guest physical memory 21. In the example shown in FIG. 1, each of the figures painted in black among the figures represents a dirty page. In addition, a plurality of quadrangular figures included in the area enclosed as the transfer buffer 121 represent dirty pages to be saved at checkpoints.

  As described above, when the fault tolerant function is valid and the state of the virtual machine 20 is saved, the virtual machine 20 temporarily stops its operation. The operation of the virtual machine 20 is resumed after the storage of the state of the virtual machine 20 in the virtual machine 40 is completed. That is, the operation of the virtual machine 20 is suspended until the state of the virtual machine 20 is stored in the virtual machine 40. When the virtual machine 40 stores the state of the virtual machine 20 in the guest physical memory 41, it is necessary to transfer data regarding the state of the virtual machine 20 from the host 10-1 to the host 10-2. On the other hand, the data copy in the memory provided in the host 10-1 is executed at a higher speed than the data transfer from the host 10-1 to the host 10-2.

  Therefore, in order to shorten the period during which the operation of the virtual machine 20 is temporarily stopped, the virtual machine management system 100-1 first stores the state of the virtual machine 20 in the transfer buffer 121. As will be described later, the transfer buffer 121 is, for example, a memory included in the virtual machine management system 100-1. When the state of the virtual machine 20 is saved in the transfer buffer 121, the virtual machine 20 resumes operation. In conjunction with the resumption of the operation of the virtual machine 20, the virtual machine management system 100-1 stores the state of the virtual machine 20 stored in the transfer buffer 121 in the guest physical memory 41 of the virtual machine 40. By doing so, the period during which the operation of the virtual machine 20 is temporarily stopped is compared with the case where the state of the virtual machine 20 is directly transferred to the guest physical memory 41 and stored in the guest physical memory 41. A short period. When the same page is repeatedly a dirty page, the virtual machine management system 100-1 transfers the difference between the content at that time and the content at the latest checkpoint to the virtual machine 40 with respect to the dirty page. Good.

  Further, when the state of the virtual machine 20 is stored as described above, communication from the virtual machine 20 to the outside is suspended until the storage of the state of the virtual machine 20 in the virtual machine 40 is completed. That is, communication to the outside that occurs after the state of the virtual machine 20 is saved at the latest checkpoint is not transmitted to the outside until the state of the virtual machine 20 is saved at the next checkpoint. Deferred. The suspended communication is actually executed after the storage of the state of the virtual machine 20 in the virtual machine 40 is completed.

  This operation is based on the following reason. When a failure or the like occurs in the host 10 or the like, and the processing of the virtual machine 20 is taken over to the virtual machine 40, the state of the virtual machine 20 saved at the latest checkpoint is taken over to the virtual machine 40. Therefore, when communication is performed in the virtual machine 20 after the state of the virtual machine 20 is saved at the latest checkpoint, there is a possibility that communication is performed in a duplicated manner when the virtual machine 40 is subsequently taken over. Occurs. There is a possibility that a contradiction occurs between the communication performed in the virtual machine 20 and the communication performed in the virtual machine 40. From this, the communication is suspended as described above.

  By the way, when the state of the virtual machine 20 is saved, the period during which the virtual machine 20 is temporarily stopped may become longer depending on the number of dirty pages. A program that is executed in the virtual machine 20 and provides some service may start providing the service after storing large-scale data in the guest physical memory 21 of the virtual machine 20. Services that are executed by such programs include services that create data caches such as databases and Web services.

  In this case, when the state of the virtual machine 20 is saved, many dirty pages are to be saved. Therefore, as the state of the virtual machine 20 is saved, the period during which the operation of the virtual machine 20 is temporarily stopped becomes longer. As a result, there may be a problem that the start of service provision by a program executed in the virtual machine 20 is delayed. In other words, due to the fault tolerant function, there is a possibility that an adverse effect such as a delay in the start of provision of services to the outside using the virtual machine 20 may occur. The virtual machine management system in each embodiment of the present invention can reduce the above-described adverse effects.

(First embodiment)
Next, a first embodiment of the present invention will be described. FIG. 2 is a diagram showing a virtual machine management system according to the first embodiment of the present invention. FIG. 3 is a flowchart illustrating an operation example of the virtual machine management system according to the first embodiment of the present invention. FIG. 4 is a flowchart illustrating a modified example of the operation of the virtual machine management system according to the first embodiment of the present invention.

  As shown in FIG. 2, the virtual machine management system 100 according to the first embodiment of the present invention includes at least a control unit 110, a service management table 111, and a checkpoint control unit 112. The virtual machine management system 100 according to the first embodiment of the present invention includes a memory transfer control unit 120, a transfer buffer 121, a network control unit 130, a communication analysis unit 131, and a hold buffer 132.

  The virtual machine management system 100 in this embodiment is used as, for example, the virtual machine management system 100-1 or 100-2 shown in FIG. That is, the virtual machine management system 100 is executed by the host 10-1 or 10-2 shown in FIG. Further, the virtual machine management system 100 manages execution of a virtual machine such as the virtual machine 20 or 40.

  Hereinafter, it is assumed that a virtual machine similar to the virtual machine 20 is executed in the virtual machine management system 100. In the virtual machine management system 100, a virtual machine similar to the virtual machine 40 may be executed. In the following description of each component of the virtual machine management system 100, “virtual machine” refers to a virtual machine executed in the virtual machine management system 100.

  Further, in the following description, execution of a program that provides an arbitrary type of service as an application program in the guest OS of the virtual machine may be referred to as execution of the service in the virtual machine. Furthermore, the availability of the service from an external server or the like or an external user may be referred to as providing the service to the outside.

  Next, each component of the virtual machine management system 100 according to the first embodiment of the present invention will be described.

  The control unit 110 determines whether or not at least one of the programs that provide the service satisfies a condition held in the service management table 111, and controls the state of the virtual machine. The control unit 110 controls the state of the virtual machine so as to enable the fault tolerant function of the virtual machine when at least one of the services executed in the virtual machine satisfies a condition held in the service management table 111 described later. To do. In addition, the control unit 110 performs control necessary for realizing a fault tolerant function, such as storage of the state of a virtual machine.

  In addition, the control unit 110 virtually disables the fault tolerant function when, for example, any of the above services stops or when the above services do not satisfy the conditions held in the service management table 111. Control the state of the machine.

  The condition for enabling the fault tolerant function includes any condition related to the operation of the program that provides the service. Details of the conditions will be described later.

  The control unit 110 determines whether or not at least one of the services satisfies a condition held in the service management table 111 by an arbitrary procedure. For example, the control unit 110 determines whether at least one of the services satisfies the above-described condition based on the result of monitoring by another component of the virtual machine management system 100 that monitors the operation of the program that provides the service.

For example, the control unit 110 determines that at least one of the services satisfies the above-described condition based on a request that the fault tolerant function should be enabled from another component that monitors the operation of the program that provides the service. May be. In addition, another component that monitors the operation of the program that provides the service writes the monitoring result in the service management table 111, and the control unit 110 refers to the result so that at least one of the services satisfies the condition described above. It may be determined whether or not
In addition, the control unit 110 may directly monitor the operation of a program that provides a service executed in the virtual machine and determine whether the service satisfies the above-described condition.

  As another example, the control unit 110 determines whether or not the service satisfies the above-described conditions based on the operation status related to the communication of the program that provides the service analyzed by the communication analysis unit 131 described later. Also good.

  When it is determined that the service satisfies the above-described condition, the control unit 110 appropriately controls the memory transfer control unit 120, the network control unit 130, and the like to enable the fault tolerant function. By performing this control, the state of the virtual machine is saved in a standby virtual machine executed by another host at intervals specified in advance as checkpoints.

  If the fault tolerant function is valid and the service unit 110 determines that the service does not satisfy the above-described condition, the control unit 110 performs control so as to invalidate the fault tolerant function. For example, the control unit 110 disables the fault tolerant function by appropriately controlling the memory transfer control unit 120, the network control unit 130, and the like. In other words, the processing related to the storage of the state of the virtual machine, which has been performed at intervals designated as checkpoints, is stopped.

  Executing a fault tolerant function may require a large load. Therefore, when the fault tolerant function is effective in the virtual machine management system 100, as described above, there is a possibility that adverse effects such as a delay in the start of provision of services executed in the virtual machine will occur. is there.

  On the other hand, the necessity of the fault tolerant function is high when a service is provided to the outside, but the necessity is sometimes small in other cases. Therefore, it is considered that the fault tolerant function is preferably made effective based on the necessity and the magnitude of the load when the program providing the service operates.

  Therefore, in the virtual machine management system 100 according to the present embodiment, the control unit 110 determines whether or not at least one of the services satisfies the condition held in the service management table 111, and controls the state of the virtual machine. . By appropriately setting the conditions, the control unit 110 can be enabled only when the necessity of the fault tolerant function is high. That is, the control by the control unit 110 makes it possible to invalidate the fault tolerant function when a process with a large load but a small necessity for the fault tolerant function is being executed in the virtual machine.

Therefore, when the control unit 110 performs the above-described control, it is possible to reduce the influence on the operation of the virtual machine due to the fault tolerant function while increasing the availability of the virtual machine.

  The service management table 111 holds a condition regarding the status of whether the fault tolerant function is valid or invalid for each of at least one service that is a target of the fault tolerant function of the virtual machine.

  As an example, the service management table 111 holds conditions for enabling the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine.

  As described above, the condition for enabling the fault tolerant function includes any condition related to the execution of the program that provides the service. This condition includes, for example, a condition related to execution of a program that provides a service, a condition related to whether or not a service is provided to the outside, and the like. As specific examples, these conditions are defined in the form of the amount of memory required for executing the program that provides the service, the type of communication performed for providing the service to the outside, and the like.

  Note that the service management table 111 may hold conditions for disabling the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine.

  In the present embodiment, as an example, the service management table 111 holds a condition regarding whether or not a service is provided to the outside for each service executed in a virtual machine. As a condition regarding whether or not the service is provided to the outside, the type of communication performed regarding the provision of the service to the outside is used.

  That is, the service management table 111 holds information related to communication transmitted / received with respect to the service when the provision of the service to the outside is started as a condition for enabling the fault tolerant function. For example, this information includes a protocol for communication performed when starting provision of a service to the outside, a packet type, an address and a port number related to the communication, and the like.

  When TCP / IP (Transmission Control Protocol / Internet Protocol) is used as a communication protocol, the service management table 111 holds, for example, either TCP or UDP (User Datagram Protocol) as the protocol type. . In this case, the service management table 111 holds an IP (Internet Protocol) address such as a service request source or a service destination as an address.

  Further, the service management table 111 may hold a valid or invalid state of the fault tolerant function. The service management table 111 may hold information on services and conditions that cause the fault tolerant function to be effective when the fault tolerant function is enabled. That is, this information is information indicating, for example, which service caused the fault tolerant function to be effective.

  Furthermore, the service management table 111 may hold information on checkpoints when the checkpoint control unit 112 described later changes checkpoint intervals. The information about checkpoints includes information about checkpoint intervals suitable for each service for one service that is the target of the fault tolerant function of the virtual machine.

  The appropriate checkpoint interval may vary depending on the services running on the virtual machine. When the state of the virtual machine is saved, communication to the outside is suspended as described above. Since the period during which communication is suspended becomes longer, there is a possibility that trouble may occur in communication to the outside when the frequency of communication to the outside is high. Therefore, for example, when the frequency of external communication is high, it is generally preferable to set a short checkpoint interval.

  On the other hand, it is generally known that there is a time locality in updating a page when a program is executed. In other words, when the program is executed, the same page may be repeatedly updated during a certain period. In such a case, by setting a long checkpoint interval, it is possible to reduce the total amount of pages to be updated as dirty pages when saving the state of the virtual machine. That is, when the frequency of writing to the memory is high, it is generally preferable to set a long checkpoint interval.

  When the fault tolerant function is valid, the checkpoint control unit 112 controls checkpoint intervals and processing at checkpoints as necessary. For example, the checkpoint control unit 112 changes the checkpoint interval according to the checkpoint interval specified for each of the services specified in the service management table 111 in response to an instruction from the control unit 110. I do.

  The memory transfer control unit 120 controls transfer of the contents of the guest physical memory from the running virtual machine to the standby virtual machine at the checkpoint. Further, the memory transfer control unit 120 may perform general control necessary for realizing the virtual machine, such as control of access to the guest physical memory in the virtual machine.

  The transfer buffer 121 is a buffer that temporarily holds the contents of the guest physical memory of the virtual machine when the virtual machine executed in the virtual machine management system 100 is operating. The memory transfer control unit 120 and the transfer buffer 121 are the same elements as those included in a virtual machine management system that can realize a general fault-tolerant function.

  When the fault tolerant function is valid, the network control unit 130 controls the above-described communication suspension operation and the like. The network control unit 130 may perform general control necessary for realizing the virtual machine, such as control of communication with the guest OS executed in the virtual machine.

  The communication analysis unit 131 analyzes the content of communication in the guest OS executed in the virtual machine when the communication management condition is defined as a condition for enabling the fault tolerant function in the service management table 111 described above. . For example, the communication analysis unit 131 analyzes whether the communication is related to the start or end of a service executed in the guest OS. When the communication in the guest OS is communication by TCP (Transmission Control Protocol), the communication analysis unit 131 analyzes the content of the communication by determining the type of the packet.

  The hold buffer 132 is a buffer that holds information regarding the held communication when the communication is put on hold when the fault tolerant function is enabled.

  Next, the operation of the virtual machine management system 100 in the first embodiment of the present invention will be described. In this operation example, it is assumed that a condition regarding communication is defined in the service management table 111 as an example of a condition for enabling the fault tolerant function. That is, the control unit 110 determines whether or not a condition for enabling the fault-tolerant function is satisfied based on the analysis result in the communication analysis unit 131.

  First, the communication analysis unit 131 analyzes communication performed in the guest OS (step S101). For example, the communication analysis unit 131 analyzes the content of communication performed by the guest OS via the network control unit 130.

  Next, the communication analysis unit 131 determines whether or not the communication analyzed in step S101 is communication related to the start of service provision to the outside performed by the guest OS of the virtual machine (step S102). For example, when the communication is TCP communication, the communication analysis unit 131 is communication related to the start of service provision to the outside when the protocol type of communication from the guest OS to be analyzed is TCP_SYN_ACK. Is determined. If the communication is based on UDP, the communication analysis unit 131 determines that the communication is related to the start of service provision to the outside at the stage of this step because the determination is difficult.

  In step S102, when it is determined that the communication to be analyzed is communication related to the start of service provision to the outside (step S102: Yes), the control unit 110 and the communication analysis unit 131 perform the process of step S103. Execute. In step S103, the control unit 110 determines whether or not a condition for enabling the fault tolerant function is satisfied. In this operation example, the control unit 110 makes a determination based on the analysis result in the communication analysis unit 131.

  Whether or not the condition for enabling the fault tolerant function is satisfied is determined by whether or not the communication analyzed in step S101 or step S102 corresponds to the condition held in the service management table 111 It is done by judging. That is, it is performed by comparing whether the address and port number related to the communication correspond to the conditions for enabling the fault tolerant function held in the service management table 111.

  As an example of the operation in step S103, the communication analysis unit 131 compares whether or not the above-described communication satisfies a condition held in the service management table 111. When the condition is met, the communication analysis unit 131 requests the control unit 110 to enable the fault tolerant function. Then, when receiving the request, the control unit 110 determines that the condition for enabling the fault tolerant function is satisfied.

  As another operation example in step S <b> 103, the communication analysis unit 131 may transmit information related to the above communication to the control unit 110. In this case, the control unit 110 may compare whether the transmission satisfies a condition held in the service management table 111.

  When it is determined in step S103 that the condition for enabling the fault tolerant function is satisfied (step S103: Yes), the control unit 110 updates the service management table 120 (step S104). For example, the control unit 110 updates the service management table 120 so as to hold information on services and conditions that cause the fault tolerant function to be effective based on the comparison result in step S103. Further, the control unit 110 may update the service management table 120 so as to indicate that the fault tolerant function is effective.

  The control unit 110 confirms whether the fault tolerant function has already been enabled (step S105). When the fault tolerant function is not enabled (step S105: No), the control unit 110 performs control to enable the fault tolerant function (step S106).

  When the fault tolerant function is already enabled (step S105: Yes), the checkpoint control unit 112 controls to change the interval between checkpoints (step S107). In this case, the checkpoint control unit 112 refers to the service management table 111, for example, and acquires information regarding checkpoint intervals suitable for the service related to the communication to be compared in step S103 and the like. If the checkpoint interval suitable for the service is longer than the checkpoint interval at that time, the checkpoint control unit 112 changes the checkpoint interval to an interval suitable for the service. To do.

  When the checkpoint interval is changed in step S107, the checkpoint control unit 112 may perform control so as to change the checkpoint interval by a procedure different from the above-described procedure. For example, the checkpoint control unit 112 determines the checkpoint interval so as to be an interval suitable for the service with the highest priority according to an arbitrary priority set for the service executed in the virtual machine. May be controlled.

  Subsequent to the processing in step S106 or S107, the network control unit 130 executes processing related to communication as necessary (step S108). The network control unit 130 executes processing related to communication including processing necessary when the fault tolerant function is effective, such as suspension of communication when the fault tolerant function is effective.

  In step S103, also when it is determined in step S103 that the conditions for enabling the fault tolerant function are not satisfied (step S103: No), the network control unit 130 executes the process of step S108 as necessary. To do.

  If it is determined in step S102 that the communication to be analyzed is not communication related to the start of service provision to the outside (step S102: No), the process of step S109 is performed. In step S109, the communication analysis unit 131 determines whether or not the communication analyzed in step S101 is communication related to termination of service provision to the outside performed by the guest OS of the virtual machine. If the communication is TCP communication, the communication analysis unit 131 determines that the communication is related to termination of service provision to the outside when the protocol type is TCP_RST or TCP_FIN, for example.

  In step S109, when it is determined that the communication to be analyzed is communication related to termination of service provision to the outside (step S109: Yes), the control unit 110 and the communication analysis unit 131 perform the process of step S110. Execute. In step S103, the control unit 110 determines whether or not a condition for disabling the fault tolerant function is satisfied.

  In this operation example, the control unit 110 performs the determination based on the analysis result in the communication analysis unit 131 in the same manner as the processing in step S103 as the processing in step S110. That is, it is performed by comparing whether the address and the port number related to the communication correspond to the conditions for disabling the fault tolerant function held in the service management table 111.

  When it is determined in step S110 that the condition for enabling the fault tolerant function is satisfied (step S110: Yes), the control unit 110 updates the service management table 120 (step S111). For example, the control unit 110 updates the service management table 120 based on the result of the comparison in step S110 so as to hold information on services and conditions that cause the fault tolerant function to be invalidated.

  Subsequently, the control unit 110 confirms whether other services that should enable the fault tolerant function are provided to the outside by referring to the service management table 111 (step S112). If another service is provided (step S112: Yes), the control unit 110 keeps the fault tolerant function enabled. Then, the checkpoint control unit 112 performs control so as to change the checkpoint interval as necessary in step S107.

  In the case where provision of other services that should enable the fault tolerant function is not performed outside, the control unit 110 performs control to disable the fault tolerant function (step S113). In conjunction with the process of step S113, the control unit 110 may update the service management table 120 to indicate that the fault tolerant function is invalid. After the processing in step S113, the network control unit 130 executes processing related to communication in step S108 as necessary.

  As described above, the virtual machine management system 100 according to the first embodiment of the present invention controls the virtual machine so as to enable the fault tolerant function when the condition for enabling the fault tolerant function is satisfied. In other words, in the present embodiment, the control unit 110 of the virtual machine management system 100 controls the virtual machine so as to enable the fault tolerant function when the conditions specified in the service management table 111 are satisfied. The conditions defined in the service management table 111 are conditions indicating that a service is provided to the outside in the virtual machine, for example.

  That is, in the virtual machine management system 100 of the present embodiment, the fault tolerant function can be enabled when the service is provided to the outside in the virtual machine according to the conditions held in the service management table 111. It becomes. In other words, in the virtual machine management system 100 of the present embodiment, the fault tolerant function may be invalidated when a process with a large load and a small possibility of affecting the outside is performed. It becomes possible.

  Therefore, it is possible to reduce the load on the virtual machine due to the fault tolerant function, such as temporary stop of the virtual machine accompanying the storage of the state of the virtual machine. As a result, it is possible to avoid delaying the provision of services to the outside due to the storage of the state of the virtual machine.

  That is, the virtual machine management system 100 according to this embodiment can provide a virtual machine management system or the like that has a small influence on the operation of the virtual machine due to the fault tolerant function.

(Modification of the first embodiment)
Various modifications are conceivable in the first embodiment of the present invention.

  The virtual machine management system 100 according to the present embodiment may include a configuration different from the configuration illustrated in FIG. 2 as a configuration required when realizing a virtual machine. Further, the virtual machine management system 100 may include a configuration different from the configuration illustrated in FIG. 2 as a configuration necessary for realizing a general fault-tolerant function. By including at least the control unit 110 and the service management table 111, the virtual machine management system 100 can reduce the influence on the operation of the virtual machine due to the fault tolerant function.

  As another example, the virtual machine management system 100 according to this embodiment may not change the checkpoint interval. In this case, when the fault tolerant function is valid, the virtual machine management system 100 stores the state of the virtual machine at a predetermined checkpoint interval. That is, the virtual machine management system 100 in this case may operate as in the flowchart shown in FIG.

  In the flowchart shown in FIG. 4, the operation relating to the adjustment of the checkpoint interval defined in S107 of the flowchart shown in FIG. 3 is omitted. If the branch defined in step S105 or step S112 is “Yes”, processing relating to communication defined in step S108 is subsequently performed. The operation of the flowchart shown in FIG. 4 is the same as that of the flowchart shown in FIG. 3 except for the points described above.

  In this case, the checkpoint control unit of the virtual machine management system 100 controls the storage of virtual machine information at the checkpoint.

  Even in such a case, the virtual machine management system 100 can affect the operation of the virtual machine by enabling the fault tolerant function when the control unit 110 satisfies the conditions specified in the service management table 111. Can be reduced.

(Second Embodiment)
Next, a second embodiment of the present invention will be described. FIG. 5 is a diagram showing a configuration of the virtual machine management system 200 in the second exemplary embodiment of the present invention.

  As shown in FIG. 5, the virtual machine management system 200 according to the second embodiment of the present invention has the same configuration as the virtual machine management system 100 according to the first embodiment of the present invention. In the virtual machine management system 200, the control unit 110 controls the state related to the fault tolerant function of the virtual machine based on the monitoring result of the virtual machine in the service monitoring unit 240.

  In general, a guest OS of a virtual machine may introduce a virtual machine management system or a program that supports the virtual machine. In the present embodiment, the service monitoring unit 240 is an example of such a program. The virtual machine management system 200 controls the virtual machine using the monitoring result in the service monitoring unit 240.

  In the present embodiment, the service monitoring unit 240 monitors the state of a process related to a service executed on the guest OS as one example of operation. That is, the service monitoring unit 240 monitors the start and end of these processes. A service or the like related to a process to be monitored by the service monitoring unit 240 may be appropriately defined in the service management table 111.

  When the process to be monitored is activated, the service monitoring unit 240 requests the control unit 110 to enable the fault tolerant function. Then, when receiving the request, the control unit 110 determines that the condition for enabling the fault tolerant function is satisfied.

  Alternatively, the service monitoring unit 240 may notify the control unit 110 when a process is started in the virtual machine. In this case, the control unit 110 compares whether the process to be notified is a process related to the condition held in the service management table 111, and determines that the condition for enabling the fault tolerant function is satisfied. May be.

  Similarly, when the process to be monitored is completed, the service monitoring unit 240 requests that the control unit 111 disable the fault tolerant function. Then, when receiving the request, the control unit 110 determines that the condition for disabling the fault tolerant function is satisfied as necessary while referring to the condition for enabling the other fault tolerant function. Alternatively, the service monitoring unit 240 may notify the control unit 110 when the process is completed in the virtual machine. Based on the notification, the control unit 110 may determine that the condition for enabling the fault tolerant function is satisfied as necessary.

As another example, the service monitoring unit 240 monitors the memory usage of the process as the state of the process related to the service executed in the guest OS.
In this case, the service monitoring unit 240 monitors, for example, whether or not a memory usage predetermined in the service management table 111 has been exceeded. Then, based on the monitoring result of the service monitoring unit 240, the control unit 110 satisfies the condition that the fault tolerant function should be enabled (or disabled) in the same procedure as the procedure in the above example. to decide.

  In the present embodiment, the communication analysis unit 131 of the virtual machine management system 200 analyzes the content of communication in the guest OS executed in the virtual machine, as in the example described in the first embodiment. In this case, several operations are assumed as the operation of the control unit 110.

  As one operation, the control unit 110 performs control related to the state of the fault tolerant function based on the monitoring result in the service monitoring unit 240. In the control related to the state of the fault tolerant function, the control unit 110 does not use the result of analyzing the communication by the communication analysis unit 131. Then, when the fault tolerant function is invalid, the communication analysis unit 131 discards a packet related to provision of a service to the outside. By performing such control, the control unit 110 can provide a service to the outside by the virtual machine only when the fault tolerant function is effective.

  The control unit 110 may control the state of the fault tolerant function based on a notification from either the communication analysis unit 131 or the service monitoring unit 240. For example, when the notification from either the communication analysis unit 131 or the service monitoring unit 240 relates to a condition for enabling the fault tolerant function, the control unit 110 determines that the condition for enabling the fault tolerant function is satisfied. May be. In this case, the control unit 110 performs control so as to enable the fault tolerant function.

  As described above, in the virtual machine management system 200 according to the second embodiment of the present invention, the control unit 110 uses the monitoring result in the service monitoring unit 240 when controlling the state of the fault tolerant function. The service monitoring unit 240 monitors the status of processes executed by the guest OS of the virtual machine. From this, the control unit 110 can control the state of the fault tolerant function based on the status of the process. Therefore, the virtual machine management system 200 in the present embodiment has the same effects as the virtual machine management system 100 in the first embodiment.

  The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. The configurations in the embodiments can be combined with each other without departing from the scope of the present invention.

10 Host 20, 40 Virtual machine 21, 41 Guest physical memory 30 Guest OS
DESCRIPTION OF SYMBOLS 50 Communication network 100, 200 Virtual machine management system 110 Control part 111 Service management table 112 Checkpoint control part 120 Memory transfer control part 121 Transfer buffer 130 Network control part 131 Communication analysis part 132 Reserve buffer 500 Information processing apparatus 501 CPU
502 ROM
503 RAM
504 Program 505 Storage device 506 Storage medium 507 Drive device 508 Communication interface 509 Communication network 510 Input / output interface 511 Bus

Claims (8)

A service management table that holds conditions relating to the state of the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine;
Control means for controlling the state of the fault tolerant function of the virtual machine based on the condition ;
E Bei and a communication analysis means for analyzing the contents of the communication in the virtual machines,
The virtual machine management system , wherein the control unit controls a state of the fault tolerant function of the virtual machine based on an analysis result of the communication content in the communication analysis unit .   2. The control unit according to claim 1, wherein the control unit controls the virtual machine to enable the fault-tolerant function when at least one of the services satisfies the condition that the fault-tolerant function should be enabled. 3. Virtual machine management system.   The control means controls the virtual machine to disable the fault tolerant function when the condition that any of the services should disable the fault tolerant function is satisfied. The virtual machine management system described in 1.   The apparatus according to claim 1, further comprising: a checkpoint control unit configured to control a checkpoint that is an interval for storing the state of the virtual machine when the fault tolerant function is enabled according to the service related to the condition. The virtual machine management system according to any one of the above.   5. The virtual machine management system according to claim 1, wherein the service management table holds a condition for enabling the fault-tolerant function related to communication of the service for each of the services. Said control means on the basis of the result of monitoring by the service monitoring unit for executing status monitoring related to the service to be executed in a virtual machine, and controls the state of the fault tolerant capabilities of the virtual machine, of the claims 1 to 5 The virtual machine management system according to any one of the above. Determining whether at least one of the services satisfies the condition based on a condition for enabling the fault tolerant function for each of at least one service that is a target of the fault tolerant function of the virtual machine;
Based on the result of the determination, control to change the state related to the fault tolerant function of the virtual machine ,
Analyzing the content of communication in the virtual machine,
A virtual machine management method for controlling a state of the fault tolerant function of the virtual machine based on an analysis result of the communication content . On the computer,
A process of determining whether at least one of the services satisfies the condition based on a condition for enabling the fault tolerant function for each of at least one service to be a target of the fault tolerant function of the virtual machine;
Based on the result of the determination, a process for controlling the virtual machine to change a state related to the fault tolerant function ;
Processing to analyze the content of communication in the virtual machine ,
The control process is a program that controls a state of the fault tolerant function of the virtual machine based on an analysis result of the content of the communication in the analysis process .

Images