JP5872433B2 - Virtual machine placement apparatus and virtual machine placement method - Google Patents

Description

  The present invention relates to a technology for arranging virtual machines having a redundant configuration on physical servers.

  In recent years, with the expansion of the system scale, it has become a challenge to reduce the procurement operation cost of the entire system, and the number of physical servers required for the entire system by arranging multiple virtual machines on one physical server The technology to reduce this is attracting attention. Here, when arranging a plurality of virtual machines on one physical server, a technique for improving the resource usage efficiency of the physical server and guaranteeing the system performance has been proposed (see Patent Document 1).

JP 2011-170787 A

  When a redundant configuration is formed in a system using virtual machines, it is necessary to ensure high availability by arranging the active virtual machine and the standby virtual machine on appropriate physical servers. In addition, when a large number of virtual machines are placed on a single physical server, the resource usage efficiency of the physical server is improved. On the other hand, resource contention between the placed virtual machines (the physical server of each virtual machine in the same physical server There is a risk that resource contention will occur and the processing capacity of each virtual machine will be reduced), and the system performance may be reduced.

  Further, since the technique described in Patent Document 1 is a method of arranging a virtual machine on a physical server without considering the types of active and standby systems of the virtual machine, the active virtual machine on the same physical server. And a standby virtual machine may exist at the same time. For this reason, the active virtual machine and the standby virtual machine of the same system may go down at the same time due to a physical server failure, or the frequency of rebuilding the standby virtual machine may increase. As a result, it becomes difficult to guarantee availability due to the redundant configuration of the system. Therefore, an object of the present invention is to improve the availability of a system while improving the resource usage efficiency of a physical server in a system composed of an active virtual machine and a standby virtual machine.

In order to solve the above-described problem, the invention according to claim 1 is a virtual machine that selects a physical server to which each of the virtual machines is arranged in a system including a pair of an active virtual machine and a standby virtual machine. For each physical server identification information, a placement apparatus, identification information of a virtual machine placed on the physical server, information indicating whether the virtual machine is an active virtual machine or a standby virtual machine, and the system Virtual machine placement information indicating the identification information of the virtual machine that is a pair of the virtual machine, physical server resource upper limit threshold information indicating the upper limit threshold of the amount of used resources of each physical server, and the standby virtual machine For each standby virtual machine, the virtual machine required resource that indicates the standby required resource amount and the required resource amount during operation. A storage unit that stores a storage amount information, and an initial placement destination selection unit that selects a physical server that is an initial placement destination of the active virtual machine and the standby virtual machine, and the initial placement destination selection unit Refers to the virtual machine placement information when selecting the physical server of the initial placement destination of each of the working virtual machines, and selects the working system from the physical servers in which the standby virtual machine is not placed. When selecting the physical server of the initial placement destination of the virtual machine and selecting the physical server of the initial placement destination of each of the standby virtual machines, the virtual machine placement information, the virtual machine required resource amount information, and the physical server Referring to the resource upper limit threshold information, calculate the total value of the standby required resource amounts of all the standby virtual machines to be placed, and increase the used resource amount. Select the number of physical server groups whose total threshold value is equal to or larger than the standby required resource amount for the calculated total value as the physical server group as the placement destination candidate of the standby virtual machine, and the virtual machine placement Information and the virtual machine necessary resource amount information, one by one from the standby virtual machines having the large standby necessary resource amount, one by one in order from among the standby virtual machines to be arranged, and the selected The physical server that satisfies all of the following conditions (1) to (3) is selected from the physical server group as the placement destination candidate as the placement destination physical server of the standby virtual machine.
(1) The number of the standby virtual machines on the physical server including the standby virtual machine of the selected placement target is equal to or less than a predetermined threshold;
(2) The upper limit threshold value of the amount of resources used by the physical server is the total value of the standby required resource amounts of the standby virtual machine on the physical server, including the selected standby virtual machine to be placed That
(3) A predetermined number N of spares selected in descending order of the amount of resources required during operation among the spare virtual machines on the physical server, including the selected spare virtual machine to be placed The total value of the required resources during operation of the host virtual machine must be less than or equal to the upper limit threshold for the amount of resources used by the physical server

  As described above, when selecting the initial placement destination of each virtual machine, the virtual machine placement device places the active virtual machine on a physical server in which the standby virtual machine is not placed (that is, for the working virtual machine). Then, the standby virtual machine is allocated to a physical server where the active virtual machine is not allocated (that is, for the standby virtual machine). As a result, it is possible to prevent both the active virtual machine and the standby virtual machine paired with the virtual machine from going down due to a physical server failure. it can. Further, the virtual machine placement apparatus places one or more standby virtual machines on one physical server. At this time, as a placement destination of the standby virtual machines, (1) according to the claim A physical server that satisfies the conditions (3) to (3) is selected. First, the virtual machine placement apparatus selects a physical server that satisfies the condition (1), thereby increasing the load caused by the operation of the standby virtual machine when a plurality of active virtual machines fail. Can be distributed among servers. Furthermore, when the virtual machine placement device selects the placement destination of the standby virtual machine, resource competition occurs between virtual machines already placed on the physical server by selecting a physical server that satisfies the condition (2). In addition, it is possible to prevent the performance of each virtual machine from deteriorating. Further, the virtual machine placement apparatus selects a physical server that satisfies the condition (3), so that a predetermined number N (N is a value desired by a system developer or the like) of standby virtual machines on the same physical server. Each can be in operation. In other words, when all of the N standby virtual machines are in an operating state, the standby virtual machine that is in an operating state (that is, the active virtual machine) is changed to another physical machine due to a lack of physical server resources. No need to move to server. As a result, the virtual machine placement apparatus can suppress the number of movements of the virtual machine and increase the availability of the system. The virtual machine placement device increases the availability of the system while placing as many spare virtual machines as possible on the physical server by selecting physical servers that satisfy the above conditions as the placement destination of the standby virtual machines. be able to. That is, the availability of the system can be increased while improving the resource usage efficiency of the physical server.

  According to a second aspect of the present invention, when the initial placement destination selection unit of the virtual machine placement device according to the first aspect does not have the physical server that satisfies all the conditions (1) to (3), a new free space is provided. A physical server is added to the placement destination candidate physical server group, and the physical server that satisfies the conditions (1) to (3) is selected again.

  By doing so, the virtual machine placement apparatus can reliably place the standby virtual machine on the physical server even when there is no physical server that satisfies the conditions (1) to (3).

  According to a third aspect of the present invention, the initial placement destination selection unit of the virtual machine placement device according to the first or second aspect is configured so that the standby virtual server is selected from the selected placement destination candidate physical server groups. When selecting the physical server to which the machine is to be placed, among the physical servers that satisfy all of the conditions (1) to (3), the active system pair that is a pair of standby virtual machines that have already been placed on the physical server. The physical server having the smallest number of matches between the physical server that is the placement destination of the virtual machine and the physical server that is the placement destination of the active virtual machine that is a pair of the selected standby virtual machine to be placed is selected.

  By doing so, the virtual machine placement apparatus can prevent the spare virtual machines that are pairs of the active virtual machines on the same physical server from being placed on the same physical server as much as possible. Thereby, the availability of the system can be further increased.

According to a fourth aspect of the present invention, the storage unit of the virtual machine placement device according to any one of the first to third aspects further includes the active virtual machine and the standby virtual machine. For each pair, failure rate information indicating a failure rate that is a probability that switching to the standby virtual machine is performed due to a failure of the active virtual machine is stored, and the initial placement destination selection unit is configured to perform the selection When selecting the physical server to which the standby virtual machine is placed, the following condition is used instead of the condition (1).
With reference to the failure rate information, the total failure rate of the spare virtual machine pair on the physical server including the spare virtual machine to be arranged is equal to or less than a predetermined threshold value.

  In this way, the virtual machine placement device causes the standby virtual machine placed on the same physical server to be in operation simultaneously due to a failure of multiple active virtual machines, increasing the load on the physical server. Can be suppressed.

  According to a fifth aspect of the present invention, the virtual machine placement apparatus according to any one of the first to fourth aspects further executes placement of the virtual machine on the physical server. The virtual machine placement execution unit refers to the virtual machine placement information and the virtual machine required resource amount information when a failure occurs in the active virtual machine, and the placement of the standby virtual machine When it is determined that the amount of used resources of the first physical server exceeds the upper limit threshold of the first physical server by operating the standby virtual machine on the first physical server to be operated, In order from the standby virtual machine having the largest amount of resources required during operation among the standby virtual machines to be used, the used resource amount in the first physical server is Until less than 1 physical server upper threshold, go to one of the second physical server is empty physical server is running.

  By doing so, the virtual machine placement apparatus can suppress resource contention in the physical server when the standby virtual machine is in an operating state. Here, the second physical server that is the migration destination of the standby virtual machine is, for example, an evacuation physical server that temporarily saves the standby virtual machine that is in the operating state.

  According to a sixth aspect of the present invention, the virtual machine placement device according to the fifth aspect further comprises a virtual machine construction unit for constructing a new standby virtual machine, and the failure of the active virtual machine Is determined to be due to a failure of the physical server, the virtual machine placement execution unit further refers to the virtual machine placement information and refers to the standby system operating on the first physical server. The virtual machine is moved to one of the second physical servers, the moved standby virtual machine is operated as a working virtual machine, and the virtual machine construction unit refers to the virtual machine arrangement information A standby virtual machine that is a pair of virtual machines to be operated as the active system is constructed on the first physical server.

  In this way, when the standby virtual machine becomes an active state and becomes the active virtual machine, the virtual machine placement apparatus can establish a pair with the active virtual machine (new active virtual machine) pair. This standby virtual machine is constructed on the first physical server (that is, the physical server on which the original standby virtual machine has been placed). The new active virtual machine is placed on the second physical server (an empty physical server other than the failed physical server). As a result, the standby virtual machine is placed on the physical server for the standby virtual machine, and the active virtual machine is placed on a different physical server. Availability can be further increased.

  According to a seventh aspect of the present invention, in the virtual machine placement device according to the sixth aspect, when it is determined that the failure of the active virtual machine is not due to the failure of the physical server, the virtual machine placement execution unit Further, referring to the virtual machine arrangement information, the standby virtual machine operating on the first physical server is moved to the physical server on which the failed active virtual machine is arranged. The moved standby virtual machine is operated as an active virtual machine, and the virtual machine construction unit further selects a standby virtual machine as a pair of virtual machines to be operated as the active system as the first virtual machine. Build on the physical server.

  In this way, when the standby virtual machine becomes an active state and becomes the active virtual machine, the virtual machine placement apparatus can establish a pair with the active virtual machine (new active virtual machine) pair. This standby virtual machine is constructed on the first physical server (that is, the physical server on which the original standby virtual machine has been placed). Then, the new active virtual machine is moved to the physical server where the failed active virtual machine is placed. As a result, the standby virtual machine is placed on the physical server for the standby virtual machine, and the active virtual machine is placed on the physical server on which the failed active virtual machine was placed. It will be. By doing so, the resource usage efficiency of the physical server can be improved.

  According to an eighth aspect of the present invention, in the virtual machine placement device according to the seventh aspect, when the standby virtual machine fails due to a failure of the physical server, the virtual machine construction unit further includes the virtual machine Referring to the placement information, constructing the spare virtual machine as a substitute for the failed standby virtual machine in any of the second physical servers, and the failure of the standby virtual machine is When it is not a failure of the physical server, the virtual machine construction unit further refers to the virtual machine arrangement information and sets the first virtual machine as a substitute for the failed standby virtual machine to the first virtual machine. Build on the physical server.

  By doing so, the virtual machine placement apparatus allows a new standby virtual machine to be added to the second physical server (free physical server) when the failure of the standby virtual machine is due to a failure of the physical server. If the failure of the standby virtual machine is not due to a failure of the physical server, a new standby system on the first physical server (the physical server on which the original standby virtual machine was placed) Build a virtual machine. In other words, since the standby virtual machine is arranged on the physical server for the standby virtual machine, the availability of the system can be further increased. In addition, when the failure of the standby virtual machine is not due to the failure of the physical server, the virtual machine placement apparatus creates a new one on the first physical server (the physical server on which the original standby virtual machine has been placed). Since the standby virtual machine is constructed, the resource usage efficiency of the physical server can be improved.

  According to the present invention, in a system composed of a working virtual machine and a standby virtual machine, the availability of the system can be ensured while improving the resource usage efficiency of the physical server.

It is a figure showing an example of composition of a virtual machine execution system of this embodiment. FIG. 2 is a diagram illustrating an outline of a selection process of a physical server that is an initial placement destination of a virtual machine by the virtual machine placement apparatus of FIG. 1. FIG. 2 is a diagram illustrating an outline of a selection process of a physical server that is an initial placement destination of a virtual machine by the virtual machine placement apparatus of FIG. 1. FIG. 2 is a functional block diagram of the virtual machine arrangement device in FIG. 1. FIG. 2 is a diagram illustrating a process of selecting a physical server that is an initial placement destination of a virtual machine by the virtual machine placement apparatus of FIG. 1. It is the figure which showed the reconstruction process of the redundant configuration of a virtual machine by the virtual machine arrangement | positioning apparatus of FIG. It is the figure which showed the reconstruction process of the redundant configuration of a virtual machine by the virtual machine arrangement | positioning apparatus of FIG.

  Hereinafter, an embodiment (this embodiment) of the present invention will be described with reference to the drawings. First, a configuration example of the virtual machine execution system 50 according to the present embodiment will be described with reference to FIG. As illustrated in FIG. 1, the virtual machine execution system 50 includes a virtual machine placement device 10, a physical server group 21 including one or more physical servers 20, and a storage device 31. The virtual machine placement apparatus 10, the physical server 20, and the storage device 31 are connected to be communicable via the network 40. In FIG. 1, only one storage device 31 is shown, but a plurality of storage devices 31 may be provided.

  The storage device 31 reads and writes virtual machines A (VM (Virtual Machine) programs) 30 a to E (VM programs) 30 e through the network 40. In the following description, when there is no need to distinguish between the virtual machine A (30a) to the virtual machine E (30e), the virtual machine A will be referred to as the virtual machine 30. In FIG. 1, five virtual machines 30 are described, but the number is not limited to five and may be four or less or six or more.

  The virtual machine placement apparatus 10 places one or more virtual machines 30 stored in the storage device 31 on the physical server 20. Then, the virtual machine placement apparatus 10 operates the virtual machine 30 placed on the physical server 20 as the active virtual machine 30 or the standby virtual machine 30. The standby virtual machine 30 is a virtual machine 30 that operates as an alternative machine when a failure or the like occurs in the active virtual machine 30. When the virtual machine placement apparatus 10 detects a failure of the active virtual machine 30 (or the physical server 20 on which the active virtual machine 30 is placed), the virtual machine placement apparatus 10 creates a standby virtual machine that becomes a pair of the active virtual machine 30. The machine 30 is operated. Here, the virtual machine placement apparatus 10 divides the standby virtual machine 30 and the active virtual machine 30 into respective types (standby systems) so as to improve the resource usage efficiency of the physical server 20 and increase the system availability. / Working system) are aggregated and arranged on a separate physical server 20. In addition, the virtual machine placement apparatus 10 operates the standby virtual machine 30 when a failure occurs in the active virtual machine 30 or moves the standby virtual machine 30 in the operating state to another physical server 20. (Rearrangement) or reconfiguration of the standby virtual machine 30. Details of the virtual machine placement apparatus 10 will be described later.

  The physical server 20 executes the virtual machine 30 placed by the virtual machine placement apparatus 10. For example, in FIG. 1, the physical server 20a executes the arranged virtual machine A (30a) and the virtual machine C (30c), and the physical server 20b includes the arranged virtual machine B (30b) and the virtual machine D ( 30d) and the virtual machine E (30e) is executed. The physical server 20m is an empty physical server in which the virtual machine 30 is not arranged.

  Next, the outline of the selection processing of the physical server 20 that is the initial placement destination of the virtual machine 30 by the virtual machine placement device 10 of FIG. 1 will be described with reference to FIG. In the initial arrangement of the active virtual machine 30 and the standby virtual machine 30, the virtual machine 30 is collectively arranged on the physical server 20 for the active virtual machine 30, and the standby virtual machine 30 is arranged. The virtual machines 30 are aggregated and arranged on the physical server 20 for the standby virtual machine 30. As a result, it is possible to avoid a situation in which both the active virtual machine 30 and the standby virtual machine 30 are down when the physical server 20 fails. That is, the availability of the system can be increased.

Furthermore, the virtual machine placement apparatus 10 places the standby virtual machine 30 so as to satisfy the conditions (a) to (d) (placement destination physical server determination criteria) when placing the standby virtual machine 30 on the physical server 20.
(A) The number of standby virtual machines 30 in the physical server 20 is set to be equal to or less than a predetermined threshold. By doing so, it is possible to distribute the load increase among the physical servers 20 when a plurality of standby virtual machines 30 are operated.
(B) The total required amount of standby resources of the virtual machine 30 (standby virtual machine 30) in the physical server 20 is set so that the resource upper limit threshold of the physical server 20 is satisfied. By doing so, it is possible to suppress the occurrence of resource contention between the standby virtual machines 30 already arranged in the physical server 20.
(C) It is possible to operate up to N virtual machines 30 on the physical server 20 (the number desired by the system developer or the like). That is, select the physical server 20 that satisfies the total resource amount required for operation of the N virtual machines 30 arranged in the physical server 20 including the standby virtual machine 30 to be arranged ≦ the resource upper limit threshold of the physical server 20 To do. In this way, when the standby virtual machine 30 of the same physical server 20 is in an operating state, the standby virtual machine 30 that is in an operating state due to a lack of resources of the physical server 20 (that is, It is not necessary to move the active virtual machine 30) to another physical server 20. Thereby, the virtual machine placement apparatus 10 can increase the availability of the system.
(D) The standby virtual machines 30 that are pairs of the active virtual machines 30 on the same physical server 20 are prevented from becoming the same physical server 20 as much as possible. Thereby, the virtual machine placement apparatus 10 can increase the availability of the system. An overview of the selection process of the physical server 20 at the initial placement destination that satisfies the condition (d) will be described with reference to FIG. 3 (see FIG. 1 as appropriate).

  Here, a case is considered where the virtual machine placement apparatus 10 selects one of the physical servers 20a and 20b, which is the physical server 20 for the standby virtual machine 30, as the initial placement destination of the standby virtual machine 30a. The standby virtual machines 30b and 30c are arranged on the physical server 20a, and the standby virtual machines 30d and 30e are arranged on the physical server 20b. The physical servers 20c, 20d, and 20e are physical servers 20 for the active virtual machine 30, and the active virtual machines 30f, 30g, and 30h are arranged on the physical server 20c, and the active server is set on the physical server 20d. Virtual machine 30i is arranged, and the active virtual machine 30j is arranged on the physical server 20e. A pair relationship between the active virtual machine 30 and the standby virtual machine 30 is as shown by an arrow in FIG.

  In such a case, the active virtual machine 30f that is a pair of the standby virtual machine 30a is arranged in the physical server 20c, and becomes a pair of the active virtual machines 30g and 30h arranged in the physical server 20c. The standby virtual machines 30b and 30c are both arranged in the physical server 20a. Here, if the virtual machine placement apparatus 10 selects the physical server 20a as the placement destination of the standby virtual machine 30a, the standby virtual machine 30a, which is a pair of the active virtual machines 30f, 30g, and 30h. 30b and 30c are all arranged in the same physical server 20a, and the availability of the system is lowered. On the other hand, if the virtual machine placement apparatus 10 selects the physical server 20b as the placement destination, both the active virtual machines 30i and 30j that form a pair of the standby virtual machines 30d and 30e are the active virtual machines 30f. Are arranged in different physical servers 20 d and 20 e and distributed to a plurality of physical servers 20. Therefore, if the virtual machine placement apparatus 10 selects the physical server 20b as the placement destination of the standby virtual machine 30a, the standby virtual machines 30 that form a pair with the active virtual machine 30 on the same physical server 20 Therefore, the availability of the system can be increased.

  As described above, the virtual machine placement apparatus 10 selects the physical server 20 that is the initial placement destination of the virtual machine 30, so that as many spare virtual machines 30 as possible are placed on the physical server 20, the system Can increase availability.

  In addition, after the initial placement of the standby virtual machine 30 and the active virtual machine 30 on each physical server 20, if any of the virtual machines 30 fails, the virtual machine placement device 10 Then, the virtual machine 30 is moved so that the standby virtual machine 30 (the active virtual machine 30) and the standby virtual machine 30 that are in an operating state are not arranged on the same physical server 20 as much as possible. Details of the processing procedure at this time will be described later.

  Next, a configuration example of the virtual machine placement apparatus 10 will be described with reference to FIG. 4 (see FIG. 1 as appropriate). The virtual machine placement apparatus 10 includes a processing unit 100, a storage unit 110, and a communication unit 120. The communication unit 120 is a communication interface, and transmits / receives information to / from the physical server 20 and the storage device 31 via the network 40 in FIG.

  The processing unit 100 includes an initial placement destination selection unit 101, a virtual machine placement execution unit 102, a virtual machine construction unit 103, and a virtual machine execution state switching unit 104. The processing unit 100 includes a CPU (Central Processing Unit) and a main memory (not shown), and each unit of the processing unit 100 is realized by developing an application program stored in the storage unit 110 in the main memory.

  The storage unit 110 is a storage medium such as a memory or a hard disk. The storage unit 110 stores virtual machine arrangement information 112 (see Table 1), physical server resource upper limit threshold information 111 (see Table 2), and virtual machine required resource amount information 113 (see Table 3). The failure rate information 114 indicated by a broken line may be deployed or not deployed, and the deployment will be described later.

  For each piece of identification information of the physical server 20, the virtual machine arrangement information 112 is information indicating the identification information of the virtual machine 30 arranged on the physical server 20 and the type of active / standby system of the arranged virtual machine 30. And the identification information of the virtual machine 30 that is a pair of the arranged virtual machines 30 (see Table 1). For example, in the virtual machine arrangement information 112 shown in Table 1, “virtual machine A” and “virtual machine C”, which are the standby virtual machines 30, are arranged on the physical server 20 with the identification information “Sa”. The virtual machines 30 that form a pair of the virtual machines 30 of the system are “virtual machine B” and “virtual machine F”. Further, it indicates that the virtual machine 30 is not arranged on the physical server 20 with the identification information “Sm”, that is, it is a free physical server. Note that the type of the standby system / working system of each virtual machine 30 and the identification information of the paired virtual machine 30 include the switching of the backup system / working system of each virtual machine 30 and the change of the paired virtual machine 30. It is rewritten when.

  The physical server resource upper limit threshold information 111 is information indicating the upper limit threshold of the used resource amount of each physical server 20 (see Table 2).

  The virtual machine required resource amount information 113 includes, for each standby virtual machine 30, the standby required resource amount required for the standby virtual machine 30 and when the standby virtual machine 30 operates (that is, This is information indicating the amount of resources required during operation (when operating as the active virtual machine 30) (see Table 3).

  The initial placement destination selection unit 101 of the processing unit 100 refers to virtual machine placement information 112 (see Table 1), physical server resource upper limit threshold information 111 (see Table 2), and virtual machine required resource amount information 113 (see Table 3). Then, the physical server 20 that is the initial placement destination of the active virtual machine 30 and the standby virtual machine 30 is selected. When selecting the initial placement destination physical server 20 of each of the active virtual machines 30, the initial placement destination selection unit 101 refers to the virtual machine placement information 112 (see Table 1) and refers to the standby virtual machine. The physical server 20 that is the initial placement destination is selected from among the physical server 20 group in which 30 is not placed (that is, from the physical server group 20 for the active virtual machine 30). In addition, when selecting the initial placement destination physical server 20 of each of the standby virtual machines 30, the initial placement destination selection unit 101 selects the physical server 20 group in which the active virtual machine 30 is not placed. The physical server 20 of the initial placement destination is selected (that is, from the physical server group 20 for the standby virtual machine 30). Here, the selection of the physical server 20 as the initial placement destination of the standby virtual machine 30 is performed as follows.

  That is, the initial placement destination selection unit 101 refers to the virtual machine placement information 112 (see Table 1) and the virtual machine required resource amount information 113 (see Table 3), and sets up all the standby virtual machines 30 to be placed. Calculate the total amount of resources required during standby. Then, the initial placement destination selection unit 101 refers to the virtual machine placement information 112 (see Table 1) and the physical server resource upper limit threshold information 111 (see Table 2), and selects the physical server 20 from among the free physical servers. The physical server 20 group in which the total value of the upper limit thresholds of the resource amount is equal to or greater than the total value of the required amount of standby resources is selected. The selected physical server group 21 becomes the physical server 20 group that is the placement destination candidate of the standby virtual machine 30. Next, the initial placement destination selection unit 101 refers to the virtual machine required resource amount information 113 (see Table 3) and the virtual machine placement information 112 (see Table 1), among the spare virtual machines 30 to be placed. Then, the physical server 20 at the placement destination is selected in order from the standby virtual machine 30 having a large amount of resources required for standby. The placement destination physical server 20 is selected by selecting a physical server 20 that satisfies the placement destination physical server determination criteria from the group of placement destination candidate physical servers 20. Details of the initial placement destination selection unit 101 will be described later.

  The virtual machine placement execution unit 102 executes placement of the virtual machine 30 on the physical server 20 via the communication unit 120. The virtual machine placement execution unit 102 places the virtual machine 30 on the physical server 20 that is the initial placement destination selected by the initial placement destination selection unit 101, or the virtual machine 30 or the physical server 20 fails due to a failure of the virtual machine 30 or the physical server 20 thereafter. The virtual machine 30 is rearranged when the movement (rearrangement) of 30 becomes necessary. Details of the virtual machine arrangement execution unit 102 will be described later.

  The virtual machine construction unit 103 reconstructs the active virtual machine 30 or the standby virtual machine 30 on the physical server 20 via the communication unit 120 when the virtual machine 30 fails. For example, the virtual machine construction unit 103 is configured to operate when the active virtual machine 30 fails and the standby virtual machine 30 that is a pair of the active virtual machine 30 starts operating. A standby virtual machine 30 that is a pair of the virtual machine 30 (original standby virtual machine 30) is newly constructed on the physical server 20.

  After the failure recovery of the virtual machine 30 or the physical server 20, the virtual machine execution state switching unit 104 switches the operating virtual machine 30 (original standby virtual machine 30) to a standby state via the communication unit 120.

<Initial location selection>
Next, the processing procedure of the virtual machine placement apparatus 10 in FIG. 4 will be described with reference to FIG. First, the selection procedure of the physical server 20 that is the initial placement destination of the virtual machine 30 by the virtual machine placement device 10 will be described (see FIG. 1 as appropriate). Here, it is assumed that the initial placement destination selection unit 101 of the virtual machine placement apparatus 10 has already selected the initial placement destination physical server 20 of the active virtual machine 30.

  First, the virtual machine placement apparatus 10 accepts input of resource usage (standby required resource amount and operating required resource amount) required for standby virtual machine 30 during standby and during operation (S1). The input resource usage is written as the standby required resource amount and the operating required resource amount of the standby virtual machine 30 in the virtual machine required resource amount information 113 (see Table 3).

  Next, the initial placement destination selection unit 101 provisionally determines the number of physical servers necessary for placement of the standby virtual machine 30 (S2). That is, first, the initial placement destination selection unit 101 refers to the virtual machine required resource amount information 113 (see Table 3) and calculates a total value of standby required resource amounts of all the standby virtual machines to be placed. To do. Then, referring to the physical server resource upper limit threshold information 111 (see Table 2), the number of physical servers 20 for securing a resource amount (upper limit threshold) equal to or greater than the standby required resource amount for the calculated total value is obtained. Make a provisional decision. For example, the initial placement destination selection unit 101 provisionally determines the number of physical servers necessary for placement of the standby virtual machine 30 by the following equation (1). Then, the initial placement destination selection unit 101 refers to the virtual machine placement information 112 (see Table 1) and selects the tentatively determined number of physical servers 20 from the free physical servers. This physical server group 20 is set as a physical server group 20 as a placement destination candidate.

Number of physical servers = the number of physical servers 20 having a resource amount (upper threshold) corresponding to the total of the standby required resource amounts of all the standby virtual machines 30 × a (1)
However, a is a positive number.

  Thereafter, the initial placement destination selection unit 101 of the virtual machine placement apparatus 10 refers to the virtual machine required resource amount information 113 (see Table 3) if there is a standby virtual machine 30 to be placed (Yes in S3). Then, one of the standby virtual machines 30 to be placed is selected from the standby virtual machines 30 in descending order of the amount of resources required during standby (the amount of resources required during standby) (S4). The initial placement destination selection unit 101 refers to the virtual machine placement information 112 (see Table 1), physical server resource upper limit threshold information 111 (see Table 2), and virtual machine required resource amount information 113 (see Table 3). Then, it is determined whether there is a physical server 20 that satisfies the placement destination physical server determination criteria in the placement destination candidate physical server 20 group (S5). If there is a physical server 20 that satisfies the placement destination physical server determination criteria (Yes in S5), the virtual machine placement execution unit 102 selects the standby virtual machine 30 to be placed selected in S4 as the physical server 20. (S6) and return to S3. On the other hand, when there is no physical server 20 that satisfies the placement destination physical server determination criteria (No in S5), the placement destination candidate physical server 20 (free physical server) is referred to by referring to the virtual machine placement information 112 (see Table 1). Is added (S7), and the process returns to S5. When the placement of all the standby virtual machines 30 scheduled to be placed in S3 is completed (No in S3), the process is terminated.

The placement destination physical server determination criterion used in S5 is composed of the following conditions (1) to (4).
(1) The number of backup virtual machines 30 on the physical server 20 including the backup virtual machines 30 to be arranged is equal to or less than a predetermined threshold (conditions corresponding to the above (a)).
(2) The total value of the standby necessary resources of the standby virtual machine 30 on the physical server 20 including the standby virtual machine 30 to be arranged is the resource amount (used resource amount) of the physical server 20 (The condition corresponding to (b) described above).
(3) A predetermined number N of standby virtual machines selected in descending order of the amount of resources required during operation among the standby virtual machines 30 on the physical server 20 including the standby virtual machines 30 to be arranged. The total value of the required resource amount during operation of the machine 30 is equal to or less than the upper limit threshold value of the used resource amount of the physical server 20 (condition corresponding to the above (c)).
(4) Of the physical servers 20 that satisfy the conditions (1) to (3) described above, the placement destination of the active virtual machine 30 that forms a pair with the standby virtual machine 30 that has already been placed on the physical server 20 The physical server 20 is the physical server 20 having the smallest number of matches between the physical server 20 and the physical server 20 that is the placement destination of the active virtual machine 30 that is a pair of the standby virtual machine 30 to be placed (described above (d)). Condition).
Note that the conditions (1) to (3) may be used as the placement destination physical server determination criteria, and the condition (4) may not be used.

  When the initial placement destination selection unit 101 selects the physical server 20 that is the placement destination of the standby virtual machine 30 by the above procedure, the virtual machine placement execution unit 102 moves to the physical server 20 that has selected each standby virtual machine 30. Deploy.

  The initial placement destination selection unit 101 selects the physical server 20 that satisfies the above conditions as the placement destination of the standby virtual machine 30, thereby improving the resource usage efficiency of the physical server 20 while increasing the system availability. Can do.

<Reconstruction>
Next, a procedure for rebuilding a redundant configuration of a virtual machine when a system failure occurs after the virtual machine placement apparatus 10 performs the initial placement of the virtual machine 30 will be described with reference to FIGS. 6 and 7. This will be described (see FIG. 4 as appropriate).

  When it is detected that a system failure has occurred in the virtual machine placement device 10 of FIG. 4 (S11), the virtual machine placement device 10 determines whether this failure is a failure of only the active virtual machine 30 (S12). ). When the virtual machine placement apparatus 10 determines that the failure is only in the active virtual machine 30 (Yes in S12), the process proceeds to S13. On the other hand, when it is determined that there is not only a failure in the active virtual machine 30, that is, when it is determined that there is also a failure in the standby virtual machine 30 (No in S12), the process proceeds to S16 in FIG. The processing after S16 will be described later.

  In S13, the virtual machine placement execution unit 102 of the virtual machine placement device 10 refers to the virtual machine placement information 112 (see Table 1) and refers to the physical server 20 of the standby virtual machine 30 (for the standby virtual machine 30). The physical server 20) determines whether or not resource contention has occurred. When it is determined that resource contention has occurred (Yes in S13), the process proceeds to S14.

  Note that whether or not resource contention has occurred in S13 is determined by whether or not the amount of resources used in the physical server 20 exceeds the upper limit threshold of the physical server 20. That is, when the amount of resources used in the physical server 20 exceeds the upper limit threshold of the physical server 20 (see the physical server resource upper limit threshold information 111), the virtual machine arrangement execution unit 102 determines that resource contention occurs. On the other hand, when the used resource amount in the physical server 20 is equal to or less than the upper limit threshold of the physical server 20, the virtual machine arrangement execution unit 102 determines that no resource conflict occurs. The used resource amount of the physical server 20 is calculated from the standby required resource amount and the operating required resource amount (see virtual machine required resource amount information 113) and the virtual machine arrangement information 112 of each virtual machine 30.

  In S <b> 14, the virtual machine placement execution unit 102 refers to the virtual machine placement information 112 (see Table 1) and the virtual machine required resource amount information (see Table 3), and the virtual machine placement execution unit 102 In order of increasing required resource amount at the time of operation, other free space is used until resource contention does not occur in the standby virtual machine 30, that is, until the used resource amount in the physical server 20 becomes equal to or less than the upper limit threshold of the physical server 20. Temporarily save to physical server. Then, it progresses to S21.

  On the other hand, in S13, when the virtual machine arrangement execution unit 102 determines that no resource contention has occurred in the physical server 20 (No in S13), the virtual machine arrangement execution unit 102 determines that the active standby virtual The machine 30 is continuously operated on the current physical server 20 (S15). Then, it progresses to S21. S21 will be described later.

  When it is determined in S21 that the system failure is due to a failure of the physical server 20 in which the active virtual machine 30 is placed (Yes in S21), the virtual machine placement execution unit 102 sends the virtual machine placement information 112 ( Referring to Table 1), all of the active standby virtual machines 30 are moved to the free physical server and are operated as the new active system (active virtual machine 30). Here, the standby virtual machine 30 in the operating state is for the standby virtual machine 30 in the operating state temporarily saved in an empty physical server in S14 or for the standby virtual machine 30 in S15. This is a standby virtual machine 30 temporarily operating on the physical server 20. That is, both the standby virtual machine 30 temporarily operating on the physical server 20 for evacuation and the standby virtual machine 30 operating on the physical server 20 for the standby virtual machine 30 are free physical. Move to the server (physical server 20 for the active virtual machine 30) and operate as the active virtual machine 30. In addition, the virtual machine construction unit 103 uses the new standby system (that is, the standby virtual machine 30 that becomes a pair of the virtual machine 30 that newly becomes the active virtual machine 30) of the original standby virtual machine 30. A new construction is performed on the placement physical server (that is, the physical server 20 on which the original standby virtual machine 30 was placed) (S22). Thereby, the redundant configuration of the virtual machine 30 is restored. Note that when the standby virtual machine 30 is moved or when the standby virtual machine 30 becomes active and switches to the active virtual machine 30, the virtual machine arrangement information 112 described above (see Table 1). This information is rewritten accordingly.

  On the other hand, when it is determined in S21 that the system failure is not due to a failure of the physical server 20 in which the active virtual machine 30 is placed (No in S21), for example, an application failure in the active virtual machine 30 When the virtual machine placement execution unit 102 determines that the failure is caused by the virtual machine placement execution unit 102, the virtual machine placement execution unit 102 moves the active standby virtual machine 30 to the placement physical server of the failed active virtual machine 30, and the new working system (working It is operated as a virtual machine 30) of the system. In addition, the virtual machine construction unit 103 newly constructs a new standby virtual machine 30 by using the placement physical server of the original standby virtual machine 30. Thereby, the redundant configuration of the virtual machine 30 is restored.

  As described above, when the standby virtual machine 30 comes to operate as the active virtual machine 30 due to a system failure, the virtual machine placement apparatus 10 becomes a virtual machine 30 (new active virtual machine 30). ) On the physical server 20 for the active virtual machine 30. Further, the standby virtual machine 30 that is a pair of the new standby virtual machine 30 is reconstructed on the physical server 20 for the standby virtual machine 30. As a result, even after a system failure occurs, the active virtual machine 30 is centrally arranged on the physical server 20 for the active virtual machine 30, and the standby virtual machine 30 is Are collectively arranged on the physical server 20 for the virtual machine 30.

  Next, with reference to FIG. 7, processing when the virtual machine placement apparatus 10 determines that the failure of the system is not the failure of only the active virtual machine 30 (No in S12 of FIG. 6) will be described (as appropriate). FIG. 1). When the virtual machine placement apparatus 10 determines that the system failure is only the failure of the standby virtual machine 30 (No in S12 in FIG. 6 → Yes in S16 in FIG. 7), the system failure is the standby system. When it is determined that the failure is caused by a failure of the physical server 20 of the virtual machine 30 (Yes in S24), the virtual machine construction unit 103 refers to the virtual machine arrangement information 112 (see Table 1) and refers to the failed standby system. A spare virtual machine that replaces the virtual machine is newly constructed on any free physical server (S25). On the other hand, when determining that the system failure is not due to the failure of the physical server 20 of the standby virtual machine 30 (No in S24), the virtual machine construction unit 103 uses the virtual machine arrangement information 112 (see Table 1). Referring to this, a spare virtual machine 30 that replaces the failed spare virtual machine 30 is newly constructed on the physical server where the original spare virtual machine 30 is placed (S26). When the virtual machine placement apparatus 10 determines that the system failure is caused by the failure of both the active virtual machine 30 and the standby virtual machine 30 (No in S16), the virtual machine construction unit 103 The active and standby virtual machines 30 are newly constructed on the server 20 (S27). Thereby, the redundant configuration of the virtual machine 30 is restored.

  As described above, even after a system failure occurs, the virtual machine placement apparatus 10 places the active virtual machine 30 on the physical server 20 for the active virtual machine 30 and sets it as a standby virtual machine. Since 30 is arranged on the physical server 20 for the standby virtual machine 30, the availability of the system can be increased.

  The storage unit 110 of the virtual machine placement apparatus 10 in FIG. 4 has a failure of the active virtual machine 30 for each system (a pair of virtual machines 30) including the active virtual machine 30 and the standby virtual machine 30. Thus, failure rate information 114 indicating a failure rate that is a probability of switching to the standby virtual machine 30 may be further stored. When the initial placement destination selection unit 101 selects the physical server 20 that is the placement destination of the standby virtual machine 30, the initial placement destination selection unit 101 uses the failure rate information 114 instead of the condition (1) of the placement destination physical server determination criterion. The physical server 20 at the placement destination is selected with reference to the failure rate of each system shown. That is, the initial placement destination selection unit 101 includes the failure rate of the system of the standby virtual machine 30 on the physical server 20 in the placement destination candidate physical server 20 group including the standby virtual machine 30 to be placed. A physical server 20 whose total value is less than or equal to a predetermined threshold value may be selected as the physical server 20 at the placement destination.

  By doing so, the virtual machine placement apparatus 10 selects the physical server 20 that is the placement destination of the standby virtual machine 30 based on the failure rate of the system, and thus a plurality of active virtual machines. It is possible to prevent the standby virtual machine 30 arranged on the same physical server 20 from operating simultaneously due to the failure of 30 and increasing the load on the physical server 20.

DESCRIPTION OF SYMBOLS 10 Virtual machine arrangement | positioning apparatus 20 Physical server 21 Physical server group 30 Virtual machine 31 Storage apparatus 40 Network 50 Virtual machine execution system 100 Processing part 101 Initial arrangement destination selection part 102 Virtual machine arrangement | positioning execution part 103 Virtual machine construction part 104 Virtual machine execution state Switching unit 110 Storage unit 111 Physical server resource upper limit threshold information 112 Virtual machine arrangement information 113 Virtual machine required resource amount information 114 Failure rate information 120 Communication unit

Claims (9)

A virtual machine placement device that selects a physical server to which each virtual machine is placed in a system comprising a pair of a working virtual machine and a standby virtual machine,
For each identification information of the physical server, identification information of a virtual machine arranged in the physical server, information indicating whether the virtual machine is an active virtual machine or a standby virtual machine, and a pair of the virtual machines in the system Virtual machine placement information indicating the identification information of the virtual machine to be used, physical server resource upper limit threshold information indicating the upper limit threshold of the amount of resource used by each physical server, and for each standby virtual machine, the spare system A storage unit that stores virtual machine required resource amount information indicating the required resource amount during standby and the required resource amount during operation of the virtual machine;
An initial placement destination selection unit that selects a physical server that is an initial placement destination of the active virtual machine and the standby virtual machine;
The initial placement destination selection unit
When selecting an initial placement destination physical server for each of the active virtual machines, the virtual machine placement information is referred to, and the active virtual machine is selected from among the physical servers on which the standby virtual machine is not placed. Select the physical server where the machine is initially placed,
When selecting the initial placement destination physical server for each of the standby virtual machines, refer to the virtual machine placement information, the virtual machine required resource amount information, and the physical server resource upper limit threshold information, The total value of the standby required resource amount of the standby virtual machine is calculated, and the total number of physical resources of which the total threshold value of the used resource amount is equal to or greater than the standby required resource amount for the calculated total value Select a server group as a physical server group as a placement destination candidate of the standby virtual machine,
With reference to the virtual machine arrangement information and the virtual machine required resource amount information, one of the standby virtual machines to be arranged is selected one by one in order from the standby virtual machine with the larger standby required resource amount. Selecting the physical server that satisfies all of the following conditions (1) to (3) from the selected physical server group of the placement destination candidates as a physical server that is the placement destination of the standby virtual machine: A virtual machine placement device.
(1) The number of the standby virtual machines on the physical server including the standby virtual machine of the selected placement target is equal to or less than a predetermined threshold;
(2) The upper limit threshold value of the amount of resources used by the physical server is the total value of the standby required resource amounts of the standby virtual machine on the physical server, including the selected standby virtual machine to be placed That
(3) A predetermined number N of spares selected in descending order of the amount of resources required during operation among the spare virtual machines on the physical server, including the selected spare virtual machine to be placed The total value of the required resources during operation of the host virtual machine must be less than or equal to the upper limit threshold for the amount of resources used by the physical server The initial placement destination selection unit
If there is no physical server that satisfies all the conditions (1) to (3), a new free physical server is added to the placement destination candidate physical server group, and the above (1) to (3) are performed again. The virtual machine placement apparatus according to claim 1, wherein the physical server that satisfies a condition is selected. The initial placement destination selection unit
Among the physical servers that satisfy all of the conditions (1) to (3) when selecting a physical server that is the placement destination of the standby virtual machine from the selected placement destination candidate physical server group, The physical server that is the placement destination of the active virtual machine that is a pair of the standby virtual machine that has already been placed on the physical server and the active virtual machine that is the pair of the selected standby virtual machine that is the placement target The virtual machine placement apparatus according to claim 1 or 2, wherein the physical server having the smallest number of matches with the placement destination physical server is selected. The storage unit further includes:
Failure rate information indicating a failure rate, which is a probability of switching to the standby virtual machine due to a failure of the active virtual machine for each pair of the active virtual machine and the standby virtual machine Remember
The initial placement destination selection unit
4. The method according to claim 1, wherein when the physical server of the selected standby virtual machine is selected, the following condition is used instead of the condition (1). The virtual machine arrangement device according to item.
With reference to the failure rate information, the total failure rate of the spare virtual machine pair on the physical server including the spare virtual machine to be arranged is equal to or less than a predetermined threshold value. The virtual machine placement device further includes:
A virtual machine placement execution unit that executes placement of the virtual machine on the physical server;
When a failure occurs in the active virtual machine,
The virtual machine placement execution unit
With reference to the virtual machine arrangement information and virtual machine required resource amount information, the first physical server is operated by the standby virtual machine on the first physical server on which the standby virtual machine is arranged. When it is determined that the used resource amount of the server exceeds the upper limit threshold value of the first physical server, the standby virtual machine having the large required resource amount during operation is sequentially selected from the standby virtual machines that are in the operating state. The system moves to one of the second physical servers that are free physical servers and operates until the amount of used resources in the first physical server becomes equal to or less than the upper limit threshold of the first physical server. The virtual machine placement device according to any one of claims 1 to 4. The virtual machine placement device further includes:
A virtual machine construction unit for constructing a new virtual machine for the standby system,
When it is determined that the failure of the active virtual machine is due to the failure of the physical server,
The virtual machine placement execution unit further includes:
Referring to the virtual machine arrangement information, the standby virtual machine running on the first physical server is moved to one of the second physical servers, and the moved standby virtual machine is moved Run as a working virtual machine,
The virtual machine construction unit
The virtual machine of the spare system used as the pair of the virtual machine operated as the active system with reference to the virtual machine arrangement information is constructed on the first physical server. Virtual machine placement device. When it is determined that the failure of the active virtual machine is not due to the failure of the physical server,
The virtual machine placement execution unit further includes:
Referring to the virtual machine placement information, the standby virtual machine operating on the first physical server is moved to the physical server on which the failed active virtual machine is placed, and the moved Run the standby virtual machine as the active virtual machine,
The virtual machine construction unit further includes:
The virtual machine placement apparatus according to claim 6, wherein a standby virtual machine that is a pair of virtual machines to be operated as the active system is constructed on the first physical server. When the standby virtual machine fails due to a failure of the physical server,
The virtual machine construction unit further includes:
With reference to the virtual machine arrangement information, construct the standby virtual machine as a substitute for the failed standby virtual machine in any of the second physical servers,
When the failure of the standby virtual machine is not the failure of the physical server,
The virtual machine construction unit further includes:
8. The virtual machine placement information is configured on the first physical server by referring to the virtual machine arrangement information, and constructing the spare virtual machine as a substitute for the failed standby virtual machine on the first physical server. Virtual machine placement device. A virtual machine placement method for selecting a physical server of a placement destination of each of the virtual machines in a system consisting of a pair of a working virtual machine and a standby virtual machine,
For each identification information of the physical server, identification information of a virtual machine arranged in the physical server, information indicating whether the virtual machine is an active virtual machine or a standby virtual machine, and a pair of the virtual machines in the system Virtual machine placement information indicating the identification information of the virtual machine to be used, physical server resource upper limit threshold information indicating the upper limit threshold of the amount of resource used by each physical server, and for each standby virtual machine, the spare system A virtual machine placement apparatus comprising a storage unit for storing virtual machine required resource amount information indicating the required resource amount during standby and the required resource amount during operation of the virtual machine,
When selecting an initial placement destination physical server for each of the active virtual machines, the virtual machine placement information is referred to, and the active virtual machine is selected from among the physical servers on which the standby virtual machine is not placed. Selecting a physical server where the machine is initially placed;
When selecting the initial placement destination physical server for each of the standby virtual machines, refer to the virtual machine placement information, the virtual machine required resource amount information, and the physical server resource upper limit threshold information, Calculating a total value of the standby required resource amount of the standby virtual machine;
Select the number of physical server groups in which the total value of the upper limit thresholds of the used resource amount is equal to or greater than the standby required resource amount for the calculated total value as the physical server group as the placement destination candidate of the standby virtual machine And steps to
With reference to the virtual machine arrangement information and the virtual machine required resource amount information, one of the standby virtual machines to be arranged is selected one by one in order from the standby virtual machine having the largest required amount of standby resources. Steps,
Selecting the physical server that satisfies all of the following conditions (1) to (3) as a physical server to which the standby virtual machine is to be placed from the selected physical server group that is the placement destination candidate: A virtual machine arrangement method characterized by executing.
(1) The number of the standby virtual machines on the physical server including the standby virtual machine of the selected placement target is equal to or less than a predetermined threshold;
(2) The upper limit threshold value of the amount of resources used by the physical server is the total value of the standby required resource amounts of the standby virtual machine on the physical server, including the selected standby virtual machine to be placed That
(3) A predetermined number N of spares selected in descending order of the amount of resources required during operation among the spare virtual machines on the physical server, including the selected spare virtual machine to be placed The total value of the required resources during operation of the host virtual machine must be less than or equal to the upper limit threshold for the amount of resources used by the physical server

Images