JP2016038888A - Information processing system, information processing method and information processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve efficiency of migration processing.SOLUTION: A storage device stores allocation information showing computer resources allocated to a virtual machine, and use information showing the computer resources used by the virtual machine. An information processing apparatus acquires the use information of computer resources used by a virtual machine operating in the apparatus from the storage device. Next, the information processing apparatus selects an information processing apparatus of a takeover destination in the case that an index of a use amount of the computer resources exceeds a threshold and that computer resources to be additionally allocated to the virtual machine are insufficient. In this case, an information processing apparatus which has more unused computer resources than the computer resources allocated to the virtual machine is selected on the basis of the allocation information in the storage device. Then, a first information processing apparatus writes allocation information for allocating more computer resources than the computer resources allocated to the virtual machine into the storage device, and makes the selected information processing apparatus take over processing of the virtual machine.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an information processing system, an information processing method, and an information processing program.

  In recent years, with the evolution of virtualization technology, the number of data centers and companies that operate by virtualizing servers is increasing. In a data center that operates a plurality of virtualized servers, it is desired to manage a plurality of virtual servers collectively.

  FIG. 1 is a diagram for explaining an example of a virtualization system using a management server. The virtualization system 100 in FIG. 1 includes a server 110a, a server 110b, and a management server 120. The server 110a, the server 110b, and the management server 120 are connected via a communication network. The server 110a includes a virtual machine (VM) 111a to VM 111c and a hypervisor 112a. The server 110b includes a VM 111d, a VM 111e, and a hypervisor 112b. The VM 111 is a virtual machine that is operated by logically dividing one server 110 on which various devices including a central processing unit (CPU), a memory, and an IO device are mounted as if it were a plurality of servers. is there. The VMs 111a to 111c are virtual machines that operate in the server 110a. The VMs 111d to 111e are virtual machines that operate in the server 110b. Each of the VMs 111 can operate an operating system (OS) and an application. The hypervisor 112 manages each VM 111. For example, the hypervisor 112a manages the VMs 111a to 111c. The hypervisor 112b manages the VMs 111d to 111e.

  The management server 120 is a server that manages the server 110a and the server 110b. The management server 120 periodically inquires each server 110 to acquire and store usage information of various devices from the hypervisor 112 or each VM 111. In addition, the management server 120 manages each VM 111 by holding management information of the VMs 111 of the servers 110a and 110b, allocation information of various devices, usage information of each VM, and the like. The allocation information of various devices includes allocation information of allocation amounts of, for example, a CPU, a memory, and an IO device that are devices that can be used by the VM 111.

  The management server 120 uses the usage information held by the device allocation amount of the device allocation amount allocated to the operating VM 111 and the device allocation amount allocated to the VM 111 to be used. Monitor certain usage. For example, the VM 111a that has a large amount of processing is increasing the amount of device usage with the amount of processing. In order to improve the processing performance of the VM 111a, it is preferable that the management server 120 allocates an additional device in the server 110a when the usage amount of the device of the VM 111a is larger than a predetermined threshold. When an additional device is allocated, if the additional device cannot be allocated to the VM 111 in the server 110a, the management server 120 performs a migration process that takes over the process of the VM 111a to the other server 110b. Migration is a process of migrating the allocation information of the VM 111 and the process to be executed to another server 110b.

  The management server 120 monitors the usage amount of devices used by the VM 111 and controls the devices used by the VM 111. In this way, the data center is managed without using manpower.

  A technique is known in which when a failure is detected in an information processing apparatus that operates a virtual machine, processing is taken over by another virtual machine (see, for example, Patent Document 1).

  As a server management method, a technique for managing the processing amount of each processing apparatus is known (see, for example, Patent Document 2).

  A technique is known in which a management server monitors a load (CPU usage, memory usage) and performs migration (see, for example, Patent Document 3).

JP 2009-80704 A Japanese Patent Application Laid-Open No. 11-85592 JP 2008-217302 A

  In the virtualization system 100 of FIG. 1, the management server 120 periodically acquires usage information and determines the usage status of various devices. Since this acquisition process is performed periodically, it is difficult for the management server 120 to immediately detect that the usage amount of the device of the VM 111 has exceeded a predetermined threshold. Similarly, even if the management server 120 runs out of devices additionally allocated to the VM 111 in the server 110a, it is difficult to immediately detect the shortage of devices. In other words, although the usage amount of the device of the VM 111 exceeds a predetermined threshold and it is better to execute the migration, there is a problem that it takes time until the migration process is actually started.

  In one aspect, the present invention aims to improve the efficiency of migration processing.

  The information processing system includes a plurality of information processing apparatuses and storage devices that can execute one or more virtual machines. The storage device stores allocation information indicating computer resources allocated to one or more virtual machines and usage information indicating computer resources used by one or more virtual machines. The first information processing apparatus among the plurality of information processing apparatuses includes an acquisition unit, a selection unit, and a processing unit. The acquisition unit acquires the use information of the computer resources used by the virtual machine that is operating in the first information processing apparatus from the storage device. When the index of the usage amount of the computer resource based on the usage information exceeds a predetermined threshold and the computer resource to be additionally allocated to the virtual machine operating in the first information processing apparatus is insufficient Process. The selection unit is a second information processing apparatus having a larger number of unused computer resources than the computer resources allocated to the virtual machine among the plurality of information processing apparatuses based on the allocation information stored in the storage device. Is selected as the information processing apparatus that takes over the processing of the virtual machine. The processing unit writes, in the storage device, allocation information for allocating more computer resources to the operating virtual machine than the computer resources allocated to the operating virtual machine. The processing unit causes the second information processing apparatus to take over the processing of the operating virtual machine.

  The migration process can be made efficient.

It is a figure explaining the example of the virtualization system using a management server. It is a figure explaining the example of the problem of the virtualization system using a management server. It is a figure explaining the example of the virtualization system which concerns on this embodiment. The example of the hardware constitutions of a server is shown. It is a figure (the 1) explaining the example of management information. It is a flowchart explaining the example of a process of the migration origin server. It is a flowchart explaining the example of a process of a migration destination server. It is FIG. (2) explaining the example of management information. It is a flowchart explaining the example (the 1) of the migration using a reservation threshold value and an upper limit threshold value. It is a flowchart explaining the example (the 2) of the migration using a reservation threshold value and an upper limit threshold value.

Hereinafter, embodiments will be described in detail with reference to the drawings.
FIG. 2 is a diagram for explaining an example of a problem of the virtualization system using the management server. In the virtualization system 100 of FIG. 2, the same components as those in FIG. The virtualization system 100 is an information processing system including a server 110a, a server 110b, and a management server 120. There may be a plurality of servers 110. The server 110a, the server 110b, and the management server 120 are connected via a communication network. The server 110a includes VMs 111a to 111c, a hypervisor 112a, and a pool 113a. The server 110b includes a VM 111d, a VM 111e, a hypervisor 112b, and a pool 113b.

  The VM 111 is a virtual machine that is operated by logically dividing a single server 110 on which various devices including a CPU, a memory, and an IO device are mounted as computer resources as if it were a plurality of servers. The VMs 111a to 111c are virtual machines that operate on the server 110a. The VMs 111d to 111e are virtual machines that operate on the server 110b. Each of the VMs 111 can operate an OS and an application. The hypervisor 112 includes an acquisition unit 114 and a processing unit 115, and manages each VM 111. For example, the hypervisor 112a manages the VMs 111a to 111c. The hypervisor 112b manages the VMs 111d to 111e. The pool 113 is a collection of computer resources that are not allocated to the VM 111 among the computer resources installed in the server 110.

  The management server 120 is a server that manages the server 110a and the server 110b. The management server 120 includes a management unit 121 and a storage unit 122. The storage unit 122 stores domain information, allocation information, and usage information of the VM 111. The domain information is information for specifying the VM 111 such as a domain name and a domain number. The allocation information is information indicating computer resources allocated to each VM 111. The usage information is information indicating computer resources used by the VM 111. Each piece of information stored in the storage unit 122 is information used by the management unit 121. The management unit 121 includes a monitoring unit 123, a determination unit 124, a selection unit 125, and an instruction unit 126. The monitoring unit 123 monitors the allocation amount of the computer resource allocated to the operating VM 111 and the usage amount of the computer resource used by the VM 111 out of the allocation amount of the computer resource at a predetermined opportunity. The monitoring unit 123 may monitor, for example, every predetermined time as a predetermined opportunity. As the usage index, the number of computer resources used by the VM 111 may be used, or the ratio of the number of used computer resources to the number of allocated computer resources may be used. Monitoring by the monitoring unit 123 is realized by acquiring the usage information of the VM 111 in each server 110 acquired by the acquisition unit 114. The monitoring unit 123 stores the acquired usage information in the storage unit 122. Based on the usage information stored in the storage unit 122, the determination unit 124 determines whether the index of the usage amount of the computer resource of each VM 111 exceeds a predetermined threshold. Here, when the usage index is the number of computer resources used by the VM 111, the number of computer resources is set as the threshold. When the usage index is the ratio of the number of used computer resources to the number of allocated computer resources, a value indicating a certain ratio is set as the threshold. The selection unit 125 selects a migration destination server based on usage information when there is a VM 111 whose computer resource usage index exceeds a predetermined threshold. The instruction unit 126 transmits an allocation amount of computer resources to be allocated to the VM 111 and a migration instruction to the server 110.

  The acquisition unit 114 of the hypervisor 112 acquires computer resource usage information as the operating state of the VM 111 in the server 110. The processing unit 115 changes the allocation amount of computer resources and performs migration processing in accordance with an instruction from the management server 120 side.

  In the virtualization system 100 of FIG. 2, the management server 120 acquires usage information and determines a usage state at a predetermined opportunity. Since this acquisition process is performed every predetermined time as a predetermined trigger, it is difficult for the management server 120 to immediately detect that the index of the usage amount of the computer resource of the VM 111 exceeds a predetermined threshold. . Similarly, even if the management server 120 runs out of computer resources additionally allocated to the VM 111 in the server 110a, it is difficult to immediately detect the shortage of computer resources. Furthermore, in the virtualization system 100, as the number of servers 110 managed by the management server 120 increases, the acquisition process and the determination process take longer. There is a problem that it takes time until the migration process is actually started even though the index of the usage amount of the computer resource of the VM 111 exceeds a predetermined threshold and it is better to execute the migration.

  Furthermore, in such a virtualized system 100, when the index of the amount of computer resources used by the VM 111 is larger than a predetermined threshold and the computer resources to be additionally allocated are insufficient in the server 110, the management server 120 It is preferable that can be detected immediately. For that purpose, the management server 120 should just increase the frequency of the monitoring process of the monitoring part 123. FIG. However, in a data center that holds many servers 110 and VMs 111, there is a problem that the network load increases when the management server 120 increases the frequency of monitoring processing. On the other hand, when the monitoring frequency is low, there is a problem that it takes time until the management server 120 detects that the computer resources to be additionally allocated are insufficient in the server 110.

  FIG. 3 is a diagram illustrating an example of a virtualization system according to this embodiment. The virtualization system 200 in FIG. 3 is an information processing system including a server 210a, a server 210b, and a storage device 220. The virtualization system 200 may further include a server 210. The server 210a, the server 210b, and the storage device 220 are connected via a communication network. The storage device 220 includes a storage unit 221 as a memory area where the server 210a and the server 210b can share and write and read information. The server 210a includes VMs 211a to 211c, a hypervisor 212a, and a pool 213a. The server 210b includes a VM 211d, a VM 211e, a hypervisor 212b, and a pool 213b.

  The VM 211 is a virtual machine that is operated by logically dividing one server 210 on which various computer resources including a CPU, a memory, and an IO device are mounted as if it were a plurality of servers. The VMs 211a to 211c are virtual machines that operate on the server 210a. The VMs 211d to 211e are virtual machines that operate on the server 210b. The server 210a and the server 210b may further operate a plurality of VMs. Each of the VMs 211 can operate an OS and an application. The hypervisor 212 manages the VM 211 in each server. For example, the hypervisor 212a manages the VMs 211a to 211c. The hypervisor 212b manages the VMs 211d to 211e. The pool 213 is a collection of computer resources that are not allocated to the VM 211 among the computer resources installed in each server 210.

  The storage device 220 includes a storage unit 221. The storage unit 221 stores management information for each server 210. The management information includes, for example, domain information, allocation information, usage information, and the like of the VM 211 executed on the server 210. Note that the storage device 220 according to the present embodiment does not manage a server like the management server 120 of FIG. The storage device 220 is a storage area used for sharing various information among the servers 210. The allocation information includes, for example, information such as an allocation amount of computer resources to each VM 211 and an increase / decrease amount of computer resources when adding a computer resource to each VM 211. The amount of computer resources allocated to the VM 211 is the amount of computer resources that the VM 211 is set to be usable. When the computer resource is added to the VM 211, the increase / decrease amount of the assignment of the computer resource is the additional VM 211 when the computer resource is used up to a predetermined threshold number or ratio in the computer resource amount assigned to the VM 211. This is the amount of computer resources allocated to.

  Details of the processing in the server 210a will be described as processing of the hypervisor 212 according to the present embodiment. The hypervisor 212a according to the present embodiment includes an acquisition unit 231, a determination unit 232, a selection unit 233, and a processing unit 234. For example, the acquiring unit 231 acquires the usage information of the VM 211 in the server 210a at predetermined time intervals. The acquisition unit 231 stores the acquired information in the storage unit 221 of the storage device 220. Based on the management information stored in the storage unit 221, the determination unit 232 determines whether the computer resource usage index used by the latest VM 211 exceeds a predetermined threshold. When the determination unit 232 uses the management information stored in the storage unit 221, the acquisition unit 231 acquires the management information from the storage device 220. The determination unit 232 further determines whether additional computer resources can be allocated to the VM 211 in the server 210.

  Here, as an example of the allocation amount of the computer resources to the VM 211a, it is assumed that there are 10 allocated CPUs. Hereinafter, a CPU is used as an example of computer resources allocated to the VM 211. Next, it is assumed that the amount of increase / decrease in CPU allocation to the VM 211a is two. Further, the predetermined threshold value of the CPU is 8 CPUs. Based on the management information stored in the storage unit 221, the determination unit 232 determines whether the CPU usage index of the latest VM 211a exceeds eight, which is a predetermined threshold value. When the index of the CPU usage amount of the VM 211a exceeds the predetermined threshold value of 8, the determination unit 232 determines whether or not an additional CPU can be allocated to the VM 211a in the server 210a. In other words, the determination unit 232 determines whether it is possible to allocate two CPUs that are the CPU increase / decrease amount in the server 210a. When there are two or more unused CPUs in the pool 213a, the determination unit 232 determines that an additional CPU can be allocated to the VM 211a in the server 210a. Then, the processing unit 234 allocates two CPUs to the VM 211a. By additionally assigning CPUs, it is possible to improve the processing performance of the VM 211a that is executing a large amount of processing. As an example, the usage index uses the number of CPUs in use among the number of assigned CPUs. As another example, the usage index may be, for example, a ratio of the number of used CPUs to the number of assigned CPUs.

  When there is no surplus CPU in the pool 213a, the selection unit 233 searches for the server 210 that holds in the pool 213 at least 12 CPUs to be allocated to the VM 211a among the servers 210 other than the server 210a, and selects them as migration destination servers. In the example of the virtualization system 200 of FIG. 3, it is assumed that the pool 213b of the server 210b holds 12 or more CPUs. The selection unit 233 selects the server 210b as the migration destination server of the VM 211a.

  The processing unit 234 includes information indicating migration of 12 CPUs and flag information indicating migration to management information corresponding to the migration destination server 210b stored in the storage unit 221 in the storage device 220. Write. The server 210b communicates with the storage device 220 every predetermined time as a predetermined opportunity. When the server 210b detects that it has been selected as the migration destination, the server 210b reserves a CPU to be allocated to the VM 211a in order to accept the VM 211a. The processing unit of the server 210b writes the latest allocation information of the server 210b in the management information. When the processing unit 234 of the server 210a confirms that the CPU for the VM 211a is assigned on the migration destination server 210b side, the processing unit 234 executes the migration of the VM 211a. The VM 211a is assigned 12 CPUs by the migration destination server 210b.

  When the migration process is completed, the processing unit 234 updates the management information of the storage unit 221 in the storage device 220 to the latest state. As with the hypervisor 212a, the hypervisor 212b also monitors each VM 211, and performs migration processing when the usage index of the VM 211 exceeds a predetermined threshold.

  In the virtualization system 200 of FIG. 3, since the hypervisor 212 in each server 210 monitors the operating state of its own VM 211, the migration is executed as soon as the computer resource usage index exceeds a predetermined threshold. Is possible. Further, since the hypervisor 212 in each server 210 monitors the operating state of its own VM 211, the analysis time does not increase even if the number of servers 210 increases. The virtualization system 200 of FIG. 3 can shorten the time from when the indicator of the usage amount of the computer resource of the VM 211 actually exceeds a predetermined threshold until the VM 211 is migrated.

  FIG. 4 shows an example of the hardware configuration of the server 210. The server 210 can be realized using, for example, an information processing apparatus (computer) as shown in FIG. 4, and includes the processor 11, the memory 12, the bus 15, the external storage device 16, and the network connection device 19. Further, as an option, the server 210 may include the input device 13, the output device 14, or the medium driving device 17.

  The processors 11a to 11n can be arbitrary processing circuits including a CPU. The processor 11 in FIG. 4 does not limit the number of processors. For example, the processors 11a to 11c are processors that execute the VM 211. Each VM 211 uses at least one processor 11. The processor 11d operates as a hypervisor 212 including an acquisition unit 231, a determination unit 232, a selection unit 233, a processing unit 234, and the like. The processors 11e to 11n are not assigned to the VM 211 and the hypervisor 212, and are unused processors 11. The processors 11e to 11n are held in the pool 213. The processor 11 can execute, for example, a program stored in the external storage device 16. The memory 12 appropriately stores data obtained by the operation of the processor 11 and data used for processing by the processor 11. The network connection device 19 is used for communication with other devices.

  The input device 13 is realized as, for example, a button, a keyboard, or a mouse, and the output device 14 is realized as a display or the like. The bus 15 connects the processor 11, the memory 12, the input device 13, the output device 14, the external storage device 16, the medium drive device 17, and the network connection device 19 so as to exchange data with each other. The external storage device 16 stores programs, data, and the like, and provides the stored information to the processor 11 and the like as appropriate. The medium driving device 17 can output the data of the memory 12 and the external storage device 16 to the portable storage medium 18 and can read programs, data, and the like from the portable storage medium 18. Here, the portable storage medium 18 may be any portable portable medium including a floppy disk, a Magnet-Optical (MO) disk, a Compact Disc Recordable (CD-R), and a Digital Versatile Disk Recordable (DVD-R). It can be a storage medium.

  FIG. 5 is a diagram (part 1) illustrating an example of management information. The management information includes domain information, allocation information, usage information, and migration information for each server. The domain information includes information such as the number of CPUs mounted on the server 210, a domain number, and a domain name. In the example of FIG. 5, the server 210a is equipped with 32 CPUs. The server 210a executes a plurality of VMs in the server, and domain numbers 1-1 to 1-4 and domain names (Ldom11 to Ldom14) are set in the settings of each VM. The server 210b executes a plurality of VMs in the server, and domain numbers 2-1 to 2-3 and domain names (Ldom21 to Ldom23) are set for each VM. Note that the VM with the domain number 1-1 is the VM 211a in FIG.

  The allocation information includes information such as priority, allocation amount, maintenance amount, increase / decrease amount, lower limit threshold, and upper limit threshold for each VM 211. The priority order is a priority level used for assigning a CPU mounted on the server 210. The allocated amount is the number of CPUs for which the VM 211 is set to be usable. The maintenance amount is the number of CPUs to be maintained at a minimum for operating the VM 211. The increase / decrease amount is the number of CPUs that are additionally allocated to the VM 211 when the usage amount of the CPU is high even though the VM 211 uses the allocated amount of CPU. The lower threshold is used to control the number of CPUs used by the VM 211. When the CPU usage index is lower than the lower limit threshold, the hypervisor 212 performs processing to reduce the number of CPUs allocated to the VM 211. The upper threshold value is a threshold value used for controlling the number of CPUs used by the VM 211. When the CPU usage index is higher than the upper threshold, the hypervisor 212 performs processing to increase the number of CPUs allocated to the VM 211. The usage information includes the number of CPUs used and the CPU usage rate indicating the ratio of the number of CPUs used among the number of CPUs allocated to the VM 211.

  The migration information includes information such as a reservation destination, a flag, and an allocation amount. The reservation destination is domain number information of the migration destination of the VM 211 by migration. The flag is information indicating that the server 210a that is the migration source has been selected as the target server for the migration process. The allocation amount is the number of CPUs allocated to the VM 211a after the migration process.

  The management information in FIG. 5 is an example when the number that is an index of the CPU usage amount of the VM 211a having the domain number 1-1 is increased from 7 to 9. In other words, the CPU usage rate of the VM 211a has increased from 70% to 90%. The determination unit 232 determines whether the CPU usage rate of the VM 211a exceeds the upper threshold value 80%. The CPU usage rate of the VM 211a exceeds the upper threshold 80%. Therefore, the hypervisor 212a tries to increase the number of CPUs allocated to the VM 211a. The determination unit 232 determines whether there is an unused CPU in the pool 213a. When there is an unused CPU in the server 210a, the processing unit 234 assigns an unused CPU to the VM 211a. When there is no unused CPU in the server 210a, the hypervisor 212a starts a migration process for moving the VM 211a to another server 210. The selection unit 233 detects a server 210b having a pool including unused CPUs. In the example of the management information in FIG. 5, there are 16 unused CPUs in the pool 213b of the server 210b. Therefore, the selection unit 233 selects the server 210b as the migration destination. In the example of FIG. 5, the determination unit 232 compares the CPU usage rate (%), which is an index of CPU usage, and the upper threshold (%), but the number of CPUs and the upper limit are used as an index of CPU usage. You may compare a threshold value (number).

  The processing unit 234 sets the migration destination domain number in the management information reservation destination corresponding to the migration source server 210a. The processing unit 234 sets a flag in the management information corresponding to the selected server 210b and the number of CPUs assigned to the VM 211a in the migration destination server 210b. In the example of FIG. 5, the processing unit 234 sets the migration destination domain number 2-4 as the reservation destination of the management information corresponding to the migration source server 210a. The processing unit 234 sets 1 as a flag in the management information corresponding to the selected server 210b, and sets the number of CPUs 12 to be assigned to the VM 211a in the migration destination server 210b.

  The hypervisor 212b of the server 210b communicates with the storage device 220 at a predetermined opportunity. The hypervisor 212b detects that a value “1” indicating the migration destination is set in the flag corresponding to the server 210b. The hypervisor 212b performs setting to allocate a CPU for the VM 211a. The hypervisor 212b writes the setting change to the management information. The processing unit 234 of the server 210a confirms that the CPU for the VM 211a has been assigned. The processing unit 234 of the server 210a executes migration of the VM 211a from the server 210a to the server 210b. The VM 211a is assigned 12 CPUs by the migration destination server 210b. A domain number 2-4 and a domain name Ldom11 are assigned to the VM 211a. When the migration process is completed, the processing unit 234 clears the migration information and updates it with the latest management information. The example of the management information in FIG. 5 may be management information such as a memory or an IO device, although it is used for the CPU usage index.

  FIG. 6 is a flowchart for explaining an example of processing of the migration source server. The acquisition unit 231 acquires the usage information of the VM 211 in the server 210a, and stores it in the storage unit 221 of the storage device 220 (step S101). Based on the management information stored in the storage unit 221, the determination unit 232 determines whether the CPU usage index of the latest VM 211 exceeds a predetermined threshold (step S102). When the CPU usage index of the VM 211 exceeds a predetermined threshold (YES in step S102), the determination unit 232 determines whether it is impossible to allocate an additional CPU to the VM 211 in the server 210a ( Step S103). If an additional CPU can be allocated to the VM 211 in the server 210a (NO in step S103), an additional CPU is allocated to the VM 211 in the server 210a (step S109). If no additional CPUs can be assigned to the VM 211 in the server 210a (YES in step S103), the selection unit 233 determines whether there are at least the number of unused CPUs to be assigned to other servers (step S104). . When there is an unused number of CPUs to be allocated to other servers (YES in step S104), the processing unit 234 includes information indicating that migration is performed to the management information stored in the storage unit 221 in the storage device 220. The flag information is written (step S105). The processing unit 234 confirms whether the migration preparation is completed at the migration destination by confirming whether the management information corresponding to the migration destination server 210b in the storage unit 221 has been updated (step S106). When the migration preparation is not completed at the migration destination (NO in step S106), the processing unit 234 repeats the process of step S106. When migration preparation is completed at the migration destination (YES in step S106), the processing unit 234 executes migration processing (step S107). The processing unit 234 updates the management information stored in the storage unit 221 in the storage device 220 with information about the latest VM (step S108). When the process of step S108 ends, the hypervisor 212 ends the process. When the process of S109 ends, the processing unit 234 passes the process to S108. If there is no unused number of CPUs to be allocated to other servers (NO in step S103), the selection unit 233 determines whether there are other server candidates (step S110). If there is another server candidate (YES in step S110), selection unit 233 returns the process to step S104. If there is no other server candidate (NO in step S110), the hypervisor 212 ends the process. When the index of the usage amount of the computer resource used by the VM 211 does not exceed the predetermined threshold (NO in step S102), the hypervisor 212 ends the process.

  As described above, since the hypervisor 212 in each server 210 monitors the usage state of the computer resources of the VM 211 in the server, migration can be executed as soon as the computer resource usage index exceeds a predetermined threshold. It is. Further, since the hypervisor 212 in each server 210 monitors the usage state of the computer resources of the VM 211 in the server, the analysis time does not increase even if the number of servers 210 increases. The virtualization system 200 of FIG. 3 can shorten the time from when the indicator of the usage amount of the computer resource of the VM 211 actually exceeds a predetermined threshold until the VM 211 is migrated.

  FIG. 7 is a flowchart for explaining an example of processing of the migration destination server. The hypervisor 212b of the server 210b determines whether information is included in the flag information corresponding to the server 210b in the management information stored in the storage unit 221 in the storage device 220 (step S201). If there is no flag information (NO in step S201), the hypervisor 212b of the server 210b ends the process. If there is flag information (YES in step S201), the hypervisor 212b of the server 210b allocates a CPU for the migrated VM 211a (step S202). The hypervisor 212b of the server 210b updates the management information stored in the storage unit 221 in the storage device 220 (step S203). The hypervisor 212b of the server 210b ends the process when S203 ends.

<Others>
FIG. 8 is a diagram (part 2) illustrating an example of management information. The allocation information in the management information of FIG. 8 further includes a reservation threshold value from the management information of FIG. The migration information in the management information of FIG. 8 further includes a monitoring time from the management information of FIG. The reservation threshold is used to detect a sign that the CPU usage index is increasing before the upper threshold. The monitoring time is an interval of time when the acquisition unit 221 acquires usage information and communicates with the storage device to write the information to the storage device.

  In the management information of FIG. 8, this is an example in which the CPU usage rate of the VM 211a having the domain number 1-1 increases to 90%. The determination unit 232 first determines whether the CPU usage rate of the VM 211a exceeds the reservation threshold value 70%. The CPU usage rate of the VM 211a exceeds the reservation threshold value 70%. The selection unit 233 reserves a server 210b having many unused CPUs as a migration destination. Since the CPU usage rate exceeds the threshold 70%, the processing unit 234 sets the VM 211a as a monitoring target. Therefore, a setting for shortening the monitoring interval is performed. The processing unit 234 sets the monitoring time corresponding to the migration source and migration destination servers in the management information to 30 s (seconds). 30 s is an example, and is an example of a time shorter than a normal monitoring interval. In the example of FIG. 8, the upper limit threshold value and the reservation threshold value of the management information may be the number of CPUs. In this case, the determination unit 232 compares the number of CPUs, which is an index of CPU usage of the VM 211a, with the number of CPUs set as the upper limit threshold and the reservation threshold.

  Next, the determination unit 232 determines whether the CPU usage rate of the VM 211a exceeds the upper limit threshold 80% at a time interval shorter than the normal monitoring interval. When the CPU usage rate of the VM 211a exceeds 80%, the hypervisor 212a attempts to increase the number of CPUs allocated to the VM 211a. The determination unit 232 determines whether there is an unused CPU in the pool 213a. When there is an unused CPU in the server 210a, the processing unit 234 assigns an unused CPU to the VM 211a. When there is no unused CPU in the server 210a, the hypervisor 212a starts a migration process for moving the VM 211a to another server 210. The selection unit 233 determines whether there are enough unused CPUs in the pool 213b of the reserved migration destination server 210b. If there are still enough unused CPUs in the reserved pool 213b of the migration destination server 210b, the processing unit 234 executes migration processing. When this migration process is completed, the processing unit 234 restores the monitoring interval.

  FIG. 9 is a flowchart illustrating an example (part 1) of migration using a reservation threshold and an upper limit threshold. The acquisition unit 231 acquires the usage information of the VM 211 in the server 210a and stores it in the storage unit 221 of the storage device 220 (step S301). Based on the management information stored in the storage unit 221, the determination unit 232 determines whether the CPU usage index of the latest VM 211 exceeds the reservation threshold (step S302). When the index of the usage amount of the CPU of the VM 211 exceeds the reservation threshold (YES in step S302), the selection unit 233 determines whether there is a server having more unused CPUs than the CPUs to be allocated to the VM 211. Determination is made (step S303). If there are more servers with unused CPUs than the CPUs to be allocated to the VM 211 (YES in step S303), the selection unit 233 reserves the server 210b as a migration destination (step S304). The processing unit 234 sets the monitoring interval of the migration source and migration destination servers 210 to be short (step S305). When the index of the usage amount of the CPU of the VM 211 does not exceed the reservation threshold value (NO in step S302), the hypervisor 212 ends the process. When there is no server having unused CPUs more than the CPUs to be allocated to the VM 211 (NO in step S303), the hypervisor 212 ends the process.

  FIG. 10 is a flowchart for explaining an example (part 2) of migration using a reservation threshold and an upper limit threshold. The acquisition unit 231 acquires the usage information of the VM 211 in the server 210a and stores it in the storage unit 221 of the storage device 220 (step S401). Based on the management information stored in the storage unit 221, the determination unit 232 determines whether or not the CPU usage index of the latest VM 211 exceeds a predetermined threshold (step S402). When the CPU usage index of the VM 211 exceeds a predetermined threshold (YES in step S402), the determination unit 232 determines whether it is impossible to allocate an additional CPU to the VM 211 in the server 210a ( Step S403). If an additional CPU can be allocated to the VM 211 in the server 210a (NO in step S403), an additional CPU is allocated to the VM 211 in the server 210a (step S409). If an additional CPU cannot be allocated to the VM 211 in the server 210a (YES in step S403), the selection unit 233 determines whether the migration destination reservation server has at least the number of unused CPUs to be allocated (step S404). ). If the migration destination reservation server has a number of unused CPUs to be allocated (YES in step S404), the processing unit 234 indicates that the management information stored in the storage unit 221 in the storage device 220 is migrated. And the flag information are written (step S405). The processing unit 234 confirms whether the migration preparation is completed at the migration destination by confirming whether the management information corresponding to the migration destination server 210b in the storage unit 221 has been updated (step S406). When the migration preparation is not completed at the migration destination (NO in step S406), the processing unit 234 repeats the process of step S406. When migration preparation is completed at the migration destination (YES in step S406), the processing unit 234 executes migration processing (step S407). The processing unit 234 updates the management information stored in the storage unit 221 in the storage device 220 with information about the latest VM (step S408). The processing unit 234 restores the monitoring interval (step S411). When the process of step S411 ends, the hypervisor 212 ends the process. When the process of S409 ends, the processing unit 234 passes the process to S408. When the migration destination reservation server does not have the number of unused CPUs to be allocated (NO in step S403), the selection unit 233 determines whether there is another server candidate (step S410). If there is another server candidate (YES in step S410), the selection unit 233 returns the process to step S404. If there is no other server candidate (NO in step S410), the hypervisor 212 ends the process. When the index of the usage amount of the computer resource used by the VM 211 does not exceed the predetermined threshold (NO in step S402), the hypervisor 212 ends the process.

  Thus, since the hypervisor 212 in each server 210 monitors the operating state of its own VM 211, migration can be executed as soon as the computer resource usage index exceeds a predetermined threshold. Further, since the hypervisor 212 in each server 210 monitors the operating state of its own VM 211, the analysis time does not increase even if the number of servers 210 increases. The virtualization system 200 of FIG. 3 can shorten the time from when the indicator of the usage amount of the computer resource of the VM 211 actually exceeds a predetermined threshold until the VM 211 is migrated.

Claims (7)

An information processing system including a plurality of information processing apparatuses capable of executing one or more virtual machines,
The information processing system includes:
A storage device that stores allocation information indicating computer resources allocated to the one or more virtual machines and usage information indicating computer resources used by the one or more virtual machines;
The first information processing device among the plurality of information processing devices is:
An acquisition unit that acquires, from the storage device, the usage information of a computer resource used by a virtual machine that is operating in the first information processing apparatus;
When the index of the usage amount of the computer resource based on the usage information exceeds a predetermined threshold and the computer resource to be additionally allocated to the operating virtual machine in the first information processing apparatus is insufficient, Second information having more unused computer resources than the computer resources allocated to the operating virtual machine among the plurality of information processing devices based on the allocation information stored in the storage device A selection unit that selects a processing device as an information processing device that takes over the processing of the operating virtual machine;
Allocation information for allocating more computer resources to the operating virtual machine than the computer resources allocated to the operating virtual machine is written in the storage device, and the operating virtual machine is stored in the second information processing apparatus. A processing unit that takes over the processing of the machine;
An information processing system comprising: The information processing system according to claim 1, wherein the second information processing apparatus takes over processing of the operating virtual machine based on the allocation information written in the storage device. The acquisition unit acquires the use information of a computer resource used by a virtual machine that is operating in the first information processing apparatus, and updates the use information of the storage device at a predetermined opportunity. The information processing system according to claim 1. In a system including a plurality of information processing apparatuses capable of executing one or more virtual machines,
The system includes a storage device that stores allocation information indicating computer resources allocated to the one or more virtual machines, and usage information indicating computer resources used by the one or more virtual machines,
A first information processing device among the plurality of information processing devices is
Obtaining the usage information of the computer resource used by the virtual machine operating in the first information processing apparatus from the storage device;
When the index of the usage amount of the computer resource based on the usage information exceeds a predetermined threshold and the computer resource to be additionally allocated to the operating virtual machine in the first information processing apparatus is insufficient, Second information having more unused computer resources than the computer resources allocated to the operating virtual machine among the plurality of information processing devices based on the allocation information stored in the storage device Select a processing device as an information processing device that takes over the processing of the operating virtual machine,
Allocation information for allocating more computer resources to the operating virtual machine than the computer resources allocated to the operating virtual machine is written in the storage device, and the operating virtual machine is stored in the second information processing apparatus. Take over the processing of the machine,
An information processing method characterized by the above. The information processing method according to claim 4, wherein the second information processing apparatus takes over the processing of the operating virtual machine based on the allocation information written in the storage device. In a system including a plurality of information processing apparatuses capable of executing one or more virtual machines,
The system includes a storage device that stores allocation information indicating computer resources allocated to the one or more virtual machines, and usage information indicating computer resources used by the one or more virtual machines,
Among the plurality of information processing devices, the first information processing device
Obtaining the usage information of the computer resource used by the virtual machine operating in the first information processing apparatus from the storage device;
When the index of the usage amount of the computer resource based on the usage information exceeds a predetermined threshold and the computer resource to be additionally allocated to the operating virtual machine in the first information processing apparatus is insufficient, Second information having more unused computer resources than the computer resources allocated to the operating virtual machine among the plurality of information processing devices based on the allocation information stored in the storage device Select a processing device as an information processing device that takes over the processing of the operating virtual machine,
Allocation information for allocating more computer resources to the operating virtual machine than the computer resources allocated to the operating virtual machine is written in the storage device, and the operating virtual machine is stored in the second information processing apparatus. Processing that takes over the processing of the machine,
Information processing program to be executed. The information processing program according to claim 6, wherein the second information processing apparatus takes over processing of the operating virtual machine based on the allocation information written in the storage device.

Images