JP4438807B2 - Virtual machine system, management server, virtual machine migration method and program - Google Patents

Description

  The present invention relates to a virtual machine system including a plurality of servers capable of executing virtualization software for constructing at least one virtual machine, a management server in the virtual machine system, a virtual machine migration method, and a program.

  In recent years, virtualization technology has attracted attention for servers, and many servers have been introduced. Famous virtualization software (VMM: Virtual Machine Monitor) to which this technology is applied include "VMware ESX Server" of VMware, "Xen" of Xen Source, and "Virtual Server" of Microsoft. Further, since a virtual machine (VM) is configured as an image on the storage, there is an advantage that it is easy to manage.

  Although it is possible for a VM to move a VM to another server by using VM migration, the same migration source VMM and migration destination VMM are usually used. In addition, when the migration source VMM and the migration destination VMM are different, it is necessary to perform conversion using a converter so that the VM running on the migration source VMM can be handled by the migration destination VMM. As this type of software, for example, “VMware Converter” of VMware, etc. can be mentioned.

  Regarding these VM migrations, several methods for finding an appropriate migration destination server have been conventionally considered. For example, Patent Document 1 can be cited.

  However, in the conventional method described in Patent Document 1 (Japanese Patent Application No. 2006-143626), the types of the migration source VMM and the migration destination VMM are not mentioned, so the types of the migration source VMM and the migration destination VMM are different. If this happens, it may not be possible to find an appropriate destination server. As a result, there arises a problem that the migration of the VM itself cannot be performed even though the migration source VMM and the server running the different VMM have sufficient resources.

On the other hand, other conventional methods solve this problem by making the type of VMM to be introduced the same, and by determining the migration destination of the VM in advance by the administrator. It was.
Japanese Patent Application No. 2006-143626

  Generally, a business system is composed of a plurality of servers, but the operating state of each of these servers is not always constant. That is, for example, when a certain server is in a low load state, there is a surplus in processing capacity, but since this surplus cannot be accommodated to other server devices, load imbalance is likely to occur between the server devices. It is difficult to operate the entire system efficiently. On the other hand, in a business system using VMM and VM, if the VM can always be arranged on an appropriate VMM according to the processing load status, the load can be redistributed, and the processing capacity of the server apparatus can be efficiently Available to: However, in reality, the migration for relocation has the following problems, and sufficient efficiency cannot be exhibited.

  The first problem is that, when the VM is migrated in the conventional method, as described above, the VMM to be migrated is limited to the same type as the VMM on which the VM is currently operating. In an environment where VMM is mixed, the optimum selection cannot be performed.

  In general, a VM is created on the premise of a specific type of VMM, and cannot operate with any other type of VMM. However, even if the migration source VMM and the migration destination VMM are different, the VM image can be converted by converting the format of the disk image, setting file, etc. (hereinafter referred to as VM image) required for VM operation to match the migration destination VMM. Can be operated with different VMMs (this process is hereinafter referred to as conversion). In the conventional method, conversion is not considered in the migration, and the migration destination is determined from the same VMM in consideration of only the CPU usage rate and the memory usage. As a result, when the VMM types are different, an appropriate migration destination server cannot be found.

  The second problem is that the downtime of the VM accompanying the conversion becomes longer. In order to perform the conversion, a processing time for the conversion is required. Usually, the target VM must be stopped (shut down) during this processing time. The downtime associated with VM conversion is relatively long compared to VM migration between identical VMMs, which is also a major factor that conversion is not considered as part of a practical migration method.

(the purpose)
An object of the present invention is to provide a virtual machine system that can determine an appropriate migration destination server without depending on the type of the VMM and suppress the downtime of the VM due to the conversion.

First virtual machine system of the present invention, in a virtual machine system comprising a plurality of servers capable of executing virtualization software to build at least one or more virtual machines, the predetermined virtualization software a predetermined servers the virtual machine that operation, from the virtualization software on multiple other servers, have a management server to migrate on the virtualization software on the optimal other server migration, the management server, from the server Obtain information about hardware, virtual machine system, and virtual machine, and evaluate the difference in hardware configuration between migration time information indicating migration process time, stop time information indicating virtual machine stop time during migration process Penalty information used for migration, migration information indicating migration methods based on the types of virtualization software and destination virtualization software, and Based on, characterized in that the virtualization software on multiple other servers, selects the virtualization software on another optimum server migration.

(Function)
With the above configuration, the management server detects virtualization software on the server that is optimal for migration from virtualization software on a plurality of servers.

  Here, the means for solving the problem will be described in more detail. The means for solving the first problem is that the VMM operates in consideration of not only the CPU usage rate and the memory usage but also the time required for the migration. The migration destination server (migration destination server) is determined. In order to make this solution possible, the following four tables are provided in the machine (management server) that issues an instruction to the server on which the VMM is operating in advance and manages it. The four tables are a conversion time table 371 in which the conversion processing time is recorded, a down time table 372 in which how many minutes the VM is stopped during the migration process, and a hardware configuration is evaluated. A penalty table 373 used for the conversion, and a conversion table 375 in which software necessary for the conversion process is recorded. In the present invention, an evaluation value is calculated based on information in these tables and information collected from the server, and by using the value, an appropriate migration destination server can be determined without depending on the type of VMM. .

  Further, as a solution to the second problem, a method of copying a VM image to be converted and converting the copied image is used. This approach can minimize the amount of files that must be updated during VM downtime. As a result, the VM downtime can be shortened.

  According to the present invention, the management server detects the virtualization software on the server that is optimal for migration from the virtualization software on the plurality of servers. Therefore, in an environment where a plurality of different VMMs are mixed, the type of VMM and the VM It is possible to find an appropriate destination server in consideration of the time required for migration.

(First embodiment)
Hereinafter, a first embodiment of the present invention will be described in detail with reference to the drawings.

(Configuration of the first embodiment)
FIG. 1 is a block diagram showing a configuration according to the present embodiment.

  Referring to FIG. 1, the present embodiment includes a server 100, a server 200, a management server 300, a storage 400, a LAN (Local Area Network) switch 500, and an FC switch 600.

  The server 100 is a computer that operates under program control, and includes a LAN module 140, an HBA (Host Bus Adapter) 150 for connecting to a SAN (Storage Area Network), and a VMM (Virtual Machine Monitor) 130. It is working.

  In the VMM 130, an information collecting unit 131, a load measuring unit 132, and an access monitoring unit 133 are operating. Further, VMs (Virtual Machines) 110 and 120 operate using the VMM 130.

  The VMM 130 is software (hypervisor) necessary for operating the VM, and has a function of allocating resources such as a CPU (Central Processing Unit) and memory of the server 100 to each VM.

  Each of the VMs 110 and 120 is a logical machine operating on the VMM 130 and is an emulation of a physical machine. That is, the VMs 110 and 120 are configured by an environment in which OSs (Operating Systems) 1101, 1201, and 2101 operate, and the applications 1102, 1202, and 2102 operate on it.

  The information collection unit 131 operates in response to an instruction from the information acquisition unit 310 in the management server 300, collects information required by the information acquisition unit 310 from the server 100, shapes the information into an appropriate form, and then sends the information to the information acquisition unit 310. It has a function to transmit information.

  The load measuring unit 132 always measures the load information (CPU usage rate and memory usage) of the server 100, receives the instruction of the information collecting unit 131, shapes the load information collected in the past, and collects information A function of transmitting information to the unit 131;

  The access monitoring unit 133 operates in accordance with an instruction from the processing command unit 330 of the management server 300, and records a function for monitoring a specific VM on the server 100 and a file updated in the monitoring target VM during access monitoring. And a function of transmitting the recorded information to the management server 300.

  Since the server 200 has the same configuration as the server 100, a description thereof will be omitted. However, the VMM 230 need not be the same as the VMM 130, and this is also a feature of the present invention.

  The management server 300 includes a storage unit 370, a LAN module 380, and an HBA 390, and an OS 360 is operating. On the OS 360, an information acquisition unit 310, a migration destination determination unit 320, a processing command unit 330, a conversion processing unit 340, The live migration unit 350 is operating.

  The management server 300 includes a conversion time table 371 (FIG. 2), a down time table 372 (FIG. 3), a penalty table 373 (FIG. 4), a parameter table 374 (FIG. 5), and a conversion table 375 (FIG. 6) as information. It is stored in the storage unit 370.

  FIG. 2 is a diagram for explaining the conversion time table 371 according to the present embodiment. The conversion time table 371 is information indicating how many minutes it takes to convert the data capacity of 1 GB when performing conversion. For example, referring to FIG. 2, it can be seen that the conversion from VMware to Xen takes 2 minutes per GB.

  FIG. 3 is a diagram for explaining the downtime table 372 according to this embodiment. The downtime table 372 is information indicating how many minutes the converted VM is stopped during the conversion process. For example, referring to FIG. 3, in order to migrate a VM on VMware to Xen, it is necessary to stop the migrating VM for 10 minutes.

  FIG. 4 is a diagram for explaining the penalty table 373 according to the present embodiment. The penalty table 373 indicates information used when calculating the evaluation value from the evaluation function. By using the information in the penalty table 373, a difference in hardware (HW) configuration between the migration source server and the migration destination server can be reflected in the evaluation value. For example, referring to FIG. 4, if one NIC is missing, the penalty value is 3.

  FIG. 5 is a diagram for explaining the parameter table 374 according to this embodiment. The parameter table 374 is information indicating various parameters used when calculating the evaluation value. By using the information in the parameter table 374, it is possible to allocate parameters for each VM, and a more appropriate evaluation value can be calculated. The value of the parameter table 374 is entered by the administrator as required when registering the VM. In addition, when the administrator does not input a value, a default value is set.

  FIG. 6 is according to the present embodiment. It is a figure explaining the conversion table 375. FIG. The conversion table 375 is information indicating which method should be used for conversion from the types of the migration source VMM and the migration destination VMM when performing the conversion. A blank place in the conversion table 375 indicates that there is no conversion means. Referring to FIG. 6, VMware to Xen can be converted using software called vm2xen. However, since nothing is written in the conversion table 375 from Xen to the physical machine, there is no conversion means. Show.

  The information acquisition unit 310 is on the management server 300 and requests the information collection units 131 and 231 of the servers 100 and 200 to collect various information (HW, VMM, VM information, and load information) from the servers 100 and 200. It has a function to organize the information.

  The migration destination determination unit 320 has a function of determining a migration destination server and a function of calculating an evaluation value of the server 200 to be evaluated using an evaluation function in order to determine the migration destination server. In order to calculate the evaluation value based on the evaluation function, the migration destination determination unit 320 converts the conversion time table 371 (FIG. 2), the down time table 372 (FIG. 3), the penalty table 373 (FIG. 4), and the parameter table 374 (FIG. 5). ) Have access to.

  The processing command unit 330 has a function of giving various commands to the servers 100 and 200.

  The conversion processing unit 340 has functions of image duplication and conversion for a VM that needs to be converted. In order to realize this function, the conversion processing unit 340 has read / write authority to the storage 400, access authority to the conversion table 375 (FIG. 5), and the conversion time table 371 (FIG. 2) and down time table 372 (FIG. 3). Has update authority.

  The live migration unit 350 has a function of determining whether or not live migration can be used, and a function of performing live migration when live migration can be used.

  The storage 400 has FC and LAN interfaces, and belongs to the same network as the servers 100 and 200 and the management server 300. In the storage 400, images of the VMs 110, 120, and 210 operating on the VMMs 130, 230 are recorded.

  The LAN switch 500 is connected to the servers 100 and 200, the management server 300, and the storage 400 to form one network.

  Servers 100 and 200, management server 300, and storage 400 are connected to FC switch 600 to form one SAN.

  Here, “VM migration” refers to a technology for moving a certain VM from the VMM currently being operated (migration source VMM) to the management of another VMM (migration destination VMM) and operating it. .

  Here, the hardware configuration of the management server 300 will be described.

  FIG. 7 is a block diagram showing a hardware configuration of the management server 300 according to this embodiment.

  Referring to FIG. 7, the management server 300 according to the present invention can be realized by a hardware configuration similar to that of a general computer device, and is a main memory such as a CPU (Central Processing Unit) 1001 and a RAM (Random Access Memory). A main storage unit 1002 used for a data work area and a temporary data save area, a communication control unit 1003 for transmitting and receiving data via the network 2000, a display unit 1004 such as a liquid crystal display, a printer and a speaker, a keyboard, A hard disk including an input unit 1005 such as a mouse, an interface unit 1006 that transmits and receives data by connecting to a peripheral device, a ROM (Read Only Memory), a magnetic disk, and a nonvolatile memory such as a semiconductor memory Auxiliary storage unit 1007 is a disk device, and a system bus 1008 for connecting the above components of the information processing apparatus to each other.

  The management server 300 according to the present invention is implemented by mounting a circuit component made of a hardware component such as an LSI (Large Scale Integration) in which a program for realizing such a function is incorporated in the management server 300. Of course, it can be realized by software by executing a program for providing each function of each component described above by the management server 300 on the computer processing apparatus.

  That is, the CPU 1001 implements each function described above in software by loading a program stored in the auxiliary storage unit 1007 into the main storage unit 1002 and executing the program, and controlling the operation of the management server 300.

  The servers 100 and 200 may have the above-described configuration, and each function described above may be realized in hardware or software.

(Operation of the first embodiment)
Next, the operation of the present embodiment will be described in detail with reference to the flowcharts of FIGS. 8, 9, 10, and 11.

  Consider a case where the VM 120 in the server 100 is migrated.

  With reference to FIG. 8, an outline of the operation of the present embodiment will be described.

  This flowchart is started when the migration source server 100 requests the management server 300 to migrate the VM 120.

  First, the management server 300 uses the information acquisition unit 310 to acquire the hardware, VMM, and load information of the servers 100 and 200 and the configuration information of the VM 120 to be migrated (S101). Details of the acquisition method will be described later.

  Based on the information acquired in S101, the migration destination determination unit 320 in the management server 300 calculates the evaluation value of the server 200 that is the migration destination target, and determines the migration destination server (S102). If all the migration destination target servers do not satisfy the condition, there is no migration destination server. The evaluation function and the migration destination server determination method will be described later.

  The migration destination determination unit 320 checks whether or not a migration destination server has been determined (S103). If there is no server that can be migrated, for example, a message indicating that migration cannot be performed is displayed on the presentation unit 1004, and the process ends. (S108).

  When the migration destination server is determined, the live migration unit 350 determines whether live migration can be used (S104). Live migration is a technology for migrating VMs with zero downtime. Whether or not live migration can be used is determined by referring to the conversion table 375. If live migration can be used, live migration is performed in the live migration unit 350 (S105), and the management server 300 ends the entire processing. If live migration is not available, the conversion processing unit 340 performs step S106. Execute.

  When live migration cannot be used, the conversion processing unit 340 converts the image format of the VM 120 to an image format that can be handled by the VMM 230 on the migration destination server 200 (S106). At the end of this operation, the VM 120 in the migration source server 100 is shut down.

  Finally, the process command unit 330 activates the VM 120 that has been migrated to the migration destination server 200 (S107), and the process ends.

  Next, with reference to FIG. 9, the various information acquisition part of this Embodiment is demonstrated in detail. This process is for the management server 300 to acquire various information from the servers 100 and 200, and is executed in S101 of FIG. In FIG. 9, the state when the management server 300 acquires information on the server 100 is described, but the same applies when acquiring information from the server 200.

  First, the information acquisition unit 310 in the management server 300 requests the information collection unit 131 on the server 100 to collect information (S201). At this time, the information acquisition unit 310 informs the information collection unit 131 of the type and format of information desired.

  Next, the information collection unit 131 executes step 203 if the information acquisition unit 310 has requested collection of hardware and VMM configuration information, and skips step 203 if the hardware and VMM information is not needed. Execute (S202).

  The information collection unit 131 acquires information regarding the hardware configuration and the VMM of the server 100 (S203). The hardware configuration information includes the number of CPUs, the number of NICs and HBAs, and connection destination information. VMM-related information includes VMM type and version information

  Next, the information collection unit 131 executes step 205 if the information acquisition unit 310 requests information collection of the VMs 110 and 120 on the server 100, and executes step 205 if no VM information is required. First, step 206 is executed (S204).

  The information collection unit 131 acquires information about the VM 120 to be migrated (S205). Information to be acquired includes the number of virtual CPUs, the correspondence between the number of virtual LANs and the physical LAN, and the correspondence between the number of virtual HBAs and the physical HBA.

  Next, the information collection unit 131 executes Step 207 if the information acquisition unit 310 needs to collect load information, and executes Step 208 if the load information is not needed (S206).

  The information collection unit 131 collects load information from the load measurement unit 132 in the server 100 (S207). The load measuring unit 132 constantly measures the load on the server 100 and the VMs 110 and 120, and after shaping the load information on the CPU and the memory of the server 100 and the VMs 110 and 120 into an appropriate format, hand over. Here, the appropriate format of the load information is a format presented by the information acquisition unit 310 of the management server 300. Specifically, it is sampled at regular intervals from load information for a certain period in the past, statistical information such as average and variance from load information for a certain period, or pattern analysis from load information for a certain period It is possible to extract the characteristics of load fluctuations using or feature extraction. Other information may be used if it can be used. Alternatively, a plurality of pieces of information can be used in combination.

  Finally, the information collection unit 131 of the server 100 collectively sends the collected various information to the information acquisition unit 310 included in the management server 300 (S208).

  Next, the migration destination determination algorithm according to the present embodiment will be described in detail with reference to FIG. This process determines the migration destination server based on the information acquired from the server and the information in the held table, and is executed in S102 of FIG.

  First, the migration destination determination unit 320 of the management server 300 calculates an evaluation value of the server 200 to be evaluated using an evaluation function (S301). Here, the following evaluation function is considered. However, parameters necessary for the following evaluation function are obtained from the parameter table 374 (FIG. 5).

と定義し、
ＣＰＵの評価関数を、

と定義し、
メモリの評価関数を、

と定義し、
コンバート処理時間の評価関数を、

と定義し、
ダウンタイムの評価関数を、

と定義し、
ペナルティの評価関数を、

と定義すると、
Ｓ３０１で算出する計算値は、

によって求められる。
ただし、以下の条件が課される。

here,

And define
CPU evaluation function

And define
Memory evaluation function

And define
The evaluation function of the conversion processing time is

And define
Downtime evaluation function,

And define
Penalty evaluation function

Defined as
The calculated value calculated in S301 is

Sought by.
However, the following conditions are imposed.

  Here, the CPU free rate and the memory free capacity are respectively the CPU usage rate currently consumed on the server 100 by the VM 120 to be migrated from the free rate (capacity) of the server to be evaluated as the migration destination. Calculated by subtracting memory usage. For this reason, when the CPU utilization rate and memory utilization amount of the server to be evaluated as the migration destination are smaller than the CPU utilization rate and memory usage amount used by the migrating VM 120, these values are negative. .

  In the above evaluation function, the part after the conversion processing time evaluation function works effectively when the configuration of the migration source server and the migration destination server is different, and is one of the most important parts of the present invention.

  The evaluation of the conversion processing time is a part for evaluating the time required for the conversion processing. The downtime evaluation is a part for evaluating the downtime of the migrated VM. Penalty evaluation is the part that evaluates hardware differences. In this evaluation function, a higher value is obtained for a server suitable as a migration destination server.

  Next, the migration destination determination unit 320 determines whether or not the evaluation value has been calculated for all the evaluation target servers. If not, the migration destination determination unit 320 calculates the evaluation value for that server, and all the evaluation target servers If the evaluation value is calculated, the following processing is executed (S302).

  Finally, the migration destination determining unit 320 determines a server having an evaluation value equal to or higher than a predetermined threshold and having the highest evaluation value as the migration destination server, and ends the process (S303). If the evaluation value is low for all the servers and does not reach the threshold value, the migration destination server ends the process without being found.

  Regarding the setting of the threshold value, for example, when selecting a condition that satisfies the conditions for all items to be evaluated, the threshold value is set to 0, and the value of L of the evaluation function is set to 0. This is because, in all the items to be evaluated, the item that does not satisfy the condition has an evaluation value of 0 or a negative value.

  Finally, with reference to FIG. 11, the conversion processing part of the present embodiment will be described in detail. This process occurs when the VMM types of the migration source server and the migration destination server are different, and corresponds to S106 in FIG.

  First, the conversion processing unit 340 of the management server 300 checks software necessary for conversion using the conversion table 375 (FIG. 6) (S401).

  Next, the command processing unit 330 of the management server 300 requests the access monitoring unit 133 of the migration source server 100 to monitor the disk access of the VM 120 to be migrated (S402). Thereafter, the access monitoring unit 133 records the updated file in the VM 120 to be migrated. The reason is that no contradiction occurs between images before and after the migration of the VM 120 to be migrated. This is because the image of the VM 120 to be migrated is copied to the storage and the conversion process is performed on the duplicated image in the next process, so that even if the file existing in the VM 120 to be migrated is updated during the conversion process This is because it is not reflected in the image after conversion. Therefore, it is necessary to maintain consistency by monitoring the disk access of the VM 120 to be migrated and applying the updated file after the conversion is completed to the converted image.

  Next, the conversion processing unit 340 of the management server 300 copies the image of the migrated VM 120 in the storage 400 (S403). The management server 300 can continue the process without shutting down the migrating VM 120 when performing the conversion process by acquiring the copy.

  The conversion processing unit 340 performs conversion processing on the image copied in S403 (S404). The software used for the conversion process is the one checked in S401.

After the conversion is completed, the conversion processing unit 340 updates the conversion time table 371 (FIG. 2) (S405). In the conversion time table 371 (FIG. 2), the conversion processing time is recorded, but the initial state is not an accurate conversion time, but an administrator or system builder inputs an approximate value. Therefore, in order to approximate the optimum value, the conversion time is updated by using the following equation.

  Here, N is the number of conversions when the migration source server and the migration destination server are the same. By applying this equation (8), the average of the time taken for conversion can be set as a new conversion time.

  Next, the instruction processing unit 330 notifies the access monitoring unit 133 of the end of the disk access monitoring of the migrating VM 120 and shuts down the migrating VM 120 (S406). After this processing, the VM 120 does not operate on the migration source server 100. Further, the access monitoring unit 133 notifies the conversion processing unit 340 in the management server 300 of the information on the file changed by the VM 120 on the migration source server 100.

  The conversion processing unit 340 in the management server 300 updates the changed portion of the converted image based on the information from the access monitoring unit 133 (S407).

  The instruction processing unit 330 activates the VM 120 that has been migrated to the migration destination server 200 (S408).

Finally, the conversion processing unit 340 in the management server 300 updates the values in the downtime table 372 (FIG. 3). The initial state of the downtime table 372 (FIG. 3) is not an accurate downtime, and an administrator or the like inputs an approximate value as in the conversion time table 371. Therefore, in order to bring it closer to a more accurate value, the following formula is used to update the downtime.

  Here, N is the same as N appearing in the conversion time update formula (formula (8)). The downtime table 372 is updated by applying the equation (9). The process ends here.

  In the configuration of the present embodiment, both servers are in a state where a VMM is installed. However, as long as the conversion table 375 indicates that the VMM is not installed, the VMM is not installed in the server. The present invention is applicable. However, in this case, since the VMM is not operating on the server, the information acquisition unit, the load measurement unit, and the access monitoring unit must be operated on the OS.

(Effects of the first embodiment)
According to the present embodiment, the following effects can be achieved.

  First, in an environment in which a plurality of different VMMs are mixed, an appropriate migration destination server can be found in consideration of the type of VMM and the time required for VM migration.

  The reason is that VMM and hardware information is acquired from all servers, the conversion processing time and the downtime of the VM to be migrated are grasped, and the evaluation value is calculated based on these information. Therefore, an appropriate migration destination server can be found by looking at the calculated evaluation value.

  Secondly, in a situation where live migration cannot be used for VM migration and conversion processing is performed, the downtime of the VM occurring during the migration processing can be shortened.

  The reason is that when the VM is converted, it is not necessary to shut down the VM to be migrated during the conversion process by duplicating the image to be converted once. The file updated in the VM during the conversion can be maintained as a difference with the image after conversion by updating the converted image as a difference.

  Third, by setting a threshold value for the evaluation value, it is possible to prevent the operation of forcibly determining the migration destination server when there is no appropriate server as the migration destination.

  Fourth, since it is not necessary to use an intermediate output when converting a VM to another VM format, it is possible to reduce the processing load of migration.

  Fifth, in general, the capacity of a program is about several MB (megabytes), but a VM is usually several GB (gigabytes) to several tens of GB. Since conversion to the VM format requires a great amount of processing time, the change processing situation becomes useless. On the other hand, according to the present embodiment, since the VM can be converted into another VM format in one step without using an intermediate output, the migration process can be performed quickly.

  Although the present invention has been described with reference to the preferred embodiments, the present invention is not necessarily limited to the above embodiments, and various modifications can be made within the scope of the technical idea. .

  For example, in the first embodiment, two servers and three VMs are shown. However, the number of servers is N, and the number of VMs existing in each server is within the upper limit of the servers. Needless to say, the present invention can be applied.

  Further, for example, by adding a new VMM entry to a table such as the conversion table 375 or the downtime table 372, a VMM (for example, Microsoft Virtual Server, etc.) not written in the first embodiment is also used. The present invention can be applied.

  Furthermore, the mathematical formula proposed in the present invention is only one example, and the present invention can be similarly handled even if another mathematical formula is used.

  Furthermore, in the first embodiment, data is stored in the SAN, but the present invention can be applied regardless of the type of storage. However, when using local storage, it takes a lot of network resources and time because the image must be copied over the network when it is duplicated.

  Furthermore, the present invention can be similarly applied when the object itself described as “server” in the present embodiment is virtually configured. For example, in the field of grid computing and the like, a basic technology for handling a plurality of computers as a single computer has appeared, but the present invention is applicable even when a VM is migrated to such a virtual object. Is applicable as well.

  The present invention can also be applied to a system using server virtualization technology in order to improve the operation efficiency of the entire system.

It is a block diagram which shows the structure by the 1st Embodiment of this invention. It is a figure explaining the conversion time table 371 by 1st Embodiment. It is a figure explaining the downtime table 372 by 1st Embodiment. It is a figure explaining the penalty table 373 by 1st Embodiment. It is a figure explaining the parameter table 374 by 1st Embodiment. It is a figure explaining the conversion table 375 by 1st Embodiment. It is a block diagram which shows the hardware constitutions of the management server 300 by 1st Embodiment. It is a flowchart which shows operation | movement of 1st Embodiment. It is a flowchart which shows operation | movement of 1st Embodiment. It is a flowchart which shows operation | movement of 1st Embodiment. It is a flowchart which shows operation | movement of 1st Embodiment.

Explanation of symbols

100, 200: Server 110, 120: VM
1101, 1201, 2101: OS
1102, 1202, 2102: Application 1103, 1203, 2103: Virtual LAN
1104, 1204, 2104: Virtual host bus adapter (HBA)
130, 230: VMM
131, 231: Information collection unit 132, 232: Load measurement unit 133, 233: Access monitoring unit 140, 240: LAN module 150, 250: Host bus adapter (HBA)
300: Management server 310: Information acquisition unit 320: Migration destination determination unit 330: Processing instruction unit 340: Conversion processing unit 350: Live migration unit 360: OS
370: Storage unit 371: Conversion time table 372: Down time table 373: Penalty table 374: Parameter table 375: Conversion table 380: LAN module 390: Host bus adapter (HBA)
400: Storage 500: LAN (Local Area Network) switch 600: FC switch 1001: CPU
1002: Main storage unit 1003: Communication control unit 1004: Presentation unit 1005: Input unit 1006: Interface unit 1007: Auxiliary storage unit 1008: System bus 2000: Network

Claims (16)

In a virtual machine system comprising a plurality of servers capable of executing virtualization software for constructing at least one virtual machine,
Management for migrating the virtual machine operated by predetermined virtualization software on a predetermined server from the virtualization software on a plurality of other servers onto the virtualization software on another server optimal for migration have a server,
The management server
Obtain information about the hardware, the virtual machine system, and the virtual machine from the server,
Migration time information indicating the migration processing time, stop time information indicating the stop time of the virtual machine during the migration process, penalty information used for evaluating a difference in hardware configuration, the virtualization software, and the migration destination And migration information indicating a migration method based on the type of virtualization software of
A virtual machine system , wherein the virtualization software on another server that is optimal for migration is selected from the virtualization software on the plurality of other servers . The management server
Obtain information about the hardware, the virtual machine system, and the virtual machine from the plurality of servers, based on the CPU usage rate, memory usage, and time required for migration,
The virtual machine system according to claim 1, wherein the virtual software on another server that is optimal for migration is selected from the virtual software on the plurality of other servers. The management server
The virtual machine system according to claim 1 or 2 , wherein an evaluation value indicating an index suitable as a migration destination is calculated based on information acquired from a plurality of the servers. The management server
4. The virtual machine system according to claim 3, wherein the migration destination server is determined or not determined based on a relationship between a predetermined threshold and the evaluation value . The management server
After determining the migration destination server, determine whether live migration is possible based on the migration information,
If possible, perform the migration using live migration,
3. If not possible, the image format of the virtual machine to be migrated is migrated to an image format that can be handled by the virtualization software on the migration destination server. Virtual machine system . The virtual machine system according to any one of claims 1 to 5, wherein a file at the end of migration existing in the virtual machine to be migrated is applied to the migrated image . The server has means for monitoring disk access of the virtual machine to be migrated;
The virtual machine system according to claim 6, further comprising recording means for recording a file updated in the virtual machine to be migrated . 8. The virtual machine system according to claim 1, wherein the virtualization software on the other server is virtualization software having a different type from the predetermined virtualization software. 9. . In a management server for managing the migration of the virtual machine in a virtual machine system comprising a plurality of servers capable of executing virtualization software for constructing at least one virtual machine,
Migrating the virtual machine operated by predetermined virtualization software on the predetermined server from the virtualization software on the plurality of other servers onto the virtualization software on another server optimal for migration;
Obtain information about hardware, the virtual machine system, and the virtual machine from a plurality of the servers,
Migration time information indicating the migration processing time, stop time information indicating the stop time of the virtual machine during the migration process, penalty information used for evaluating a difference in hardware configuration, the virtualization software, and the migration destination And migration information indicating a migration method based on the type of virtualization software of
A management server, wherein the virtualization software on another server optimal for migration is selected from the virtualization software on the plurality of other servers . Obtain information about the hardware, the virtual machine system, and the virtual machine from the plurality of servers, based on the CPU usage rate, memory usage, and time required for migration,
The management server according to claim 9, wherein the virtualization software on another server that is optimal for migration is selected from the virtualization software on the plurality of other servers . The management server according to claim 9 or 10, wherein an evaluation value indicating an index suitable as a migration destination is calculated based on information acquired from a plurality of the servers . 12. The management server according to claim 11, wherein the migration destination server is determined or not determined based on a relationship between a predetermined threshold and the evaluation value . After determining the migration destination server, determine whether live migration is possible based on the migration information,
If possible, perform the migration using live migration,
The image format of the virtual machine to be migrated is migrated to an image format that can be handled by virtualization software on the migration destination server when it is not possible. Management server . The management server according to any one of claims 9 to 13, wherein a file at the end of migration existing in the virtual machine to be migrated is applied to the image after migration . A virtual machine by a virtual machine system comprising a plurality of servers capable of executing virtualization software for constructing at least one virtual machine, and a management server for migrating the virtual machines onto the virtualization software on another server A migration method,
The virtualization software on the other server that is optimal for migration from the virtualization software on the plurality of other servers to the virtual machine operated by the predetermined virtualization software on the predetermined server. Have a transition step to move up,
In the transition step,
Obtain information about the hardware, the virtual machine system, and the virtual machine from the server,
Migration time information indicating the migration processing time, stop time information indicating the stop time of the virtual machine during the migration process, penalty information used for evaluating a difference in hardware configuration, the virtualization software, and the migration destination And migration information indicating a migration method based on the type of virtualization software of
A virtual machine migration method comprising: selecting the virtualization software on another server most suitable for migration from the virtualization software on a plurality of the other servers . A program executed on a management server that manages migration of the virtual machine in a virtual machine system including a plurality of servers capable of executing virtualization software for constructing at least one virtual machine,
In the management server,
Processing for migrating the virtual machine operated by predetermined virtualization software on the predetermined server from the virtualization software on the plurality of other servers onto the virtualization software on another server optimal for migration And execute
In the management server,
Obtain information about hardware, the virtual machine system, and the virtual machine from a plurality of the servers,
Migration time information indicating the migration processing time, stop time information indicating the stop time of the virtual machine during the migration process, penalty information used for evaluating a difference in hardware configuration, the virtualization software, and the migration destination And migration information indicating a migration method based on the type of virtualization software of
A program for executing a process of selecting the virtualization software on another server optimal for migration from the virtualization software on the plurality of other servers .

Images