JP5391152B2 - Server system and method for selecting virtual server migration method - Google Patents

Description

  The present invention relates to migration of virtual servers between physical servers.

  A technology related to server virtualization is known (for example, Patent Document 1).

JP 2008-176482 A

  Virtual servers can be migrated between physical servers. As a virtual server migration method, there are a method for performing migration with emphasis on reliability (hereinafter referred to as “reboot method”) and a method for performing migration with emphasis on migration speed (hereinafter referred to as “live method”).

  The virtual server migration method is selected by an operator (human). The conditions required for the migration destination physical server differ depending on the migration method. This must be taken into account by the operator. For this reason, it is difficult to appropriately select the migration method. If the migration method is inappropriate, virtual server migration may fail.

  Therefore, an object of the present invention is to increase the possibility that an appropriate migration method is selected.

  A virtual server migration device is provided. The virtual server migration apparatus is an apparatus for migrating a migration target virtual server among a plurality of virtual servers of a plurality of physical servers from a physical server having the virtual server to a physical server different from the physical server. The virtual server migration device is any one of a plurality of physical servers or a device connected to a plurality of physical servers.

  As a virtual server migration method, a first migration method that depends on a hardware relationship that is a relationship between a hardware configuration of a migration source physical server and a hardware configuration of a migration destination physical server; There is a second migration method that does not depend on the hardware relationship.

  The virtual server migration apparatus stores migration policy information representing a migration policy for each of one or more virtual servers. The virtual server migration apparatus includes two or more including (A) a first physical server (a physical server having a migration target virtual server) and a second physical server (a migration destination physical server of the migration target virtual server). (B) Among the first migration method and the second migration method, the hardware configuration of the first physical server and the hardware configuration of the second physical server are identified. The migration method is selected according to the hardware relationship and the migration policy corresponding to the virtual server to be migrated. The virtual server migration apparatus can migrate the migration target virtual server from the first physical server to the second physical server by the selected migration method.

  According to the present invention, it is possible to increase the possibility that an appropriate migration method is selected.

1 shows a configuration of a virtual server migration environment according to an embodiment of the present invention. 6 shows a processing procedure for virtual server migration method selection according to an embodiment of the present invention. An example of a virtual server migration policy setting screen concerning one embodiment of the present invention is shown. 4 shows an example of a virtual server migration policy table according to an embodiment of the present invention. The example of the virtual server transfer instruction | indication screen which concerns on one Embodiment of this invention is shown. 4 shows an example of a hardware specification table according to an embodiment of the present invention. The example of the hardware operation condition table which concerns on one Embodiment of this invention is shown. 6 shows a procedure for determining a virtual server migration method according to an embodiment of the present invention. The example of the detail of step (802) of FIG. 8 is shown. The example of the detail of step (803) of FIG. 8 is shown. The example of the detail of step (804) of FIG. 8 is shown.

  Hereinafter, an embodiment of the present invention will be described. The outline is as follows.

  A virtual server migration machine that performs virtual server migration stores migration policy information. The migration policy information is, for example, for each virtual server. It is possible to set in advance a migration policy corresponding to the reliability required for the system constructed in each virtual server, and when a human performs the migration work of a virtual server, the migration target virtual There is no need to determine whether the server is a system that requires high reliability.

  Further, the virtual server migration machine collects the hardware configuration of all physical servers and the hardware operation status of all physical servers when performing the virtual server migration work. The virtual server migration machine determines whether the hardware configuration (for example, specifications) of the migration source physical server (physical server having the migration target virtual server) and the hardware configuration of the migration destination candidate physical server are the same. Further, the virtual server migration machine grasps the hardware operation status of the migration source physical server and the hardware operation status of the migration destination candidate physical server. Therefore, when performing a migration operation of a virtual server, a human determines whether the hardware configuration of the migration source physical server is the same as the hardware configuration of the migration destination candidate physical server, It is not necessary to determine whether or not the required performance (required performance) is satisfied when the migration target virtual server is migrated to the server.

  Furthermore, in this embodiment, a new migration method is provided as a virtual server migration method. The new migration method is an intermediate migration method between the Live method and the Reboot method (hereinafter referred to as “Sleep method”). In the present embodiment, any one of the live method, the reboot method, and the sleep method can be selected. The characteristics of the three transition methods are as follows.

  <(1) Live system>.

The hypervisor (migration source hypervisor) of the migration source physical server receives a migration instruction specifying the migration method, the migration target virtual server and the migration destination physical server from the virtual server migration machine. When the migration source hypervisor receives the migration instruction, the following (1-1) to (1-3),
(1-1) Even if the migration target virtual server performs I / O (Input / Output) of data to a storage device (for example, a file system), the I / O is transferred regardless of the stage of the I / O. To stop,
(1-2) Move the migration target virtual server to the migration destination physical server without ending (without shutting down).
(1-3) Transfer information about the virtual server to be migrated to the hypervisor (migration destination hypervisor) of the migration destination physical server.
I do.

  According to the Live method, there is an advantage that the stop time of the guest OS (OS (Operating System executed on the migration target virtual server)) is short (for example, within several seconds). Therefore, it seems that the user terminal accessing the migration target virtual server almost stopped.

  However, the Live method has a drawback that migration is impossible unless the hardware configuration (for example, specifications) of the migration source physical server and the hardware configuration of the migration destination physical server are exactly the same. This is because the virtual server is migrated without being terminated (without being rebooted on the migration destination physical server), and therefore the environment in which the migration target virtual server operates (for example, information on the migration target virtual server is stored) If the migration source physical server and the migration destination physical server are different, the migration target virtual server must operate on the migration destination physical server in the same way as it was operated on the migration source physical server. It is because it is not possible. Specifically, for example, when information indicating the location of the file referred to by the migration target virtual server is stored in a certain location in the memory of the migration source physical server, if the hardware configuration is different, This is because information indicating the location of the file referred to by the migration target virtual server is not stored in the same location in the migration destination physical server.

  In addition, the Live method has a drawback in that a case where data recovery processing is required after migration may occur. This is because data writing may fail depending on the timing of data I / O of the migration target virtual server, and in this case, consistency cannot be achieved before and after migration.

  Therefore, the Live method is considered suitable for, for example, migrating a virtual server for a non-reliable system to a physical server having the same configuration as the migration source physical server.

  <(2) Reboot method>.

The migration source hypervisor receives a migration instruction designating the migration method, the migration target virtual server, and the migration destination physical server from the virtual server migration machine. When the migration source hypervisor receives the migration instruction, the following (2-1) to (2-3),
(2-1) Terminate (shut down) the virtual server to be migrated,
(2-2) Transfer information about the virtual server to be migrated to the migration destination hypervisor.
(2-3) Instruct the migration destination hypervisor to reboot,
I do. In response to the reboot instruction, the migration destination hypervisor reboots the migration target virtual server. As a result, the migration target virtual server operates on the migration destination physical server.

  The Reboot method has an advantage that there is no possibility of inconsistency in data. This is because if the virtual server to be migrated is performing data I / O to the storage device, after the I / O has been normally completed (for example, to the storage device remaining in the memory of the virtual server) This is because the guest OS is shut down after unwritten data is written to the storage device.

  Further, the Reboot method has an advantage that migration is possible even if there is a difference between the hardware configuration of the migration source physical server and the hardware configuration of the migration destination physical server. This is because the migration target virtual server is rebooted on the migration destination physical server, and at that time, the environment (for example, information) necessary for the operation of the virtual server is constructed according to the hardware configuration of the migration destination physical server. Because it will be.

  However, the Reboot method has a drawback that the guest OS has a long stop time (for example, around 10 minutes). This is because it takes time to shut down and reboot the migration target virtual server. When viewed from the user terminal that accesses the migration target virtual server, it feels that the system has been down for a long time.

  Therefore, it is considered that the Reboot method is suitable for, for example, systematically transferring a virtual server for a highly reliable system to a server having a different hardware configuration (for example, planned stop operation).

  <(3) Sleep method>.

The migration source hypervisor receives a migration instruction designating the migration method, the migration target virtual server, and the migration destination physical server from the virtual server migration machine. When the migration source hypervisor receives the migration instruction, the following (3-1) to (3-3),
(3-1) If the virtual server to be migrated performs I / O (Input / Output) of data to a storage device (for example, a file system), the I / O is delimited by the I / O (checkpoint) And let me stop,
(3-2) Move the migration target virtual server to the migration destination physical server without ending (without shutting down).
(3-3) Transfer information related to the virtual server to be migrated to the hypervisor of the migration destination physical server (migration destination hypervisor).
I do.

  The Sleep method has an advantage that there is no fear of inconsistency in data. This is because the I / O is canceled at that break.

  However, according to the Sleep method, there is a drawback that migration is impossible unless the hardware configuration of the migration source physical server and the hardware configuration of the migration destination physical server are exactly the same. This is because the virtual server is migrated without ending as described above.

  According to the Sleep method, when the hardware configuration of the migration destination physical server is the same as the hardware configuration of the migration source physical server, the virtual server can be migrated faster than the Reboot method and more safely than the Live method. is there. Therefore, the Sleep method is considered suitable for, for example, a case where a virtual server for a highly reliable system is migrated to another server at a timing other than planned stop operation.

  According to the present embodiment, a new transition method “Sleep method” is provided. The virtual server migration machine stores migration policy information for each virtual server. The migration server computer collects the hardware configuration and operation status of all physical servers, and based on the hardware configuration and operation status and migration policy information of the migration target virtual server, the migration method and migration destination physical server Select. Therefore, the migration target server can be migrated to an appropriate physical server by an appropriate migration method.

  Hereinafter, an embodiment of the present invention will be described in detail.

  In the following description, various types of information may be described using the expression “xxx table”, but the various types of information may be expressed using a data structure other than a table. In order to show that it does not depend on the data structure, the “xxx table” can be called “xxx information”.

  In the following description, the process may be described using “program” as a subject. However, a program is executed by a processor (for example, a CPU (Central Processing Unit)), so that a predetermined process is appropriately performed. Since the processing is performed using a storage resource (for example, a memory) and / or a communication interface device (for example, a communication port), the subject of processing may be a processor. The processing described with the program as the subject may be processing performed by the virtual server migration machine 13 described later. Further, the processor may include a hardware circuit that performs part or all of the processing performed by the processor. The computer program may be installed on each computer from a program source. The program source may be, for example, a program distribution server or a storage medium.

  FIG. 1 shows a configuration of a virtual server migration environment according to an embodiment of the present invention.

  The virtual server migration environment 1 includes a plurality of server blades and a virtual server migration machine 13. In FIG. 1, as a plurality of server blades, there are a server blade 2 in which a virtual server to be migrated exists, and server blades 3 and 4 that can be candidates for migration destinations of virtual servers.

  The server blades 2 to 4 will be described by taking the server blade 2 as an example. The configurations and functions of the server blades 3 and 4 may be the same as those of the server blade 2.

  The server blade 2 is an example of a physical server. The server blade 2 has hardware 9 that is a physical resource. Examples of the hardware 9 include one or more CPUs (Central Processing Units) 10, one or more memories 11, and one or more I / O devices (input / output devices) 12. The one or more I / O devices 12 may include, for example, a communication interface device and a physical storage device (for example, a hard disk drive).

  The hypervisor 5 is executed by at least one CPU 10. The hypervisor 5 can create a virtual server or delete a virtual server. In the example of FIG. 1, the hypervisor 5 manages a plurality of virtual servers 6, 7 and 8. A part of the hardware 9 (for example, the CPU 10 and the memory 11) is allocated to each virtual server.

  A system administrator inputs virtual server migration policy input data (hereinafter referred to as migration policy data) 25, which is data representing a virtual server migration policy, to the virtual server computer 13.

  The virtual server migration machine 13 is an apparatus (for example, a computer) connected to a plurality of server blades. For example, the virtual server migration machine 13 may be connected to a plurality of server blades via a communication network (for example, a LAN (Local Area Network)). The virtual server migration machine 13 may be housed in a chassis having a plurality of server blades, or may be located away from the chassis.

  The virtual server migration machine 13 includes one or more CPUs 14, one or more memories 15, and one or more I / O devices 16. The one or more I / O devices 16 may include, for example, a communication interface device (for example, an interface device connected to a plurality of server blades) and a physical storage device (for example, a hard disk drive).

  The memory 15 includes a virtual server migration method selection program (hereinafter referred to as selection program) 17, a virtual server migration policy table (hereinafter referred to as migration policy table) 18, a hardware specification table (hereinafter referred to as specification table) 19, and hardware operation status A table (hereinafter, operating status table) 20 is stored. The selection program 17 is executed by the CPU 14.

  The selection program 17 selects a virtual server migration method and a migration destination server blade. The selection program 17 includes a virtual server migration policy table control unit (hereinafter referred to as migration policy control unit) 21, a hardware specification table control unit (hereinafter referred to as specification control unit) 22, and a hardware operation status table control unit (hereinafter referred to as operation status control). And a virtual server migration destination / migration method determining unit (hereinafter referred to as a migration determining unit) 24.

  The migration policy control unit 21 creates a migration policy table 18 on the memory 15 based on the inputted migration policy data 25 and stores the table 18 in the I / O device 16. Further, the migration policy control unit 21 reads the migration policy table 18 in the I / O device 16 and stores the table 18 in the memory 15.

  The specification control unit 22 creates a specification table 19 representing the hardware in the plurality of server blades on the memory 15.

  The operating status control unit 23 creates an operating status table 20 representing the operating status of hardware in a plurality of server blades on the memory 15.

  The migration determination unit 24 determines the optimal migration destination server blade and migration method of the virtual server based on the migration policy table 18, the specification table 19, and the operation status table 20. The migration determination unit 24 migrates the migration target virtual server to the determined optimal migration destination server blade by the determined optimal migration method.

  FIG. 2 shows a processing procedure for selecting a virtual server migration method according to this embodiment. The flow of virtual server migration will be described with reference to this figure.

  <Input migration policy>.

  In step (200), the system administrator activates the virtual server migration machine 13.

  In step (201), the virtual server migration machine 13 (migration policy control unit 21) waits for input of the migration policy data 25 from the system administrator.

  In step (202), the system administrator inputs the migration policy data 25.

  In step (203), the virtual server migration machine 13 (migration policy control unit 21) creates the migration policy table 18 based on the inputted migration policy data 25. The virtual server migration machine 13 (migration policy control unit 21) stores the created migration policy table 18 in the I / O device 16.

  In step (205), the system administrator ends the virtual server migration machine 13.

  <Virtual server migration>.

  In step (210), the system administrator activates the virtual server migration machine 13.

  In step (211), the virtual server migration machine 13 waits for a virtual server migration instruction from the system administrator.

  In step (212), the system administrator instructs the virtual server migration machine 13 to migrate the virtual server.

  In step (213), the virtual server migration machine 13 (migration policy control unit 13) reads the migration policy table 18 from the I / O device 16 and stores the table 18 in the memory 15.

  In step (220), the virtual server migration machine 13 (specification control unit 22) requests a plurality of server blades to acquire information indicating the hardware specifications of the server blade (hereinafter referred to as HW specification information) (hereinafter referred to as specification acquisition). Request).

  In step (221), the server blade that has received the specification acquisition request collects the hardware specifications of the server blade. For example, the memory of the server blade stores information (HW specification information) indicating the hardware specification of the server blade, and the HW specification information is acquired.

  In step (222), the server blade responds (transmits) the acquired HW specification information to the virtual server migration machine 13.

  In step (223), the virtual server migration machine 13 (specification control unit 22) creates the specification table 19 on the memory 15 based on the HW specification information from each server blade.

  In step (230), the virtual server migration machine 13 (operation status control unit 23) requests a plurality of server blades to acquire information indicating the hardware operation status of the server blade (hereinafter, HW operation status information) (hereinafter, HW operation status information). , Operational status acquisition request).

  In step (231), the server blade (for example, hypervisor) that has received the operation status acquisition request collects the hardware operation status of the server blade. For example, for example, the memory of the server blade stores information (HW operating status information) indicating the hardware operating status (for example, CPU usage rate) of the server blade, and the HW operating status information is acquired.

  In step (232), the server blade (eg, hypervisor) responds (transmits) the acquired HW operation status information to the virtual server migration machine 13.

  In step (233), the virtual server migration machine 13 (operation status control unit 23) creates the operation status table 20 on the memory 15 based on the HW operation status information from each server blade.

  In step (240), the virtual server migration machine 13 (migration decision unit 24) determines the migration destination server blade and migration method of the migration target virtual server based on the migration policy table 18, the specification table 19 and the operation status table 20. To decide.

  In step (241), the virtual server migration machine 13 (migration determination unit 24) transfers to the server blade (migration source server blade) having the migration target virtual server using the migration method determined for the determined migration destination server blade. A request for virtual server migration (virtual server migration request) is issued.

  In step (242), in response to the virtual server migration request, the migration source server blade migrates the migration target virtual server to the migration destination server blade identified from the request by the migration method identified from the request. To do.

  In step (243), the migration source server blade reports the completion of the virtual server migration to the virtual server migration machine 13 after the migration of the virtual server is completed.

  FIG. 3 shows an example of a virtual server migration policy setting screen (hereinafter referred to as a migration policy setting screen) according to the present embodiment.

  The migration policy setting screen 300 is a GUI (Graphical User Interface) that accepts an input of a migration policy from the system administrator. The data according to the input to this screen 300 is the migration policy data 25 described above.

  The migration policy setting screen 300 includes, for example, migration source server information 310, migration destination server information 320, a setting button 330, and a completion button 340. After setting information in the migration source server information 310 and the migration destination server information 320, when the setting button 330 is pressed, the migration policy table 18 is created based on the set information (migration policy data 25). When the completion button 340 is pressed after the setting is completed, the migration policy setting screen 300 ends.

  The migration source server information 310 is an area for setting information regarding the migration source server blade and the migration target virtual server. The migration source server information 310 includes, for example, a server blade number 311, a virtual server number 312, and a virtual server migration policy 313.

  The server blade number 311 is an area in which the number of the migration source server blade is designated. In that area, for example, a number desired by the system administrator is selected from the existing server blade numbers using a pull-down menu.

  The virtual server number 312 is an area in which the number of the migration target virtual server is specified. In that area, for example, a number desired by the system administrator is selected from the existing virtual server numbers in a pull-down menu.

  The virtual server migration policy 313 is designated by the migration policy for migrating the migration target virtual server (the virtual server designated by the virtual server number 312) in the migration source server blade (the server blade designated by the server blade number 311). It is an area to be. In this area, a policy desired by the system administrator is selected from among “Safety”, “Speed”, and “Not set” with a radio button, for example. “Safety” means a transition method that emphasizes safety over transition speed. “Speed” means a transition method in which the transition speed is more important than safety. “Not set” means that the virtual server is not migrated. The system administrator can select either “Speed” or “Safety” as a policy depending on whether to give priority to the transition speed or safety.

  The migration destination server information 320 is an area in which information regarding the migration destination server blade is set. The migration destination server information 320 includes, for example, a server blade number 321.

  The server blade number 321 is an area in which the server blade number of the migration destination server blade is designated. In this area, for example, a number desired by the system administrator is selected from the existing server blade numbers using a pull-down menu. Options include “unset” and “automatic selection” in addition to the server blade number. “Not set” means that the migration destination server blade is not selected. “Automatic selection” means that any server blade of the virtual server migration machine 13 is selected as the migration destination.

  FIG. 4 shows an example of the migration policy table 18 according to the present embodiment.

The migration policy table 18 exists for the maximum number of server blades. The table 18 includes, for example, the following information for each virtual server included in the server blade corresponding to the table 18:
(*) A virtual server number 401 which is a virtual server number,
(*) Migration policy 402, which is information indicating the migration policy of the virtual server,
(*) Migration destination server blade number 403 that is the migration destination server blade number of the virtual server,
Have Information 401 to 403 is information input to the migration policy setting screen 300, respectively.

  FIG. 5 shows an example of a virtual server migration instruction screen (hereinafter, migration instruction screen) according to the present embodiment.

  The migration instruction screen 500 is a GUI that receives the migration of the virtual server from the system administrator. The screen 500 includes, for example, migration source server information 510, an execution button 520, and a cancel button 530. When the execution button 520 is pressed after the information is set in the migration source server information 510, the migration of the virtual server is started. When the cancel button 530 is pressed, the migration instruction screen 500 ends.

  The migration source virtual server information 510 is an area in which information regarding the migration source server blade and the migration target virtual server is set. The information 510 includes, for example, a server blade number 511 and a virtual server number 512.

  The server blade number 511 is an area in which the number of the migration source server blade is designated. In this area, for example, a number desired by the system administrator is selected from the existing server blade numbers using a pull-down menu.

  The virtual server number 512 is an area in which the number of the migration target virtual server is designated. In this area, for example, a number desired by the system administrator is selected from the existing virtual server numbers using a pull-down menu.

  FIG. 6 shows an example of the specification table 19 according to the present embodiment.

  There are as many specification tables 19 as there are server blades. The table 19 includes, for example, information regarding the hardware specifications of the server blade corresponding to the table 19 (hereinafter referred to as “target server blade” in the description of FIG. 6). Specifically, for example, the table 19 includes CPU information 610, mounted memory information 620, and mounted I / O information 630.

  The CPU information 610 is information related to the CPU specifications of the target server blade. The CPU information 610 includes, for example, a CPU type (information indicating the CPU type) 611, a chip set (information indicating the CPU operating frequency) 612, a core number (information indicating the number of CPU cores) 613, and a socket number (CPU socket). Information 614).

  The on-board memory information 620 is information regarding the specifications of the memory included in the target server blade. The mounted memory information 620 includes, for example, a memory type 621 (information indicating the type of memory mounted in the memory slot) for each memory slot.

  The mounted I / O information 630 is information related to the specifications of the I / O device that the target server blade has. The mounted I / O information 630 is inserted into, for example, an onboard I / O type (information indicating the type of an onboard I / O device) 631 and a slot I / O type (I / O slot (expansion slot)). Information 632) indicating the type of the I / O device that is present.

  FIG. 7 shows an example of the operation status table 20 according to the present embodiment.

  There are as many operating status tables 20 as there are server blades. The table 20 includes, for example, information related to the operating status of hardware included in a server blade corresponding to the table 20 (hereinafter referred to as “target server blade” in the description of FIG. 7). Specifically, for example, the table 20 includes a CPU operating status 710, a memory operating status 720, and a mounted I / O operating status 730.

  The CPU operating status 710 is information representing the operating status of the CPU included in the target server blade. The CPU operating status 710 includes, for example, a CPU operating rate per unit time (information indicating the CPU operating rate per unit time) 711.

  The memory operating status 720 is information indicating the operating status of the memory included in the target server blade. The memory operating status 720 includes, for example, a used memory capacity (information indicating a memory capacity used by the virtual server) 721 for each virtual server included in the target server blade.

  The installed I / O operation status 730 is information indicating the operation status of the I / O device included in the target server blade. The mounted I / O operation status 730 includes, for example, used mounted I / O information (information including the number of the virtual server using the I / O device) 731 for each I / O device included in the target server blade. Have.

  FIG. 8 shows a processing procedure for determining a virtual server migration method according to this embodiment. In the following description, the server blade with the number i may be described as “server blade #i”.

  The process shown in this figure is the details of step (240) in FIG. 2, and is the procedure of the process performed by the migration determining unit 24 of the virtual server migration machine 13. The migration target virtual server and the migration source server blade are designated by the migration instruction (migration instruction issued using the screen 500 in FIG. 5) input in step (212) in FIG.

  In step (800), the migration determination unit 24 determines whether or not the migration policy for the migration target virtual server is “unset” based on the migration policy table 18 corresponding to the migration source server blade. When the migration policy is “not set”, step (811) is performed, and when the migration policy is other than “not set”, step (801) is performed.

  In step (801), the migration determination unit 24 determines whether or not the migration destination server blade of the migration target virtual server is “unset” based on the migration policy table 18 corresponding to the migration source server blade. When the migration destination server blade is “not set”, step (811) is performed, and when the migration destination server blade is other than “not set”, step (802) is performed.

  In step (802), the migration determining unit 24 creates a designated migration destination blade list. The designated migration destination blade list has To [i] for the maximum number of server blades (i is the number of the server blade as described above). If server blade #i is the migration destination server blade (or candidate), “OK” is set for To [i], and if server blade #i cannot be the migration destination server blade, about To [i] “NG” is set. Server blades that cannot be the migration destination server blade are migration source server blades and server blades that are not designated as migration destinations. Therefore, when “automatic selection” is designated as the migration destination server blade, “OK” is set for To [i] of all server blades other than the migration source server blade. That is, the designated migration destination server blade list is a list of migration destination server blades (and their candidates).

  In step (803), the migration determining unit 24 creates an identical specification hardware list. The same specification hardware list has Spec [i] corresponding to the maximum number of server blades. If the hardware specification of server blade #i and the hardware specification of the migration source server blade are the same, “OK” is set for Spec [i], and if the hardware specifications are not identical, Spec [ “NG” is set for i]. Whether the hardware specifications are the same can be determined by comparing the specification table 19 for the migration source server blade with the specification table 19 for the server blade #i. For example, if the CPU, memory, and I / O device with the same specifications are in both server blades (migration source server blade and server blade #i) even if the slots are different, “OK” is displayed for Spec [i]. Is set.

  In step (804), the migration determining unit 24 creates an operable blade list. The operable blade list has Able [i] corresponding to the maximum number of server blades. When the server blade #i is operable even if the migration target virtual server is migrated to the server blade #i (for example, even if the migration target virtual server is migrated, the CPU usage rate does not exceed a predetermined threshold) If presumed), “OK” is set for Able [i]. On the other hand, when the migration target virtual server is migrated to the server blade #i, the server blade #i becomes inoperable (for example, when the migration target virtual server is migrated, the CPU usage rate exceeds a predetermined threshold). NG ”is set for Able [i]. Whether or not the migration target virtual server is migrated to server blade #i corresponds to both server blades (migration source server blade and server blade #i). The determination can be made based on the operation status table 20.

  In step (805), the migration determining unit 24 sets a server blade for which “OK” is set for all of To [i], Spec [i], and Able [i] (that is, a migration destination or a migration destination candidate). And a server blade that has the same hardware specifications as the migration source server blade and can operate even if the migration target virtual server is migrated). If there is a corresponding server blade, step (806) is performed, and if there is no corresponding server blade, step (807) is performed.

  In step (806), the migration determining unit 24 determines whether or not the migration policy of the migration target virtual server is “Speed” based on the migration policy table 18 corresponding to the migration source server blade. When the transition policy is “Speed”, step (808) is performed, and when the transition policy is other than “Speed”, step (809) is performed.

  In step (807), the migration determining unit 24 sets the server blade for which “OK” is set for To [i] and Spec [i] (that is, the migration destination or the migration destination candidate, and the hardware specification is the migration source). Search for a server blade that is different from the server blade but can be operated even if the migration target virtual server is migrated. If there is a corresponding server blade, step (810) is performed, and if there is no corresponding server blade, step (811) is performed.

  In step (808), the migration determining unit 24 determines the migration method to be the Live method.

  In step (809), the transition determining unit 24 determines the transition method to be the sleep method.

  In step (810), the migration determining unit 24 determines the migration method as the Reboot method.

  In step (811), the migration determining unit 24 fails the virtual server migration.

  According to the flow shown in FIG. 8, if the hardware configuration is the same at the migration source and the migration destination, either the Live method or the Sleep method can be selected. Which migration method is used depends on whether the migration speed or safety is important. The Sleep method has higher safety but slower transition speed than the Live method. Therefore, if the transition policy is “Speed”, the Live method is selected, and if the transition policy is “Safety”, the Sleep method is selected.

  Further, according to the flow shown in FIG. 8, a server blade having the same hardware configuration as the migration source is preferentially set as a migration destination over a server blade having a hardware configuration that is not identical to the migration source. As a result, it is possible to reduce the possibility that the Reboot method, which has substantially the same safety as the Sleep method but has a slower transition speed than the Sleep method, is selected.

  Although the Reboot method may be selected even if the hardware configuration is the same, as described above, the safety of the Reboot method and the safety of the Sleep method are substantially the same. If the hardware configuration is the same, the Reboot method is not selected.

  FIG. 9 shows an example of details of step (802) of FIG.

  In step (900), the migration determination unit 24 sets 0 to i (migration destination blade number).

  In step (901), the migration determining unit 24 determines whether i is less than the maximum number of blades. If i is less than the maximum number of blades, step (902) is performed, and if i is greater than or equal to the maximum number of blades, the processing flow shown in FIG. 9 ends.

  In step (902), the migration determining unit 24 determines whether i is the same as the migration source server blade number (hereinafter, migration source blade number). If i is the same as the migration source blade number, step (908) is performed, and if i is a number other than the migration source blade number, step (908) is performed.

  In step (903), the migration determining unit 24 determines whether or not the migration destination server number 403 (see FIG. 4) corresponding to the migration target virtual server is “automatic selection”. When the migration destination server number 403 is “automatic selection”, step (907) is performed, and when the migration destination server number 403 is other than “automatic selection”, step (904) is performed.

  In step (904), the migration determining unit 24 determines whether i is the same as the number (migration destination designation blade number) represented by the migration destination server number 403 corresponding to the migration target virtual server. If i is the same as the migration destination designation blade number, step (905) is performed, and if i is a number other than the migration destination blade number, step (906) is performed.

  In step (905), the migration determining unit 24 sets “OK” for To [i].

  In step (906), the migration determining unit 24 sets “NG” for To [i].

  In step (907), the migration determining unit 24 sets “OK” for To [i].

  In step (908), the migration determining unit 24 sets “NG” for To [i].

  In step (909), the migration determining unit 24 adds 1 to i. Step (901) is performed for i after the update.

  FIG. 10 shows an example of details of step (803) of FIG.

  In step (1000), the migration determining unit 24 sets 0 to i, and sets the migration source blade number to the parameter j.

  In step (1001), the migration determining unit 24 determines whether i is less than the maximum number of blades. If i is less than the maximum number of blades, step (1002) is performed, and if i is greater than or equal to the maximum number of blades, the processing flow shown in FIG.

  In step (1002), the migration determining unit 24 compares the CPU information 610 of the blade #i (see FIG. 6) with the CPU information 610 of the blade #j. When the CPU information 610 of blade #i matches the CPU information 610 of blade #j (when the CPU type, chipset, number of cores, number of sockets, etc. are the same), step (1003) is performed. On the other hand, if at least part of CPU information 610 of blade #i does not match CPU information 610 of blade #j (if any of CPU type, chipset, number of cores, number of sockets, etc. is not the same), step (1006) is performed.

  In step (1003), the migration determining unit 24 compares the mounted memory information 620 of blade #i with the mounted memory information 620 of blade #j. When the mounted memory information 620 of the blade #i matches the mounted memory information 620 of the blade #j (for example, when the combination of the slot number and the memory type all match each other), the step (1004) is performed. On the other hand, if at least part of the mounted memory information 620 of blade #i does not match the mounted memory information 620 of blade #j, step (1006) is performed.

  In step (1004), the migration determining unit 24 compares the mounted I / O information 630 of blade #i with the mounted I / O information of blade #j. When the mounted I / O information 630 of blade #i matches the mounted I / O information 630 of blade #j (for example, when the combination of slot number and I / O device type all match each other), the step ( 1005) is performed. On the other hand, if at least part of the mounted I / O information 630 for blade #i does not match the mounted I / O information 630 for blade #j, step (1006) is performed.

  In step (1005), the migration determining unit 24 sets “OK” for Spec [i].

  In step (1006), the migration determining unit 24 sets “NG” for Spec [i].

  In step (1007), the migration determining unit 24 adds 1 to i. Step (1001) is performed for i after the update.

  FIG. 11 shows an example of details of step (804) of FIG.

  In step (1100), the migration determination unit 24 sets i to 0, sets the migration source blade number to the parameter j, and sets the migration target virtual server number to the parameter k. Hereinafter, the migration target virtual server is referred to as “virtual server #k”.

  In step (1101), the migration determining unit 24 determines whether i is less than the maximum number of blades. If i is less than the maximum number of blades, step (1102) is performed, and if i is greater than or equal to the maximum number of blades, the processing flow shown in FIG. 11 ends.

  In step (1102), the migration determining unit 24 compares the CPU operating status 710 of blade #i with the CPU operating status 710 of blade #j. When the CPU operating rate of blade #i is lower than the CPU operating rate of blade #j (for example, the sum of the CPU operating rate of blade #i and the CPU operating rate of virtual server #k is the CPU operating rate of blade #j Step (1103) is performed when the difference from the CPU operating rate by the virtual server #k is smaller. On the other hand, when the CPU operating rate of blade #i is equal to or higher than the CPU operating rate of blade #j, step (1106) is performed.

  In step (1103), the migration determining unit 24 compares the memory operating status 720 of blade #i with the memory operating status 720 of blade #j. When the used memory capacity of blade #i is smaller than the used memory capacity of blade #j (for example, the sum of the used memory capacity of blade #i and the used memory capacity of virtual server #k is equal to the used memory capacity of blade #j) If the difference from the used memory capacity by the virtual server #k is smaller), step (1104) is performed. On the other hand, if the used memory capacity of blade #i is greater than or equal to the used memory capacity of blade #j, step (1106) is performed.

  In step (1104), the migration determining unit 24 compares the mounted I / O operating status 730 of blade #i with the mounted I / O operating status 730 of blade #j. If the value corresponding to the used I / O slot number of the virtual server k (value in the mounted I / O operating status 730 of blade #i) is “unused”, step (1105) is performed. In other words, any I / O device (I / O device in blade #i) to which the same number as the I / O device number used for the virtual server in blade #j is assigned in blade #i If it is not used for the virtual server, step (1105) is performed. On the other hand, if the virtual server number corresponding to the used I / O slot number of the virtual server k (virtual server number in the mounted I / O operating status 730 of blade #i) is not “unused”, step (1106) is performed. Is called.

  In step (1105), the migration determining unit 24 sets “OK” for Able [i].

  In step (1106), the migration determining unit 24 sets “NG” for Able [i].

  Step (1107) The migration determining unit 24 adds 1 to i. Step (1101) is performed for i after the update.

  The embodiment of the present invention has been described above, but this is an example for explaining the present invention, and the scope of the present invention is not limited to this embodiment. The present invention can be implemented in various other forms.

  For example, any server blade of a plurality of server blades may function as a virtual server migration machine. In this case, a device other than a plurality of server blades may not be a virtual server migration machine.

  2, 3, 4 ... Server blade 6, 7, 8 ... Virtual server 13 ... Virtual server migration machine

Claims (17)

A plurality of physical servers having hypervisors running a plurality of virtual servers;
A virtual server migration apparatus that migrates a virtual server to be migrated from the plurality of virtual servers from a physical server having the virtual server to a physical server different from the physical server;
The virtual server migration device is one of the plurality of physical servers, or a device connected to the plurality of physical servers, and represents a migration policy for each of one or more virtual servers. Remember information,
As a virtual server migration method, a first migration method that depends on a hardware relationship that is a relationship between a hardware configuration of a migration source physical server and a hardware configuration of a migration destination physical server; There is a second migration method that does not depend on the hardware relationship,
The virtual server migration device
(A) Two or more physical servers including a first physical server that is a physical server having the migration target virtual server and a second physical server that is a migration destination physical server of the migration target virtual server Identify each hardware configuration,
(B) Of the first migration method and the second migration method, the hardware relationship between the hardware configuration of the first physical server and the hardware configuration of the second physical server, and the migration policy information Representing, selecting a migration method according to a target migration policy that is a migration policy corresponding to the virtual server to be migrated,
Server system. The server system according to claim 1,
In the second migration method, the virtual server is terminated on the migration source physical server, the virtual server is migrated from the migration source physical server to the migration destination physical server, and the virtual server is booted on the migration destination physical server. Method,
As the first migration method, there are a first first migration method and a second first migration method,
In the first first migration method, I / O (Input / Output) performed by a virtual server is canceled even if it is not a delimiter, and the migration-destination physical server to the migration-destination physical server without ending the virtual server Is a method of migrating virtual servers to
The second first migration method is a method in which I / O performed by a virtual server is canceled at the break, and the virtual server is migrated from the migration source physical server to the migration destination physical server without terminating the virtual server. Yes,
In the virtual server migration device, in (B),
(B1) This means that the hardware configuration of the first physical server and the hardware configuration of the second physical server are the same, and the target migration policy prioritizes the migration speed over safety. If it is a policy, select the first first transition method,
(B2) This means that the hardware configuration of the first physical server is the same as the hardware configuration of the second physical server, and that the target migration policy prioritizes safety over migration speed. If it is a policy, select the second first transition method,
(B3) If the hardware configuration of the first physical server and the hardware configuration of the second physical server are not the same, the second migration method is selected regardless of the target migration policy.
Server system. The server system according to claim 2,
The virtual server migration device
(C) identifying the hardware operating status of each of two or more physical servers including the first and second physical servers;
The second physical server is a physical server having a hardware operating status lower than the hardware operating status of the first physical server.
Server system. The server system according to claim 3,
The virtual server migration device
(D) selecting a physical server as the second physical server from physical servers other than the first physical server;
In the virtual server migration device, in (D), the hardware configuration is the same as the hardware configuration of the first physical server, rather than the physical server having a hardware configuration different from the hardware configuration of the first physical server. Are preferentially selected as the second physical server,
Server system. The server system according to claim 4,
The migration policy information is information having a migration policy designated by an administrator.
Server system. The server system according to claim 1,
As the first migration method, there are a first first migration method and a second first migration method,
The first first transition method is a transition method whose safety is lower than that of the second transition method and whose transition speed is faster than that of the second first transition method.
The second first transition mode is slower manner than the safety is the same as the second migration method migration speed is faster than the second transition method first first transition mode,
Server system. The server system according to claim 6, wherein
In the virtual server migration device, in (B),
(B1) This means that the hardware configuration of the first physical server and the hardware configuration of the second physical server are the same, and the target migration policy prioritizes the migration speed over safety. If it is a policy, select the first first transition method,
(B2) This means that the hardware configuration of the first physical server is the same as the hardware configuration of the second physical server, and that the target migration policy prioritizes safety over migration speed. If it is a policy, the second first transition method is selected.
Server system. The server system according to claim 7,
The virtual server migration device
(D) selecting a physical server as the second physical server from physical servers other than the first physical server;
In the virtual server migration device, in (D), the hardware configuration is the same as the hardware configuration of the first physical server, rather than the physical server having a hardware configuration different from the hardware configuration of the first physical server. Is preferentially selected as the second physical server,
If the hardware configuration of the first physical server and the hardware configuration of the second physical server are not the same in (B), the virtual server migration device, regardless of the target migration policy, Select the second migration method,
Server system. The server system according to claim 1,
The virtual server migration device
(C) identifying the hardware operating status of each of two or more physical servers including the first and second physical servers;
The second physical server is a physical server having a hardware operating status lower than the hardware operating status of the first physical server.
Server system. A method of migrating a virtual server to be migrated from a plurality of virtual servers executed by a hypervisor of a plurality of physical servers from a physical server having the virtual server to a physical server different from the physical server,
(X) storing migration policy information representing a migration policy for each of the one or more virtual servers;
(A) Two or more physical servers including a first physical server that is a physical server having the migration target virtual server and a second physical server that is a migration destination physical server of the migration target virtual server Identify each hardware configuration,
(B) Of the first migration method and the second migration method, the hardware relationship between the hardware configuration of the first physical server and the hardware configuration of the second physical server, and the migration policy information Representing the migration method according to the migration policy corresponding to the migration target virtual server,
As a virtual server migration method, a first migration method that depends on a hardware relationship that is a relationship between a hardware configuration of a migration source physical server and a hardware configuration of a migration destination physical server; There is a second migration method that does not depend on the hardware relationship.
Virtual server migration method. The virtual server migration method according to claim 10,
In the second migration method, the virtual server is terminated on the migration source physical server, the virtual server is migrated from the migration source physical server to the migration destination physical server, and the virtual server is booted on the migration destination physical server. Method,
As the first migration method, there are a first first migration method and a second first migration method,
In the first first migration method, I / O (Input / Output) performed by a virtual server is canceled even if it is not a delimiter, and the migration-destination physical server to the migration-destination physical server without ending the virtual server Is a method of migrating virtual servers to
The second first migration method is a method in which I / O performed by a virtual server is canceled at the break, and the virtual server is migrated from the migration source physical server to the migration destination physical server without terminating the virtual server. Yes,
In (B) above,
(B1) This means that the hardware configuration of the first physical server and the hardware configuration of the second physical server are the same, and the target migration policy prioritizes the migration speed over safety. If it is a policy, select the first first transition method,
(B2) This means that the hardware configuration of the first physical server is the same as the hardware configuration of the second physical server, and that the target migration policy prioritizes safety over migration speed. If it is a policy, select the second first transition method,
(B3) If the hardware configuration of the first physical server and the hardware configuration of the second physical server are not the same, the second migration method is selected regardless of the target migration policy.
Virtual server migration method. The virtual server migration method according to claim 11 ,
(C) identifying the hardware operating status of each of two or more physical servers including the first and second physical servers;
The second physical server is a physical server having a hardware operating status lower than the hardware operating status of the first physical server.
Virtual server migration method. The virtual server migration method according to claim 12,
(D) selecting a physical server as the second physical server from physical servers other than the first physical server;
In (D), a physical server having the same hardware configuration as the hardware configuration of the first physical server is given priority over a physical server having a hardware configuration different from the hardware configuration of the first physical server. To select as the second physical server,
Virtual server migration method. The virtual server migration method according to claim 13,
The migration policy information is information having a migration policy designated by an administrator.
Virtual server migration method. The virtual server migration method according to claim 10,
As the first migration method, there are a first first migration method and a second first migration method,
The first first transition method is a transition method whose safety is lower than that of the second transition method and whose transition speed is faster than that of the second first transition method.
The second first transition mode is slower manner than the safety is the same as the second migration method migration speed is faster than the second transition method first first transition mode,
Virtual server migration method. A computer program that causes a migration target virtual server among a plurality of virtual servers executed by a hypervisor of a plurality of physical servers to be executed by a device that migrates from the physical server having the virtual server to a physical server different from the physical server Because
(A) Two or more physical servers including a first physical server that is a physical server having the migration target virtual server and a second physical server that is a migration destination physical server of the migration target virtual server Identify each hardware configuration,
(B) Of the first migration method and the second migration method, the hardware relationship between the hardware configuration of the first physical server and the hardware configuration of the second physical server, and stored in the storage resource Selecting a migration method in accordance with a migration policy corresponding to the migration target virtual server represented by the migration policy information being migrated,
Causing the device to
As a virtual server migration method, a first migration method that depends on a hardware relationship that is a relationship between a hardware configuration of a migration source physical server and a hardware configuration of a migration destination physical server; There is a second migration method that does not depend on the hardware relationship.
Computer program. A virtual server migration apparatus that migrates a virtual server to be migrated from a plurality of virtual servers executed by a hypervisor of a plurality of physical servers from a physical server having the virtual server to a physical server different from the physical server. And
Storage resources,
A processor connected to the storage resource,
The storage resource stores migration policy information representing a migration policy for each of the one or more virtual servers;
As a virtual server migration method, a first migration method that depends on a hardware relationship that is a relationship between a hardware configuration of a migration source physical server and a hardware configuration of a migration destination physical server; There is a second migration method that does not depend on the hardware relationship,
The processor is
(A) Two or more physical servers including a first physical server that is a physical server having the migration target virtual server and a second physical server that is a migration destination physical server of the migration target virtual server Identify each hardware configuration,
(B) Of the first migration method and the second migration method, the hardware relationship between the hardware configuration of the first physical server and the hardware configuration of the second physical server, and the migration policy information Representing, selecting a migration method according to a target migration policy that is a migration policy corresponding to the virtual server to be migrated,
Virtual server migration device.

Images