JP2011039740A - Server management system, server management method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a server management system for suppressing costs, and for preventing service level agreement (SLA) violation. <P>SOLUTION: A server management system 1 automatically migrates a system that is likely to cause SLA(Service Level Agreement) violation among a plurality of systems configured of virtual machines, to a highly available environment based on information on a service level stored in a service level management table. Specifically, a cluster is configured by adding a virtual machine to physical servers 10 and 20 when the system can be redundantly configured, and the system is migrated to an FT server 30 if not so. After the end of the calculation unit period of the SLA, the system migrated to the FT server 30 or the redundantly configured system is restored to the original single configuration system. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a server management system, a server management method, and a program. More particularly, the present invention relates to a server management system, a server management method, and a program that reduce costs and avoid service level agreement (SLA) violations.

  Many companies outsource to a service provider called a data center in order to reduce maintenance management costs for the construction and maintenance of a hosting environment such as an online system on a network.

  Usually, in such a data center, there is a service quality contract with a customer. This is generally called service level agreement (SLA), and the side providing the data center needs to eliminate system troubles and maintain the service level according to the SLA.

  2. Description of the Related Art Conventionally, it is known to determine resources to be allocated to each user while confirming the operation status of a computer based on SLA (see, for example, Patent Documents 1 and 2).

JP-A-2005-174201 JP 2002-024192 A

  However, the conventional SLA has the following problems. In other words, if a redundant configuration is adopted in order to increase availability, there is a problem that the installation cost and management cost of equipment increase. Another problem is that once a failure occurs in a system in which a failure has occurred, the service level is significantly reduced and an SLA violation is likely to occur.

  In order to prevent such an increase in cost, it is necessary to reduce costs by operating in an environment that is not redundant as much as possible. In order to prevent SLA violations, it is necessary to take measures to improve availability so that a failure does not recur in a system having a failure history.

  The present invention has been made to solve the above-described problems, and provides a server management system, a server management method, and a program that reduce costs and avoid service level agreement (SLA) violations.

  In order to achieve the above object, in the server management system according to the first aspect of the present invention, a predetermined state value in a system configured by virtual machines on a physical server is determined based on a service level agreement (SLA). SLA violation determining means for determining whether or not the system has a risk of violating the SLA by determining whether or not the violation value is exceeded, and a risk of violating the SLA by the SLA violation determining means A redundant configuration determining means for determining whether or not the system determined to be redundant can be configured redundantly, and determining that the system can be configured redundantly by the redundant configuration determining means If the virtual machines constituting the system are arranged on a plurality of physical servers to make a redundant configuration, the system is made redundant. If it is determined that it is is impossible, comprises a virtual machine migration unit for migrating the virtual machine to configure the system on a given FT (Fault Tolerant) server from on the physical server, the.

  The server management system includes: an SLA re-violation determining unit that determines whether the redundant system configured by the virtual machine migration unit has a risk of violating the SLA again; and the SLA re-violation determining unit When it is determined that there is no possibility that the system again violates the SLA, the virtual machine on any physical server of the plurality of physical servers on which the virtual machines constituting the system are arranged Virtual machine reduction means for reducing the number of virtual machines.

  The server management system updates the multiplicity by adding 1 each time the system is made redundant by the virtual machine migration unit, while the virtual machine reduction unit reduces the virtual machine. The multiplicity update means for subtracting and updating the multiplicity by 1 for each time, the redundancy configuration determining means, wherein the multiplicity updated by the multiplicity update means is determined by the SLA violation determining means. If it is less than the maximum multiplicity of the system determined to be in danger of violating the SLA, it may be determined that the system can be configured redundantly.

  Furthermore, in order to achieve the above object, the server management method according to the second aspect of the present invention provides that a predetermined state value in a system configured by virtual machines on a physical server is based on a service level agreement (SLA). The SLA violation determining step for determining whether the system is at risk of violating the SLA by determining whether or not the specified violation value is exceeded, and the SLA violation determining step violates the SLA. It is possible to make the system redundant by the redundant configuration determining step for determining whether the system determined to be at risk of being redundant can be configured redundantly and the redundant configuration determining step. If it is determined that the virtual machine constituting the system is arranged on a plurality of physical servers to make a redundant configuration, A virtual machine migration step of migrating the virtual machine constituting the system from the physical server to a predetermined FT (Fault Tolerant) server when it is determined that the system cannot be redundantly configured. .

  In order to achieve the above object, a program according to the third aspect of the present invention is based on a service level agreement (SLA) in which a predetermined state value in a system constituted by a virtual machine on a physical server is stored in a computer. The SLA violation determination procedure for determining whether or not there is a risk of the system violating the SLA by determining whether or not the violation value exceeds the predetermined violation value, and the SLA violation determination procedure It is possible to make a redundant configuration of the system by a redundant configuration determination procedure for determining whether it is possible to make a redundant configuration of a system determined to be in danger of violating, and the redundant configuration determination procedure. When it is determined that the virtual machine is present, the virtual machine constituting the system is arranged on the plurality of physical servers to make a redundant configuration, while the system A virtual machine migration procedure for migrating the virtual machine constituting the system from the physical server to a predetermined FT (Fault Tolerant) server .

  According to the present invention, it is possible to provide a server management system, a server management method, and a program that can suppress costs and avoid service level agreement (SLA) violations.

It is a block diagram which shows the structural example of a server management system. It is a block diagram which shows the structural example of a service level manager. It is a block diagram which shows the structural example of a structure management part. It is a figure which shows the structural example of a service level management table. It is a figure which shows the structural example of a service status management table. It is a figure which shows the structural example of a system configuration table. It is a figure which shows the structural example of a system management table. It is a figure which shows the structural example of a resource management table. It is a figure which shows the structural example of a migration schedule table. It is a figure which shows the structural example of a log | history management table. It is a flowchart which shows an example of a system configuration control process. It is a flowchart which shows an example of a performance monitoring process. It is a flowchart which shows an example of VM rearrangement plan determination processing. It is a flowchart which shows an example of VM rearrangement plan determination (degeneration) processing. It is a flowchart which shows an example of VM rearrangement processing. It is a timing chart which shows the operation example of a server management system. It is a timing chart which shows the operation example of a server management system.

  Hereinafter, modes for carrying out the present invention will be described.

  First, the configuration of the server management system according to the present embodiment will be described with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration example of a server management system.

  As shown in FIG. 1, the server management system 1 includes a physical server 10 (S01), a physical server (S02) 20, an FT (Fault Tolerant) server (FT1) 30, a shared disk 40, and a load balancer 50. The service level manager 100 is connected to service user terminals 70, 80, and 90 via the Internet 60.

  The physical server 10 includes, for example, a general-purpose server computer, and includes a hypervisor (HV1) 11 that executes a virtual machine (VM) on the real resource of the physical server 10 and allocates the real resource.

  The physical server 20 includes, for example, a general-purpose server computer, and includes a hypervisor (HV2) 21 that executes a virtual machine on the real resource of the physical server 20 and allocates the real resource.

  The FT server 30 has two (or three or more) hardware components such as a CPU (Central Processing Unit), a main memory, a hard disk, a network interface, a PCI (Peripheral Component Interconnect) slot, and a power supply. The configuration is made redundant. The FT server 30 includes a hypervisor 31 (HVft) that executes virtual machines on the real resources of the FT server 30 and allocates real resources.

  The shared disk 40 is composed of, for example, a hard disk or the like, and records actual files (image files) of data and virtual machine images.

  The shared disk 40 has a configuration that can be executed in the FT server 30 environment. In the case of a system that supports such an environment, an image file for the normal physical servers 10 and 20 and an image file for the FT server 30 are provided. Both are prepared. Then, from the shared disk 40, either the normal image file for the physical servers 10 and 20 or the image file for the FT server 30 is allocated to either the physical server 10, 20 or the FT server 30. Selected.

  The load balancer 50 distributes the traffic in the server management system 1 provided by a plurality of servers such as the physical servers 10 and 20.

  The service level manager 100 manages the service level, manages the virtual machine, manages the actual configuration, and performs recombination for maintaining the service level. The service level manager 100 includes, for example, a general-purpose server computer. .

  FIG. 2 is a block diagram illustrating a configuration example of the service level manager.

  As illustrated in FIG. 2, the service level manager 100 includes a service level management unit 110, a service monitoring unit 120, a relocation determination unit 130, a system configuration control unit 140, and a configuration management unit 150. .

  The service level management unit 110 is realized by, for example, a CPU, a ROM (Read Only Memory), a RAM (Random Access Memory), and the like, and records a service level agreement (SLA) and a state value for each service. to manage.

  The service monitoring unit 120 is realized by, for example, a CPU, a ROM, a RAM, and the like, and collects information related to a service level by monitoring each service status.

  The rearrangement determining unit 130 is realized by, for example, a CPU, a ROM, a RAM, and the like. When the state value managed by the service level management unit 110 exceeds a predetermined violation value or warning value, the risk of SLA violation ( The virtual machine placement destination (migration destination) schedule is determined.

  The system configuration control unit 140 is realized by, for example, a CPU, a ROM, a RAM, and the like, and performs rearrangement according to the placement destination of the virtual machine determined by the rearrangement determination unit 130.

  The configuration management unit 150 is realized by, for example, a hard disk or the like, and manages a table that records the resource information of the physical servers 10 and 20, the service, and the placement status of the virtual machines.

  FIG. 3 is a block diagram illustrating a configuration example of the configuration management unit.

  As shown in FIG. 3, the configuration management unit 150 includes a service level management table 151, a service status management table 152, a system configuration table 153, a system management table 154, a resource management table 155, and a migration schedule table 156. A history management table 157.

  The service level management table 151 records items to be managed as SLA for each system, and issues a prescribed value (service level) and warning that should be kept to a minimum such as response time (response time) and stop time. The value (warning value) and the risk value (violation risk value and warning risk value) set according to the penalty in the contract are stored.

  FIG. 4 is a diagram illustrating a configuration example of the service level management table.

  In the present embodiment, the service level management table 151 includes a service level ID (Identification Data), a system ID, a management item, a business ID, a service level (violation value), a violation, as shown in FIG. A risk value, a warning value, a warning risk value, and a management unit are stored in association with each other.

  The service status management table 152 shown in FIG. 3 manages the status of each item to be managed as an SLA for each system, and is appropriately updated by monitoring of the service monitoring unit 120.

  FIG. 5 is a diagram illustrating a configuration example of the service status management table.

  In the present embodiment, as shown in FIG. 5, the service status management table 152 includes a service level ID, a system ID, a management item, a business ID, a status value, the number of violations, the number of warnings, and the previous time. The update date and time are stored in association with each other.

  The system configuration table 153 shown in FIG. 3 represents the system configuration such as the virtual machines constituting each system, the actual image file of the virtual machines, whether or not automatic handling is possible, and the maximum multiplicity that can be clustered. Is.

  FIG. 6 is a diagram illustrating a configuration example of the system configuration table.

  In the present embodiment, as shown in FIG. 6, the system configuration table 153 includes a system ID, a system name, automatic handling, a maximum multiplicity, a system configuration machine name, and normal physical servers 10 and 20. And the image file for the FT server 30 are stored in association with each other.

  The system management table 154 shown in FIG. 3 is for managing the state of each system.

  FIG. 7 is a diagram illustrating a configuration example of the system management table.

  In the present embodiment, as shown in FIG. 7, the system management table 154 stores a system ID, a countermeasure status, and a multiplicity in association with each other. In the figure, Null represents a state in which no multiplexing is supported.

  The resource management table 155 shown in FIG. 3 manages on which server each virtual machine is operating.

  FIG. 8 is a diagram illustrating a configuration example of the resource management table.

  In this embodiment, as shown in FIG. 8, the resource management table 155 stores a virtual machine ID, a system ID, and a server ID in association with each other. In the figure, a virtual machine ID is assigned to each unit in which an image file corresponding to a virtual machine name actually operates as a virtual machine.

  The migration schedule table 156 shown in FIG. 3 manages the migration schedule and is updated by the rearrangement determination unit 130. When a new virtual machine is added, if there is no free space, the virtual machine is deleted and then added, so such a migration schedule table 156 is necessary.

  FIG. 9 is a diagram illustrating a configuration example of the migration schedule table.

  In the present embodiment, as shown in FIG. 9, the migration schedule table 156 stores a virtual machine ID, a migration source server, a migration destination server, and a migration reason in association with each other.

  FIG. 10 is a diagram illustrating a configuration example of the history management table.

  In the present embodiment, as shown in FIG. 10, the history management table 157 stores the date and time, the virtual machine ID, the migration source server, the migration destination server, and the migration reason in association with each other.

  Next, system configuration control processing executed by a server management system having the above configuration will be described with reference to the drawings.

  FIG. 11 is a flowchart illustrating an example of a system configuration control process.

  In the system configuration control process, the service level manager 100 first executes a performance monitoring process as shown in FIG. 11 (step S1).

  FIG. 12 is a flowchart illustrating an example of the performance monitoring process.

  When the performance monitoring process is started, the service monitoring unit 120 first measures the status of each system as shown in FIG. 12, and provides information on each status such as a predetermined response time (response time) as a service. It is collected as information relating to the level (step S11). More specifically, the service monitoring unit 120 measures the response time (response time) for each request, periodically checks the usage status of PING (Packet Internet Groper) and resources, etc. Collect relevant information.

  Next, the service level management unit 110 updates the status value stored in the service status management table 152 shown in FIG. 5 using the information collected in the process of step S11 (step S12).

  Then, the service level management unit 110 periodically performs each of the management unit described in the service level management table 151 illustrated in FIG. 4 and the previous update date / time described in the service status management table 152 illustrated in FIG. After resetting the number of service violations and the number of warnings (step S13), the performance monitoring process is terminated.

  In the process of step S13, the management unit and the previous update date / time are read from the service level management table 151 shown in FIG. 4 and the service status management table 152 shown in FIG. 5, respectively. It is determined whether or not (calculation unit period) has elapsed. If it is determined that the SLA calculation unit period has elapsed, it is determined that there is no predetermined time problem, and the number of violations and the number of warnings corresponding to the previous update date and time are reset. The update date / time is updated to the current date / time.

  After updating each situation (state value) in the performance monitoring process of step S1 shown in FIG. 11, the service level manager 100 executes an SLA violation possibility detection process (step S2). In the SLA violation possibility detection process, the service level management unit 110 refers to the service level management table 151 shown in FIG. 4 and detects a service that has reached a violation value or a warning value.

  In step S2, the violation value, the warning value, and the status value are read from the service level management table 151 shown in FIG. 4 and the service status management table 152 shown in FIG. 5, respectively. It is determined whether or not the value is exceeded. When it is determined that the state value exceeds the violation value or the warning value, it is determined that there is a possibility of SLA violation, and the system ID corresponding to the state value is read.

  If a service that has reached the violation value and the warning value is detected (step S3; Yes), the process proceeds to the next VM relocation plan determination process (step S4).

  FIG. 13 is a flowchart illustrating an example of the VM relocation plan determination process.

  When the VM relocation plan determination process is started, the relocation determination unit 130 first, as shown in FIG. 13, for a system having a migration target service (a service that has reached a violation value and a warning value), as shown in FIG. The system configuration table 153 shown in FIG. 4 is checked to determine whether the clustering is possible or the migration to the FT server 30 is possible, thereby determining the migration mode (step S41).

  In step S41, the maximum multiplicity and multiplicity corresponding to the system ID are read from the system configuration table 153 and the system management table 154 shown in FIG. 6, respectively, and whether the multiplicity has reached the maximum multiplicity. It is determined whether or not. Here, when it is determined that the multiplicity has not reached the maximum multiplicity, it is determined that the redundant configuration is possible, and it is determined to shift to the environment of the redundant configuration (cluster configuration). On the other hand, when it is determined that the multiplicity has reached the maximum multiplicity, it is determined that a redundant configuration cannot be obtained, and it is determined to shift to the FT server 30.

  Next, the rearrangement determination unit 130 determines whether the transition mode determined in the process of step S41 is transition to a redundant configuration environment or transition to the FT server 30 (step S42).

  Here, when it is determined that the migration mode determined in the process of step S41 is the migration to the redundant configuration environment (step S42; Yes), the migration is performed with reference to the resource management table 155 shown in FIG. Possible resource candidates are extracted (step S43). Then, the relocation determination unit 130 calculates the migration cost for each resource candidate, and determines the migration destination server as one of the physical servers 10 and 20 (step S44).

  The migration cost is calculated according to the availability of physical resources, the performance status of services on the physical resources, the presence or absence of virtual machines belonging to the same system, and the like. For example, when priority is given to the case where there is a vacancy in the migration destination server, the following migration cost calculation method can be considered.

  When there is a vacancy in the migration destination physical servers 10 and 20, the virtual machines constituting the same system are not collected in the same physical servers 10 and 20, and the virtual machine migrated with a similar warning (management type) In order not to collect the machines on the same physical servers 10 and 20, the migration cost to the physical servers 10 and 20 is calculated as shown in the following equation 1.

  Here, a to c are weighting factors, which are determined according to the scale of the physical servers 10 and 20 constituting the server management system 1 and the virtual machines that are actually operating.

  On the other hand, when there is no space in any of the migration destination physical servers 10 and 20, the SLA violation risk is calculated for all the multiplexed systems.

  The rearrangement determination unit 130 weights the SLA violation risk of the system i with respect to the number of violations and the number of warnings for the SLA (0 to j) set in each system, as shown in the following formula 2. calculate.

  Compared with the SLA violation risk of the system i that has newly issued a warning, if the system i that has newly issued a warning exceeds the SLA violation risk of any system, among the virtual machines included in the system, The virtual machine in the physical server 10 or 20 with the lowest cost is specified. Then, the identified virtual machine is unmounted from the physical servers 10 and 20, and a new virtual machine is arranged with the physical servers 10 and 20 as migration destinations.

  When it is determined that the migration mode is the migration to the FT server 30 in the process of step S42 (step S42; No), after determining the migration destination in the process of step S44, the relocation determining unit 130 The migration is recorded in the migration schedule table 156 shown in FIG. 9 (step S45), and the VM relocation plan determination process is terminated.

  Here, when there is no space to accommodate the virtual machine, a process of degenerating a virtual machine that has already been clustered or FTed and replacing it with the virtual machine that is to be handled this time is performed. Specifically, when it is determined that there is no free space for the virtual machine, the virtual machine to be degenerated is determined with reference to the service status management table 152 shown in FIG. Is recorded in the migration schedule table 156 shown in FIG.

  There is a possibility that a warning is generated in a plurality of services and the virtual machine is rearranged. In that case, referring to the migration schedule table 156 shown in FIG. 9, it is determined whether or not the migration destination is duplicated. If the migration destination is duplicated, the service from which resource candidates that can be migrated later are extracted separately. You can select the migration destination.

  When the service that has reached the violation value (service level) and the warning value is not detected in the process of step S3 shown in FIG. 11 (step S3; No), after executing the process of step S4, the service level management unit 110 executes a decline transition condition detection process (step S5). In the decline transition condition detection process, the service level management unit 110 determines whether there is a service in which the SLA calculation unit period has elapsed in the process of step S13 shown in FIG. (Step S6).

  Here, when it is determined that there is a service in which the number of violations and the number of warnings are reset (step S6; Yes), the relocation determination unit 130 executes a VM relocation plan determination (degeneration) process (step S7).

  FIG. 14 is a flowchart illustrating an example of VM relocation plan determination (degeneration) processing.

  When the VM relocation plan determination (degeneration) process is started, the relocation determination unit 130 relocates the virtual machine by an immediate violation or warning in order to secure a free resource as shown in FIG. The virtual machines to be degenerated (stopped) are sequentially determined from the physical servers 10 and 20 for those that have no possibility of rearrangement (step S71). As in the case of the VM relocation plan determination process shown in FIG. 13, the migration cost may be calculated and the virtual machine to be stopped may be determined.

  Here, in order to determine that there is no immediate violation or warning in the VM relocation plan decision (degeneration) processing, the smaller the number of occurrences of relocation, the safer it is. The number of occurrences of virtual machines and the number of virtual machines to be degenerated are safer, but if the number of occurrences of relocation is large, migration costs will occur and it is better to avoid them as much as possible. The evaluation value shown in Equation 3 is calculated, and a virtual server that is equal to or less than a predetermined threshold value may be determined to be degenerated.

  Note that d and e are predetermined constants that become weights.

  Thereafter, the relocation determination unit 130 records the reduction in the migration schedule table 156 shown in FIG. 9 (step S72), and then ends the VM relocation plan determination (degeneration) processing.

  If it is determined in step S6 that there is no service in which the number of violations and the number of warnings have been reset (step S6; No), after executing the VM relocation plan determination (degeneration) processing in step S7, system configuration control The unit 140 confirms the migration schedule table 156 shown in FIG. 9 and executes a VM relocation process for performing relocation of each virtual machine (step S8).

  FIG. 15 is a flowchart illustrating an example of the VM relocation process.

  When the VM rearrangement process is started, the system configuration control unit 140 first confirms the migration schedule table 156 shown in FIG. 9 as shown in FIG. 15, and determines the virtual server ID, migration server, migration destination server, The reason for the transition is specified (step S81).

  Here, when the migration reason specified in the process of step S81 is migration to the cluster configuration environment or migration to the FT server 30 (step S82; No), the image of the virtual machine corresponding to the virtual machine ID The file is specified from the system configuration table 153 shown in FIG. 6, and the migration destination (placement destination) server is instructed to read the image file of the target virtual machine from the shared disk 40, and the virtual machine is placed on the migration destination server. (Step S83).

  Further, the system configuration control unit 140 activates the virtual machine on the migration destination server (step S84).

  Thereafter, when it is confirmed that the activation of the virtual machine has been completed (step S85; Yes), if the reason for migration is the migration to the cluster configuration environment (step S86; Yes), the system corresponding to the virtual machine ID The ID is specified from the resource management table 155 (step S87).

  Then, the system configuration control unit 140 updates the system management table 154 by incrementing the multiplicity corresponding to the system ID specified in the process of step S87 by 1 (step S88), and then determines the VM relocation. End the process.

  On the other hand, when the reason for migration is migration to the FT server 30 (step S86; No), migration processing is executed (step S89). In the migration process in step S89, the virtual machine currently operating on the migration source server specified in the process in step S83 is stopped, while memory information (not shown) is read and the FT server 30 By continuing the processing on the virtual machine, the virtual machine is migrated to the FT server 30.

  When the migration reason identified in the process of step S81 is degeneration of the virtual server (step S82; Yes), the virtual server indicated by the virtual server ID is stopped on the migration source server (physical servers 10 and 20). A stop command for instructing execution of the process is issued (step S90).

  After that, when the stop notification is received from the migration source server (physical servers 10 and 20) and the stop of the virtual server is confirmed (step S91; Yes), the system configuration control unit 140 sets the system ID corresponding to the virtual machine ID. It is specified from the resource management table 155 (step S92).

  Then, the system configuration control unit 140 updates the system management table 154 by decrementing the multiplicity corresponding to the system ID identified in the process of step S92 (step S93), and then determines the VM relocation. End the process.

  Next, an operation example of the server management system that executes the system configuration control process will be described with reference to the drawings.

  16 and 17 are timing charts showing an operation example of the server management system.

  In this operation example, in the initial state, as shown in FIG. 16A, the virtual machines 001M01a and 001M02a configuring the system 001 are the physical server (SO1) 10, and the virtual machine 002M01a configuring the system 002 is the physical server ( It is assumed that it is operating at SO2) 20. Further, it is assumed that the system 003 is clustered and the virtual machine 003M01a is operating on the physical server (SO1) 10 and the 003M01b is operating on the physical server (SO2) 20.

  The service level management unit 110 detects the stop state of the system 002 in the performance monitoring process of step S1 shown in FIG. 11, and determines whether or not the stop time exceeds the warning time in the SLA violation possibility detection process of step S2. Determine.

  When the service level management unit 110 determines that the stop time has exceeded the warning time (step S3; Yes), the relocation determination unit 130 migrates the virtual machine 002M01a in the VM relocation plan determination process in step S4. Determine the destination (relocation destination) server.

  In the operation example shown in FIG. 16B, since it is determined that the migration to the FT server 30 is possible in the process of step S41 shown in FIG. 13, the FT server 30 is determined as the migration destination server (step S42). No), the physical server 20 as the migration source server and the FT server 30 as the migration destination server are recorded in the migration schedule table 156 shown in FIG. 9 in the process of step S45.

  Then, the system configuration control unit 140 refers to the system configuration table 153 illustrated in FIG. 6 in the process of step S83 illustrated in FIG. 15 and refers to the image file 002M01_f.for the FT server 30 corresponding to the virtual machine 002M01 from the shared disk 40. The vmx is read by the FT server 30, the virtual machine 002M01 is arranged on the FT server 30, and the virtual machine 002M01 on the FT server 30 is activated in the process of step S84.

  Thereafter, the service level management unit 110 refers to the service status management table 152 shown in FIG. 5 in the process of step S13 shown in FIG. 12, and the SLA calculation unit period corresponding to the system 003 has passed. If it is determined that there is no problem with time, the number of violations and the number of warnings are reset.

  In the operation example shown in FIG. 16C, the virtual machine 003 is determined as the virtual machine to be degenerated (stopped) from the physical servers 10 and 20 in the process of step S71 shown in FIG. 14, and in the process of step S72. Recording in the migration schedule table 156 shown in FIG. 9 is performed.

  Then, the system configuration control unit 140 stops the virtual machine 003 in the process of step S90 and removes the virtual machine 003 from the physical server 20.

  Next, the service level management unit 110 detects a decrease in the response time of the product search in the system 001 in the performance monitoring process in step S1 shown in FIG. 11, and the response time in the SLA violation possibility detection process in step S2. It is determined whether or not the warning time has been exceeded.

  If the service level management unit 110 determines that the response time has exceeded the warning time (step S3; Yes), the relocation determination unit 130 migrates the virtual machine 002M01a in the VM relocation plan determination process in step S4. Determine the destination (relocation destination) server.

  In the operation example shown in FIG. 17, since the multiplicity is within the range of the maximum multiplicity in the process of step S41 shown in FIG. 13, it is determined that clustering is possible. The server 20 is determined as the migration destination server (step S44), and the physical server 10 as the migration source server and the physical server 20 as the migration destination server are recorded in the migration schedule table 156 shown in FIG.

  Then, the system configuration control unit 140 refers to the system configuration table 153 illustrated in FIG. 6 in the process of step S83 illustrated in FIG. 15, and the image files corresponding to the virtual machines 001M01 and 001M02 configuring the system 001 from the shared disk 40. 001M01_f. vmx and 001M02_f. The vmx is read by the physical server 20, the virtual machines 001M01 and 001M02 are arranged on the physical server 20, and the virtual machine 001M01 and the virtual machine 001M01b on the FT server 30 are activated in the process of step S84.

  As described above, according to the server management system 1 according to the present embodiment, the hosting environment provided to the user is constructed using the machine virtualization technology, and the service level management table 151 and the service status management table 152 are stored. Based on the information related to the service level such as the stored SLA information and failure history, among the systems configured with virtual machines, the system that is at risk of SLA violation is prevented from causing SLA violation. Automatically migrate to a highly available environment.

  Specifically, the system configuration table 153 and the system management table 154 are used to manage whether or not the system can be configured redundantly. If the redundant configuration is possible, a virtual machine is added to the physical servers 10 and 20 to create a cluster. On the other hand, if a redundant configuration cannot be obtained, the configuration shifts to the physically robust FT server 30.

  In addition, after the SLA calculation unit period ends, by reducing the number of virtual machines that have a low risk of violating the SLA from the physical servers 10 and 20, the system moved to the FT server 30 or the redundant system is changed to the original single configuration. return.

  In this way, by appropriately switching to a highly available environment according to the SLA violation situation, it is possible to reduce costs and avoid SLA violations.

  As a result, before the risk of SLA violation occurs, the system is operated in an environment without redundancy, so that it can be operated at low cost. Moreover, the risk can be reduced by migrating a system having a business risk such as a penalty due to SLA violation to a high availability environment based on the failure occurrence history or the like. Furthermore, a system in which SLA is held with high probability can be provided to customers at a low price.

  In addition, this invention is not limited to the said embodiment, A various deformation | transformation and application are possible. Hereinafter, modifications of the above-described embodiment applicable to the present invention will be described.

  In the above embodiment, the program for executing the system configuration control process has been described as functioning as the service level manager 100 according to the above embodiment by applying it to an existing general-purpose computer in advance. However, the present invention is not limited to this. For example, it is stored in a computer-readable recording medium (for example, a flexible disk, a CD (Compact Disc) -ROM, a DVD (Digital Versatile Disc) -ROM, etc.) and distributed. Or a program downloaded from a storage on a network such as the Internet may be applied to an existing general-purpose computer to function as the service level manager 100 according to the embodiment.

  Further, when the system configuration control process is executed by sharing an OS (Operating System) and an application program, or by cooperation between the OS and the application program, only the application program may be stored in a recording medium or storage. . It is also possible to superimpose a program on a carrier wave and distribute it via a network. For example, the program may be posted on a bulletin board (BBS: Bulletin Board System) on the network, and the program may be distributed via the network. Then, this program may be activated and executed in the same manner as other application programs under the control of the OS, so that the above processing can be executed.

DESCRIPTION OF SYMBOLS 1 Server management system 10 Physical server 11 Hypervisor 20 Physical server 21 Hypervisor 30 FT server 31 Hypervisor 40 Shared disk 50 Load balancer 60 Internet 70 Service user terminal 80 Service user terminal 90 Service user terminal 100 Service level manager 110 Service level management Unit 120 service monitoring unit 130 relocation determining unit 140 system configuration control unit 150 configuration management unit 151 service level management table 152 service status management table 153 system configuration table 154 system management table 155 resource management table 156 migration schedule table 157 history management table

Claims (5)

The system violates the SLA by determining whether or not a predetermined state value in the system constituted by virtual machines on the physical server exceeds a violation value determined based on SLA (service level agreement). SLA violation determination means for determining whether there is a risk of
Redundant configuration determining means for determining whether it is possible to redundantly configure a system that is determined to be in danger of violating the SLA by the SLA violation determining unit;
When it is determined by the redundant configuration determining means that the system can be configured redundantly, the virtual machine configuring the system is arranged on a plurality of physical servers to be configured redundantly. A virtual machine migration means for migrating the virtual machine constituting the system from the physical server to a predetermined FT (Fault Tolerant) server when it is determined that the system cannot be redundantly configured;
A server management system comprising: SLA re-violation determining means for determining whether or not the redundantly configured system by the virtual machine migration means has a risk of violating the SLA again;
When it is determined by the SLA re-violation determining means that the system is not likely to violate the SLA again, any physical of the plurality of physical servers on which the virtual machines constituting the system are arranged Virtual machine reduction means for reducing the virtual machines on the server;
The server management system according to claim 1, further comprising: Each time the system is made redundant by the virtual machine migration means, the multiplicity is incremented by 1 and updated. On the other hand, every time the virtual machine is reduced by the virtual machine reduction means, the multiplicity is incremented by 1. It further comprises multiplicity update means for subtracting and updating,
The redundant configuration determination unit is configured such that the multiplicity updated by the multiplicity update unit is less than the maximum multiplicity of the system determined to be in danger of violating the SLA by the SLA violation determination unit. Determining that the system can be configured redundantly;
The server management system according to claim 2. The system violates the SLA by determining whether or not a predetermined state value in the system constituted by virtual machines on the physical server exceeds a violation value determined based on SLA (service level agreement). SLA violation determination step for determining whether there is a risk of
A redundant configuration determining step for determining whether the system determined to be in danger of violating the SLA by the SLA violation determining step can be configured redundantly;
When it is determined in the redundant configuration determination step that the system can be configured in a redundant configuration, the virtual machines constituting the system are arranged on a plurality of physical servers to be configured in a redundant configuration. A virtual machine migration step of migrating the virtual machine constituting the system from the physical server to a predetermined FT (Fault Tolerant) server when it is determined that the system cannot be redundantly configured;
A server management method comprising: On the computer,
The system violates the SLA by determining whether or not a predetermined state value in the system constituted by virtual machines on the physical server exceeds a violation value determined based on SLA (service level agreement). SLA violation determination procedure for determining whether there is a risk of
A redundant configuration determination procedure for determining whether it is possible to redundantly configure a system determined to be in danger of violating the SLA by the SLA violation determination procedure;
When it is determined by the redundant configuration determination procedure that the system can be configured redundantly, the virtual machines constituting the system are arranged on a plurality of physical servers to be configured redundantly. A virtual machine migration procedure for migrating the virtual machine constituting the system from the physical server to a predetermined FT (Fault Tolerant) server when it is determined that the system cannot be redundantly configured;
A program for running

Images