JP6254963B2 - Distributed processing system and distributed processing method - Google Patents

Distributed processing system and distributed processing method Download PDF

Info

Publication number
JP6254963B2
JP6254963B2 JP2015017931A JP2015017931A JP6254963B2 JP 6254963 B2 JP6254963 B2 JP 6254963B2 JP 2015017931 A JP2015017931 A JP 2015017931A JP 2015017931 A JP2015017931 A JP 2015017931A JP 6254963 B2 JP6254963 B2 JP 6254963B2
Authority
JP
Japan
Prior art keywords
server
db server
virtual machine
ap
db
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
JP2015017931A
Other languages
Japanese (ja)
Other versions
JP2016143202A (en
Inventor
康次郎 玉置
康次郎 玉置
勝 真田
勝 真田
小沢 誠司
誠司 小沢
武之 岡本
武之 岡本
Original Assignee
日本電信電話株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電信電話株式会社 filed Critical 日本電信電話株式会社
Priority to JP2015017931A priority Critical patent/JP6254963B2/en
Publication of JP2016143202A publication Critical patent/JP2016143202A/en
Application granted granted Critical
Publication of JP6254963B2 publication Critical patent/JP6254963B2/en
Application status is Active legal-status Critical
Anticipated expiration legal-status Critical

Links

Images

Description

  The present invention relates to a distributed processing technique for a system composed of a plurality of servers, and more particularly to a file update technique for the system.

  In the following description, “AP” represents an application, and “DB” represents a database. “Act” (Active) represents operation, operation, or online, and “Stb” (Standby) represents standby or offline. The term “server” includes an AP server and a DB server.

  As a technique related to system file update, for example, there is a technique disclosed in Patent Document 1. Patent Document 1 discloses a technique in which N file update servers are prepared as additional resources in an N-unit system, and the files are updated in parallel on these N servers. According to such a method, the file of each server in the system can be updated in a short time without affecting the service.

  Distributed system that consists of an Act system consisting of N + 1 AP servers and an Act / Stb system consisting of an active (0 system) / standby system (1 system) DB server and performs a distributed process to provide a predetermined service Regarding processing systems, there is a desire to update the files of the AP server and the DB server while providing services. When updating the file, it is indispensable to perform a predetermined test (for example, 1 call test) on the entire system through the AP server to the DB server in order to maintain the quality of service.

JP 2011-96161 A

  However, conventionally, with respect to the file update of the DB server, it is not possible to perform a preliminary test on the entire system before switching between the active system and the standby system. In other words, for a standby DB server that cannot handle one test call because it is offline, the file system is first updated and then switched to the active / standby system for the first time. Then, the one-call test of the entire system is ready to be performed. However, if the switching is completed, the terminal (end user) can use the service (release is completed). Therefore, for the former standby DB server switched to the active system, a one-call test has to be performed after the service is provided. As a result, there is a problem that even if the file is updated, a problem inherent in the entire system cannot be found until the service provision is started.

  However, there is no disclosure of any technology that can solve the problems related to the preliminary test described above, including the technology of Patent Document 1.

  In view of the above circumstances, an object of the present invention is to make it possible to perform a test relating to file update of the entire system without affecting the service.

  In order to solve the above-mentioned problem, the invention of claim 1 is directed to a plurality of AP servers, an operational DB server, and a standby DB server in which one or a plurality of virtual machines for providing a predetermined service are created. A storage in which a DB area referred to by the active DB server and the standby DB server is set, a distribution device that selects a virtual machine of the AP server to which an access from an external device is distributed, and the AP A distributed processing system comprising: a server, an operational DB server, the standby DB server, the storage, and a management device that controls the distribution device, wherein the management device tests the DB area. An area setting unit for setting a test area for the storage in the storage, the AP server, the active DB server, and the standby DB server A virtual machine control unit that creates a new virtual machine and assigns an address to the new virtual machine of the AP server, and each of the AP server, the active DB server, and the standby DB server. A file update unit that performs a file update process for each new virtual machine, a new virtual machine of the AP server to which the address assignment has been performed, and the active DB server that is in an operating state and refers to the test area A predetermined test for the new virtual machine is performed, the new virtual machine of the AP server to which the address is assigned, and the test area that is referred to when the standby state is switched to the active state A test execution unit that performs the predetermined test for a new virtual machine of the standby DB server. To.

  The invention according to claim 5 is a configuration in which one or a plurality of virtual machines for providing a predetermined service are created, a plurality of AP servers, an active DB server, a standby DB server, and the active DB A storage in which a DB area referred to by the server and the standby DB server is set, a distribution device that selects a virtual machine of the AP server to which an access from an external device is distributed, the AP server, and the active DB In a distributed processing system comprising: a server, the standby DB server, the storage, and a management device that controls the distribution device, the management device provides a test area for testing with respect to the DB area. In each of the area setting step to be set, and the AP server, the active DB server, and the standby DB server, A virtual machine control step of creating a virtual machine and assigning an address to the new virtual machine of the AP server; and each of the new virtual machines of the AP server, the active DB server, and the standby DB server A file update step for performing a file update process, a new virtual machine of the AP server to which the address assignment has been performed, and a new virtual machine of the active DB server that is in an operating state and refers to the test area A new virtual machine of the AP server to which the address assignment has been performed, and a new of the standby DB server that refers to the test area when switching from the standby state to the active state Performing a test execution step of performing the predetermined test on a virtual machine; A distributed processing method according to symptoms.

  According to the first and fifth aspects of the present invention, the two tests by the test execution unit (test execution step) can target a new virtual machine that is independent of the virtual machine that currently provides the service. Therefore, these tests can be performed before the service release by the new virtual machine is completed. Therefore, even if the file is updated, it is possible to find a problem inherent in the entire system before the service provision is started. Therefore, the distributed processing system according to the present embodiment makes it possible to perform a test relating to file update of the entire system without affecting the service.

  The invention according to claim 2 is the distributed processing system according to claim 1, wherein the management device uses the file update process to all addresses of virtual machines of the AP server to which the access is distributed. And an address information update unit that updates the address of the new virtual machine of the AP server.

  According to the second aspect of the present invention, it is possible to update all the virtual machine addresses of the AP server to which access is distributed to the new virtual machine address of the AP server at the same timing by the file update process. Therefore, the time required for updating the file of the AP server can be shortened. At this time, since a new virtual machine is created in each AP server, it is not necessary to add an AP server as a surplus resource.

  The invention according to claim 3 is the distributed processing system according to claim 1 or 2, wherein the management device does not satisfy a predetermined condition even after completion of the file update processing. In each of the AP server, the active DB server, and the standby DB server, the virtual machine originally created to provide the predetermined service and the virtual machine controller It is characterized by coexisting with a new virtual machine.

  According to the invention of claim 3, since the virtual machine and the new virtual machine created in each of the AP server, the active DB server, and the standby DB server can operate independently without affecting each other, Update file switching can be simplified, and labor can be greatly reduced.

  The invention of claim 4 is the distributed processing system according to any one of claims 1 to 3, wherein the management device is the AP server, the active DB server, and the standby system. A resource allocation unit that allocates resources from surplus resources of the AP server, the active DB server, and the standby DB server for each new virtual machine of the DB server, and the resource allocation unit Is characterized in that a resource capable of processing at least one call is allocated to each of the new virtual machines of the AP server, the active DB server, and the standby DB server.

  According to the invention of claim 4, when a test is performed on a new virtual machine created in the AP server, the active DB server, and the standby DB server, the resource allocation can be minimized.

  According to the present invention, it is possible to perform a test relating to file update of the entire system without affecting the service.

It is a figure which shows an example of a structure of a distributed processing system. It is a figure which shows an example of address information. It is a flowchart which shows the process sequence of a management apparatus. It is a figure explaining the preliminary test of the whole distributed processing system. It is a figure explaining the preliminary test of the whole distributed processing system. It is a figure explaining the service release in a distributed processing system. It is a figure explaining a mode that VM is switched simultaneously about a some AP server. It is a figure explaining switchback of the updated VM at the time of trouble occurrence. It is a figure explaining the resource allocation image of VM in operation in a distributed processing system, and the new VM used as the object of file update. It is a figure which shows the computer which runs a management program.

  EMBODIMENT OF THE INVENTION The form (embodiment) for implementing this invention is demonstrated in detail, referring drawings.

〔overall structure〕
As shown in FIG. 1, the distributed processing system according to the present embodiment includes a load balancer that receives access from N + 1 AP servers 10 serving as active servers, two DB servers 11, and a terminal 30 for using a service. (Sorting device) 20 and a management device 40 that controls a test by the test device 50 are provided.

  Each AP server 10 is an operation server, and provides a predetermined service by executing processing based on access from the terminal 30. The AP server 10 includes a virtual machine (VM) and a host OS (Operating System) that creates and controls the VM. In FIG. 1, only three AP servers 10 are illustrated, AP servers 10A to 10C, but in reality, N + 1 AP servers 10 are prepared.

  The DB server 11 operates a predetermined database management system and manages data necessary for providing services. The DB server 11 is switched to the active system (0 system) or the standby system (1 system) depending on the operation status. The DB server 11 includes a virtual machine and a host OS that creates and controls a VM. For convenience of explanation, it is assumed in FIG. 1 that there are two DB servers 11A and 11B as the DB server 11, but three or more DB servers 11 may be prepared.

  The storage 12 is an external storage device that stores data managed by the DB server 11. In the storage 12, a DB area 12a serving as a work area for reading and writing data managed by the DB server 11 is set. In addition, a test area 12b which is a snapshot of the writable DB area 12a is set in the storage 12.

  When the load balancer 20 receives an access from the terminal 30 via a network such as the Internet, the load balancer 20 distributes the access to a VM of any AP server 10. Details of the load balancer 20 will be described later.

  The management device 40 controls the AP server 10, the DB server 11, the storage 12, the load balancer 20, and the test device 50. For example, the management device 40 performs file update of the AP server 10 and file update of the DB server 11, and instructs the load balancer 20 to change the access distribution destination from the terminal 30.

The management device 40 creates two VMs as VM1 and VM2 in each of the AP server 10 and each of the DB servers 11. The VM 1 is a VM that is currently operating for providing services, and the VM 2 is a VM (new virtual machine) to be updated that is newly created for file update. Details of the management device 40 will be described later.
In addition, regarding the DB server 11, the management device 40 can register the operating state (Act) and the standby state (Stb) for each VM created in one DB server 11.

  The test apparatus 50 performs a test on the AP server 10 and the DB server 11 (which may further include the storage 12). For example, when the file update of the AP server 10 and the DB server 11 is performed, the test apparatus 50 performs a test (for example, one call) to confirm whether or not the AP server 10 and the DB server 11 operate normally. Test).

(Load balancer 20)
When the load balancer 20 receives an access from the terminal 30 (external device) or the like via the network, the load balancer 20 distributes the access to a VM of any AP server 10. The load balancer 20 includes address information 21 and a distribution unit 22.

  The address information 21 is information indicating an address (IP address) of a VM that is a distribution destination of access from the terminal 30. For example, RIP (Real IP address) is used as the IP address. Further, a VIP (Virtual IP address) is set in the load balancer 20. For example, as illustrated in FIG. 2, the address information 21 indicates the RIP of a VM that is a distribution destination of access addressed to the VIP of the load balancer 20. The address information 21 shown on the left side of FIG. 2 indicates that the destination of the access addressed to the VIP “aaa” is the VM of one of the addresses RIP1 “RIP1A, RIP1B, RIP1C,. Here, RIP1A, RIP1B, RIP1C,... Are RIPs assigned to the VM1 (sometimes referred to as “old VM”) of the AP servers 10A, 10B, and 10C, respectively. The load balancer 20 distributes access from the terminal 30 or the like to the VMs 1 of the AP servers 10A, 10B, 10C,.

  This address information 21 is updated by the management device 40. As a result, in the address information 21 shown on the right side of FIG. 2, the distribution destination of the access addressed to the VIP “aaa” is changed to the VM of any address of RIP2 “RIP2A, RIP2B, RIP2C,. Here, RIP2A, RIP2B, RIP2C,... Are RIPs assigned to the VMs 2 (may be referred to as “new VMs”) of the AP servers 10A, 10B, and 10C, respectively. The load balancer 20 distributes access from the terminal 30 or the like to the VMs 2 of the AP servers 10A, 10B, 10C,.

  When the address information 21 of the load balancer 20 is updated and the association with the VIP is changed from the RIP1 assigned to the VM1 to the RIP2 of the new VM2, the access addressed to the VIP can be distributed to the VM2. As a result, in each AP server 10, the access addressed to the VIP from the terminal 30 is switched from VM 1 to VM 2.

  The distribution unit 22 distributes access from the terminals 30. Specifically, when receiving an access from the terminal 30, the distribution unit 22 refers to the address information 21, selects an access destination VM, and distributes the access to the selected VM.

(Management device 40)
The management device 40 includes an input / output unit 41, a communication unit 42, a control unit 43, and a storage unit 44.

  The input / output unit 41 receives various instruction inputs and setting information inputs to the management apparatus 40 from an unillustrated HMI (Human Machine Interface) apparatus or the like. For example, the input / output unit 41 receives a file update instruction input of each AP server 10 or an input of setting information related to the load balancer 20 from an HMI device or the like. The input / output unit 41 is realized by an input / output interface that performs data input / output with an HMI device or the like.

  The communication unit 42 performs data communication with the load balancer 20, the AP server 10, the DB server 11, the storage 12, and the like. For example, the communication unit 42 transmits information related to the update of the address information 21 in the load balancer 20 to the load balancer 20, transmits a file update instruction to the AP server 10 and the DB server 11, The backup data is received from the DB server 11. The communication unit 42 is realized by a communication interface that performs data communication with an HMI device or the like using a predetermined communication protocol.

  The control unit 43 controls the entire management apparatus 40, and includes a virtual machine control unit 431, a resource allocation unit 432, a file update unit 433, a load balancer control unit 434, an address information update unit 435, and a backup unit 436. And an area setting unit 437 and a test execution unit 438. These parts will be described later. The control unit 43 is realized by a program execution process by a CPU (Central Processing Unit) included in the management apparatus 40 or by dedicated hardware.

  The storage unit 44 stores new VM backup data, addresses of servers and VMs in the system, resources allocated to the respective servers and VMs, setting information of the load balancer 20, and the like. The storage unit 44 is realized by a storage unit such as an HDD (Hard Disk Drive) or a RAM (Random Access Memory) included in the management device 40.

  The virtual machine control unit 431 controls the VMs of the AP server 10 and the DB server 11 via the communication unit 42. Specifically, when the virtual machine control unit 431 receives an instruction input for updating the files of the AP server 10 and the DB server 11 via the input / output unit 41, the virtual machine control unit 431 sends a file update instruction to each of the AP server 10 and the DB server 11. A new VM is created, and an address (for example, RIP) is assigned to the new VM. Also, for example, when an instruction to delete the old VM is received from each of the AP server 10 and the DB server 11 via the input / output unit 41 after the start of the operation of the new VM, each of the AP server 10 and the DB server 11 receives the old VM. Delete the VM.

  The resource allocation unit 432 allocates resources to the VMs of the AP server 10 and the DB server 11. For example, when a new VM is created in each of the AP server 10 and the DB server 11, the resource allocation unit 432 allocates resources to the new VM. The resources are allocated to the new VM from resources (surplus resources) that are not allocated to the operating VM in the AP server 10 or the DB server 11 in which the new VM is created. In each of the AP server 10 and the DB server 11 in the system, for example, a resource equivalent to one server divided by the number of servers (N + 3) constituting the distributed system is reserved as a spare resource. Has been. Therefore, even when a resource that can be used by a VM running on one server is fully used, it is possible to allocate a resource to a new VM using this spare resource.

  The file update unit 433 performs file update processing for each new VM. Specifically, for each of the AP server 10 and the DB server 11, the file update unit 433 changes the VM running on each of the AP server 10 and the DB server 11 to a new VM created with the VM. Copy and perform a file update process for this new VM.

  The load balancer control unit 434 closes the load balancer 20 and releases the blockage. “Blocking of the load balancer 20” means that access distribution by the load balancer 20 is stopped for a predetermined period. For example, the load balancer control unit 434 closes the load balancer 20 after the file update process to the new VM is performed by the file update unit 433, and updates the address information 21 of the load balancer 20 by the address information update unit 435. After the operation is performed, the blockage is released.

  The address information update unit 435 updates the VM address in the address information 21 of the load balancer 20 to a new VM address after the load balancer control unit 434 closes the load balancer 20. Thereby, for example, the address is switched as shown in FIG. When the load balancer 20 receives access from the terminal 30 after the load balancer 20 is released from blocking, the access from the terminal 30 is distributed based on the updated address information 21. That is, access from the terminal 30 is distributed to a new VM that has been updated.

  The backup unit 436 performs backup processing for each new VM of the AP server 10 and the DB server 11. Specifically, after the file update to the new VM of each of the AP server 10 and the DB server 11, the test apparatus 50 performs a test on the new VM, and when the normal operation is confirmed, Take a backup of each VM. The backup data is stored in a predetermined area of the storage unit 44.

  The area setting unit 437 sets a DB area 12 a referred to by the DB server 11 in the storage 12 in order to provide a service. Further, the area setting unit 437 sets the test area 12b referred to by the DB server 11 in the storage 12 in order to perform the test after the file update.

  The test execution unit 438 causes the test apparatus 50 to perform a test for file update.

〔processing〕
The processing procedure of the management apparatus 40 related to file update will be described mainly with reference to FIG.

First, the virtual machine control unit 431 receives a system file update instruction input via the input / output unit 41 (step S1).
Next, the management apparatus 40 prepares update files for the N + 1 AP servers 10 and the two DB servers 11 in the system (step S2).
Next, the area setting unit 437 sets the test area 12b in the storage 12 (step S3).
Next, the virtual machine control unit 431 instructs the AP server 10 and the DB server 11 to create a new VM (step S4).
Next, the resource allocation unit 432 allocates resources to the new VMs of the AP server 10 and the DB server 11 (step S5).
Next, the virtual machine control unit 431 assigns an address to each new VM (step S6).

  Next, the file update unit 433 instructs the new VM to update the file (step S7). For example, the file update unit 433 transmits the update file to each of the AP server 10 and the DB server 11. Then, each of the AP server 10 and the DB server 11 copies a VM running on its own server to a new VM, and then updates the file using this update file.

Next, the test execution unit 438 instructs the test apparatus 50 to perform a test on each new VM (VM after file update) of the AP server 10 and the DB server 11 (step S8).
If it is confirmed in step S8 that the new VM operates normally, the backup unit 436 performs backup processing for the new VM (step S9).

Next, the load balancer control unit 434 closes the load balancer 20 (step S10).
Next, the address information update unit 435 updates the address information 21 of the load balancer 20 (step S11). That is, the address information update unit 435 changes the VM address in the address information 21 to a new VM address.
Next, the load balancer control unit 434 releases the blockage of the load balancer 20 (step S12). As a result, the load balancer 20 distributes access with reference to the updated address information 21, so that access to a new VM is distributed.
Thus, the processing procedure of the management apparatus 40 regarding the file update is completed.

(Preliminary examination)
The test related to the file update shown in FIG. 3 includes a new VM of the AP server 10 and a new VM of the active DB server 11 that is in an operating state and refers to the test area 12b even during service provision. And a new VM of the AP server 10, and a new DB server 11 of the standby system that refers to the test area 12b when the standby state (Stb) is switched to the active state (Act). It can be implemented as a one-call test for VMs.

  As shown in FIG. 4, in the distributed processing system, the load balancer 20, the VM 1 of the AP server 10, the VM 1 (Act) of the DB server 11A, and the DB area 12a of the storage 12 are linked in an online state. (See arrows a1 and a2). Therefore, a predetermined service by the VM 1 of the AP server 10 is being provided to the terminal 30.

  When there is a file update instruction, the test area 12b is set in the storage 12 as a snapshot of the DB area 12a (see step S3 in FIG. 3). Further, regarding the DB servers 11A and 11B, a VM2 to be updated for file update, which is different from the already created VM1, is created (see step S4 in FIG. 3), and a predetermined amount of resources are allocated (step in FIG. 3). (See S5).

  The VM 2 of the DB server 11A mounts the test area 12b of the storage 12 (makes it available), enables the database management system of the DB server 11A (see arrow b2), and accepts a connection from the AP server 10. . As a result, the management apparatus 40 registers the VM 2 of the (0 system) DB server 11A in the storage unit 44 as the operating state (Act). The VM 1 of the DB server 11A mounts the DB area 12a of the storage 12, enables the database management system of the DB server 11A (see arrow a2), and accepts connection from the AP server 10. Further, the management apparatus 40 registers the VM 2 of the (system 1) DB server 11B in the storage unit 44 as a standby state (Stb).

  The management apparatus 40 performs processing so as to associate the VM 2 of the DB server 11A and the VM 2 of the DB server 11B (in FIG. 4, this association is indicated by “HA”), and high availability for service provision after file update (HA: High Availability). Note that the management apparatus 40 performs the same association with the VM1 of the DB server 11A and the VM1 of the DB server 11B, and realizes high availability related to the service being provided.

  On the other hand, for each AP server 10, a VM2 to be updated is created for file update, which is different from the already created VM1 (see step S4 in FIG. 3), and a predetermined amount of resources are allocated (step in FIG. 3). (See S5). The management device 40 registers the VM2 of the AP server 10 in the storage unit 44 in association with the VM2 of the DB server 11A. In each AP server 10, RIP1 is assigned to VM1, whereas RIP2 is assigned to VM2 (see step S6 in FIG. 3). The DB servers 11A and 11B are also assigned predetermined IP addresses for each VM, but are not shown in FIG.

  Therefore, the test apparatus 50 can perform a one-call test on the VM 2 of the AP server 10 that cooperates with the VM 2 of the DB server 11A that can refer to the test area 12b. The VM 2 of the AP server 10, the VM 2 (Act) of the DB server 11 A, and the test area 12 b (see the arrows b 1 and b 2) are linked in an online state, and the VM 1 of the AP server 10 and the VM 1 of the DB server 11 A ( Act) and the DB area 12a (see arrows of reference signs a1 and a2) exist independently. For this reason, the one-call test for the VM 2 of the AP server 10, the VM 2 (Act) of the DB server 11A, and the test area 12b is performed by the VM 1 of the AP server 10, the VM 1 (Act) of the DB server 11A, and the DB area 12a. It is carried out without interruption (see step S8 in FIG. 3).

  Further, as shown in FIG. 5, the VM 2 of the DB server 11B mounts the test area 12b of the storage 12 and enables the database management system of the DB server 11B (see the arrow b4). Accept connection. The VM 2 of the DB server 11A unmounts the test area 12b of the storage 12 (see the broken arrow). The management device 40 registers the VM2 of the AP server 10 in the storage unit 44 in association with the VM2 of the DB server 11B. As a result, the management apparatus 40 changes the VM2 of the DB server 11A from the operating state (Act) to the standby state (Stb), and changes the VM2 of the DB server 11B from the standby state (Stb) to the operating state (Act). The active system is switched for VM2 between 11A and 11B.

  Therefore, the test apparatus 50 can perform a one-call test on the VM 2 of the AP server 10 that cooperates with the VM 2 of the DB server 11B that can refer to the test area 12b. The VM 2 of the AP server 10, the VM 2 (Act) of the DB server 11 B, and the test area 12 b (see the arrows b 3 and b 4) are linked in an online state, the VM 1 of the AP server 10 and the VM 1 of the DB server 11 A ( Act) and the DB area 12a (see arrows of reference signs a1 and a2) exist independently. For this reason, the one-call test for the VM 2 of the AP server 10, the VM 2 (Act) of the DB server 11A, and the test area 12b is performed by the VM 1 of the AP server 10, the VM 1 (Act) of the DB server 11A, and the DB area 12a. It is carried out without interruption (see step S8 in FIG. 3).

  If stable operation is confirmed in the one-call test described with reference to FIGS. 4 and 5, the VM that provides the service is switched from VM1 to VM2 to be updated as shown in FIG. For each AP server 10, the load balancer 20 updates the association between VIP and RIP1 in the address information 21 to the association between VIP and RIP2 (see step S10 to step S12 in FIG. 3; see arrow b5 in FIG. 6). .)

  Further, the VM 1 of the DB server 11A unmounts the DB area 12a of the storage 12 (see the broken arrow), while the VM 2 of the DB server 11A mounts the DB area 12a of the storage 12 and operates the database management system of the DB server 11A. The connection from the AP server 10 is accepted. As a result, the management apparatus 40 sets the VM2 of the DB server 11B to the standby state (Stb) and sets the VM2 of the DB server 11A to the operating state (Act). Thereby, the service release by the VM 2 with the updated file is completed. Thereafter, when it is determined that a predetermined condition such as stable service operation is satisfied, the virtual machine control unit 431 deletes the VM1 before the file update.

  One call test (see FIG. 4) targeting the new VM of the AP server 10 and the new VM of the active DB server 11 that is in the active state and refers to the test area 12b, and the new AP server 10 The one-call test (see FIG. 5) for the VM and the new VM of the standby DB server 11 that refers to the test area 12b when the standby state (Stb) is switched to the active state (Act). You can target a new virtual machine that is independent of the virtual machine that currently provides the service. Thus, these tests can be performed before the service release shown in FIG. 6 is completed. For this reason, even if file update is performed, it is possible to find a problem inherent in the entire system before the service provision is started. Therefore, the distributed processing system according to the present embodiment makes it possible to perform a test relating to file update of the entire system without affecting the service.

(Parallel execution of file update of multiple AP servers)
As already described, in the distributed processing system of the present embodiment, when the management apparatus 40 updates each file of the AP server 10, the VM 1 (in FIG. 7, “ A new VM 2 (described as “scheduled to be updated” in FIG. 7) is created separately from “production”. In addition, the management apparatus 40 performs new RIP allocation and resource allocation for each of the new VMs 2. Then, the file is updated for this new VM2. Thereafter, the management device 40 causes the load balancer 20 to change the RIP corresponding to the VIP to the RIP of the new VM2, and distributes the access addressed to the VIP to the VM2. Thereby, the access addressed to the VIP from the terminal 30 with respect to each of the AP servers 10 is simultaneously switched from VM1 to VM2.

  That is, as shown in FIG. 7, the address information update unit 435 of the management device 40 stores RIP1A, RIP1B, RIP1C,... Assigned to each VM2 of the AP server 10 while the load balancer 20 is blocked. , RIP2A, RIP2B, RIP2C,... Can be updated at a time (see step S11 in FIG. 3). Such file update is possible because the VM1 and VM2 created in each of the AP servers 10 can operate independently from the VM1 and VM2 of the other AP servers 10, so that the RIP1 to RIP2 This is because the change can be freely performed in each AP server 10.

  Conventionally, file updates in a plurality of servers had to be performed sequentially one by one in order to continue providing services. For this reason, as the number of servers increases, the time required for file update increases, and the quality of service for file update decreases. In addition, in the past, in order to provide a service of the same quality as before the file update, it was necessary to add one server as a surplus resource to the server where the file was updated. As a result, server installation costs and operation costs in the distributed processing system have increased.

  The management apparatus 40 according to the present embodiment causes the address information update unit 435 to function, and by file update processing, all the addresses of the VM 1 of the AP server 10 that is the access distribution destination are the same as the addresses of the VM 2 of the AP server 10. Since the update can be performed at the timing, the time required for the file update of the AP server 10 can be shortened. At this time, since the VM 2 is created in each of the AP servers 10, it is not necessary to add an AP server that becomes a surplus resource.

(Switch back when unexpected trouble occurs)
Even after the file update process, the VM1 before the file update remains until a predetermined condition such as stable service provision is satisfied (see FIG. 6). If an unexpected trouble occurs before VM1 is deleted after confirmation of stable operation, it is necessary to switch back the update file. At this time, conventionally, since files were updated sequentially for each server, it was necessary to switch back the updated files one by one, which required a lot of time. Also, switching back an update file requires a different procedure and another verification from the time of file update, and requires a lot of labor. For example, in the file update process, when the procedure “Delete function A” → “Add function B” is taken, when the file is updated “Delete function B” → “Add function A” in the update file reversion It is necessary to take a different procedure. In consideration of the influence of the function A to be deleted on the function B and other functions, the method of leaving the function A to be deleted cannot be taken even after the file update process. Also, since the update contents are different, the verification method is also different.

  In the present embodiment, VM1 and VM2 coexist in each of the AP server 10 and the DB server 11 until the predetermined condition (stable operation confirmation) is satisfied even after the file update process is completed. As shown in FIG. 8, when an unexpected trouble has occurred, when the update file is switched back, the address information update unit 435 of the management device 40 first sets each of the AP servers 10 while the load balancer 20 is blocked. The address information 21 is updated so as to change RIP2A, RIP2B, RIP2C,... Assigned to the VM2 to RIP1A, RIP1B, RIP1C,.

  Thereafter, the VM 2 of the DB server 11A unmounts the DB area 12a of the storage 12, while the VM 1 of the DB server 11A mounts the DB area 12a of the storage 12 and enables the database management system of the DB server 11A (reference number a2). The connection from the AP server 10 is accepted. Thereby, the management apparatus 40 puts VM1 of the DB server 11A into an operating state (Act).

  Therefore, since VM1 and VM2 created in each of the AP server 10 and the DB server 11 can operate independently without affecting each other, it is possible to simplify update file reversion and greatly reduce labor. Can be reduced.

(Effective use of server resources)
Next, with reference to FIG. 9, an explanation will be given of the VMs in operation in the AP server 10 and the DB server 11 and new VM resource allocation targets for file update, with a specific example of the processing of the management device 40. To do. Here, the operating VM of the AP server 10 and the DB server 11 is referred to as VM1, and the new VM subject to file update is referred to as VM2.

  First, the management apparatus 40 instructs creation of a VM 2 and resource allocation in a state where the VM 1 of each of the AP server 10 and the DB server 11 in the system is operating (step S21). Next, after the test apparatus 50 performs a preliminary test on the VM 2 of each of the AP server 10 and the DB server 11 (step S22), the load balancer 20 is blocked and the allocation setting of the AP server 10 to the VM 2 is performed (step S22). S23). That is, the management device 40 closes the load balancer 20 and changes the access distribution destination in the address information 21 to the address of the VM 2. Thereafter, the management device 40 releases the load balancer 20 from being blocked. Thereafter, the management apparatus 40 instructs the AP server 10 and the DB server 11 to delete the VM 1 (step S24). As a result, the VM 2 operates.

  Here, overcommitment is applied to the effective resource allocation of the VMs in the AP server 10 and the DB server 11. In other words, from the creation of VM2 to the deletion of VM1, the same physical resources are allocated to VM1 and VM2 in terms of settings, but resources are allocated according to access to VM1 and VM2 in effect. Therefore, for example, as indicated by reference numeral 201, physical resources (except for the host OS) are allocated to VM1 and VM2 for setting. In other words, the server as a whole allocates about twice the physical resources in the settings.

  Here, by applying overcommitment to the effective resource allocation of the VM, for example, the VM1 is in operation, but the VM2 is in a state of receiving a preliminary test from the test apparatus 50 (FIG. 9). In the state of step S22), as indicated by reference numeral 202, the effective allocation resource (Y) of VM2 is a resource necessary for the test. In other words, the resource (X) that is the same as that of the VM 1 is not necessary because the VM 2 only needs to be allocated a resource necessary for the test (for example, a resource that can execute processing for one call).

  Here, if the product of “the maximum transaction value that can be processed with resource Y” and “the number of servers that make up the distributed processing system” is equal to or less than the maximum value of transactions that can be processed by one server, the system Even if the file is updated, additional resources to the server are not required.

  Note that effective allocation of resources to each VM is performed by the VM itself according to the transaction. However, since the maximum number of transactions per server does not change before and after the file update, X + Y in FIG. 9 does not exceed the physical resources of the server as a result.

  Therefore, according to the present embodiment, when a test is performed on a new VM created in the AP server 10 and the DB server 11, resource allocation can be minimized.

  According to the distributed processing system of the present embodiment, the above-described system-wide preliminary test is performed by creating the VM 1 provided for service provision and the VM 2 scheduled for file update in each of the AP server 10 and the DB server 11. Parallel execution of file update of AP servers, switch-back in case of unexpected trouble, and effective use of server resources can be realized simultaneously without introducing a new server.

(program)
In addition, it is possible to create a program in which processing executed by the management device 40 according to the above embodiment is described in a language that can be executed by a computer. In this case, the same effect as the above-described embodiment can be obtained by the computer executing the program. Further, such a program may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer and executed to execute the same processing as in the above embodiment. Hereinafter, an example of a computer that executes a management program that realizes the same function as the management apparatus 40 will be described.

  FIG. 10 is a diagram illustrating a computer that executes a management program. As shown in FIG. 10, the computer 1000 includes, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These units are connected by a bus 1080.

  The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores a boot program such as BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1090. The disk drive interface 1040 is connected to the disk drive 1100. A removable storage medium such as a magnetic disk or an optical disk is inserted into the disk drive 1100, for example. For example, a mouse 1110 and a keyboard 1120 are connected to the serial port interface 1050. For example, a display 1130 is connected to the video adapter 1060.

  Here, as shown in FIG. 10, the hard disk drive 1090 stores, for example, an OS 1091, an application program 1092, a program module 1093, and program data 1094. Each table described in the above embodiment is stored in the hard disk drive 1090 or the memory 1010, for example.

  Further, the management program is stored in the hard disk drive 1090 as a program module in which a command executed by the computer 1000 is described, for example. Specifically, a program module describing each process executed by the management apparatus 40 described in the above embodiment is stored in the hard disk drive 1090.

  Further, data used for information processing by the management program is stored in the hard disk drive 1090 as program data, for example. Then, the CPU 1020 reads out the program module 1093 and the program data 1094 stored in the hard disk drive 1090 to the RAM 1012 as necessary, and executes the above-described procedures.

  The program module 1093 and the program data 1094 related to the management program are not limited to being stored in the hard disk drive 1090. For example, the program module 1093 and the program data 1094 are stored in a removable storage medium and read by the CPU 1020 via the disk drive 1100 or the like. May be. Alternatively, the program module 1093 and the program data 1094 related to the management program are stored in another computer connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network), and are transmitted via the network interface 1070. It may be read by the CPU 1020.

≪Others≫
The test by the test apparatus 50 may be performed via the load balancer 20 as shown in FIGS. However, the test apparatus 50 may directly access the AP server 10 without going through the load balancer 20 and perform the test of this embodiment for the AP server 10, the DB server 11, and the storage 12.

  The external device of the present embodiment is not limited to the terminal 30 and includes, for example, a server device that can access the load balancer 20 via a network.

  In the present embodiment, over-commit is applied to the effective resource allocation of the VMs in the AP server 10 and the DB server 11, and the effective resource allocation (Y) of the VM 2 is the resource required for the test. For example, a resource capable of executing processing for one call is used. However, for each of the AP server 10 and the DB server 11, physical resources corresponding to the sum of VM1 and VM2 may be prepared in advance, and overcommitment may be unnecessary. Even in this case, the physical resources on the setting that are allocated to the VM1 and the VM2 by the same amount are not exceeded, so the VM1 (production) and the VM2 (update schedule) coexist for each of the AP server 10 and the DB server 11. It becomes possible. Therefore, it is possible to realize the preliminary test of the entire system, the parallel execution of the file update of a plurality of AP servers, and the switch back in case of an unexpected trouble, which have already been described.

A technique obtained by appropriately combining various techniques described in the present embodiment can also be realized.
The software described in this embodiment can be realized as hardware, and the hardware can also be realized as software.
In addition, hardware, software, flowcharts, and the like can be changed as appropriate without departing from the spirit of the present invention.

10, 10A, 10B, 10C AP server 11, 11A, 11B DB server 20 Load balancer (distribution device)
21 Address information 22 Sorting unit 30 Terminal (external device)
DESCRIPTION OF SYMBOLS 40 Management apparatus 41 Input / output part 42 Communication part 43 Control part 44 Storage part 50 Test apparatus 431 Virtual machine control part 432 Resource allocation part 433 File update part 434 Load balancer control part 435 Address information update part 436 Backup part 437 Area setting part 438 Test department

Claims (5)

  1. A plurality of AP servers, an active DB server, and a standby DB server in which one or a plurality of virtual machines for providing a predetermined service are created;
    A storage in which a DB area referred to by the active DB server and the standby DB server is set;
    A distribution device that selects a virtual machine of the AP server that is a distribution destination of access from an external device;
    A distributed processing system comprising the AP server, the operational DB server, the standby DB server, the storage, and a management device that controls the distribution device,
    The management device
    An area setting unit for setting a test area for testing in the storage with respect to the DB area;
    A virtual machine control unit that causes each of the AP server, the active DB server, and the standby DB server to create a new virtual machine and assigns an address to the new virtual machine of the AP server;
    A file update unit that performs a file update process for each of the new virtual machines of the AP server, the active DB server, and the standby DB server;
    A new test is performed on a new virtual machine of the AP server to which the address assignment has been performed, and a new virtual machine of the active DB server that is in an operating state and refers to the test area,
    The new virtual machine of the AP server to which the address is assigned and the predetermined virtual machine of the standby DB server that refers to the test area when the standby state is switched to the active state A test execution unit for performing the test,
    A distributed processing system characterized by that.
  2. The management device
    An address information update unit that updates all addresses of the virtual machine of the AP server to which the access is distributed by the file update process to the address of the new virtual machine of the AP server;
    The distributed processing system according to claim 1.
  3. The management device
    Even after the completion of the file update process, until the predetermined condition is satisfied, each of the AP server, the active DB server, and the standby DB server originally provides the predetermined service. Coexisting the created virtual machine and the new virtual machine created by the virtual machine control unit;
    The distributed processing system according to claim 1, wherein the system is a distributed processing system.
  4. The management device
    For each of the new virtual machines of the AP server, the active DB server, and the standby DB server, resources are allocated from the surplus resources of the AP server, the active DB server, and the standby DB server. A resource allocation unit for performing,
    The resource allocation unit allocates a resource capable of executing processing for at least one call to each of the new virtual machines of the AP server, the active DB server, and the standby DB server;
    The distributed processing system according to any one of claims 1 to 3, wherein the system is a distributed processing system.
  5. A plurality of AP servers, an active DB server, and a standby DB server in which one or a plurality of virtual machines for providing a predetermined service are created;
    A storage in which a DB area referred to by the active DB server and the standby DB server is set;
    A distribution device that selects a virtual machine of the AP server that is a distribution destination of access from an external device;
    In a distributed processing system comprising the AP server, the operational DB server, the standby DB server, the storage, and a management device that controls the distribution device,
    The management device is
    An area setting step for setting a test area for testing in the storage with respect to the DB area;
    A virtual machine control step of causing each of the AP server, the active DB server, and the standby DB server to create a new virtual machine and assigning an address to the new virtual machine of the AP server;
    A file update step for performing a file update process for each of the new virtual machines of the AP server, the active DB server, and the standby DB server;
    A new test is performed on a new virtual machine of the AP server to which the address assignment has been performed, and a new virtual machine of the active DB server that is in an operating state and refers to the test area,
    The new virtual machine of the AP server to which the address is assigned and the predetermined virtual machine of the standby DB server that refers to the test area when the standby state is switched to the active state Performing a test execution step for performing the test;
    A distributed processing method.
JP2015017931A 2015-01-30 2015-01-30 Distributed processing system and distributed processing method Active JP6254963B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2015017931A JP6254963B2 (en) 2015-01-30 2015-01-30 Distributed processing system and distributed processing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2015017931A JP6254963B2 (en) 2015-01-30 2015-01-30 Distributed processing system and distributed processing method

Publications (2)

Publication Number Publication Date
JP2016143202A JP2016143202A (en) 2016-08-08
JP6254963B2 true JP6254963B2 (en) 2017-12-27

Family

ID=56568775

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2015017931A Active JP6254963B2 (en) 2015-01-30 2015-01-30 Distributed processing system and distributed processing method

Country Status (1)

Country Link
JP (1) JP6254963B2 (en)

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05257673A (en) * 1992-03-11 1993-10-08 Fujitsu Commun Syst Ltd Old file reserving method at the time of updating file
JPH0644200A (en) * 1992-07-24 1994-02-18 Fujitsu Ltd System for controlling distribution processing
JP4426736B2 (en) * 2001-04-27 2010-03-03 株式会社日立製作所 Program correction method and program
JP2010003022A (en) * 2008-06-19 2010-01-07 Fujitsu Ltd File updating method
JP5648186B2 (en) * 2011-08-22 2015-01-07 インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Machines Corporation Apparatus and method for generating data for replaying a transaction
JP6075882B2 (en) * 2014-01-22 2017-02-08 日本電信電話株式会社 Distributed processing system and distributed processing method

Also Published As

Publication number Publication date
JP2016143202A (en) 2016-08-08

Similar Documents

Publication Publication Date Title
US8719415B1 (en) Use of temporarily available computing nodes for dynamic scaling of a cluster
JP5401922B2 (en) Virtual system control program, method and apparatus
CN1831775B (en) Systems and methods for multi-level intercept processing in a virtual machine environment
US8214686B2 (en) Distributed processing method
JP4056471B2 (en) System for transferring to a processor
JP5497201B2 (en) Method for allocating resources, computer program for allocating resources, and system for allocating resources
JP4295184B2 (en) Virtual computer system
US8370473B2 (en) Live multi-hop VM remote-migration over long distance
JP5459811B2 (en) Method and system for virtual storage migration and virtual machine monitor
US9329909B1 (en) Dynamically modifying a cluster of computing nodes used for distributed execution of a program
ES2734248T3 (en) Procedure to fork or migrate a virtual machine
US20180081891A1 (en) Sharing of Snapshots among Multiple Computing Machines
US9268590B2 (en) Provisioning a cluster of distributed computing platform based on placement strategy
US7784060B2 (en) Efficient virtual machine communication via virtual machine queues
US8370493B2 (en) Saving program execution state
JP5142678B2 (en) Deployment method and system
US9965304B2 (en) Delayed hardware upgrades in virtualization systems
US20130339956A1 (en) Computer system and optimal arrangement method of virtual machine in computer system
US20070258388A1 (en) Virtual server cloning
US8260840B1 (en) Dynamic scaling of a cluster of computing nodes used for distributed execution of a program
JP4769484B2 (en) Method and system for migrating virtual machines
JPWO2008146677A1 (en) Virtual machine management apparatus, virtual machine management method, and virtual machine management program
JP2008015768A (en) Storage system and data management method using the same
US20140137125A1 (en) Application migration with dynamic operating system containers
JP5427574B2 (en) Virtual computer migration management method, computer using the migration management method, virtualization mechanism using the migration management method, and computer system using the migration management method

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20170224

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20171115

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20171128

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20171201

R150 Certificate of patent or registration of utility model

Ref document number: 6254963

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R150