JP5447626B1 - Cluster system and cluster system update method - Google Patents

Abstract

When updating host software (SW) and control device SW accompanying changes in interface (IF) specifications between a host computer (host) and a control device, both without stopping the entire system A cluster system that can update the system is provided.
A cluster system includes a plurality of hosts and a control device including a first control unit and a second control unit shared by the hosts. SW for controlling IF with the control unit and the second control unit is provided, the first control unit and the second control unit are provided with SW for controlling IF with the host, and the host is the first control unit. The SW of the own computer is updated one by one in accordance with the completion of the update of the SW, and the second control unit updates the second control unit in accordance with the completion of the update of all the SWs of the host. Update the SW.
[Selection] Figure 1

Description

  The present invention relates to a technique for updating software in a cluster system.

  In a cluster system, even when a failure occurs or when software is updated, a technique for continuing processing without stopping the system is known.

  For example, Patent Document 1 discloses a technique for replacing a processor board without stopping the system when a plurality of processor boards that are communicably connected to the system bus have failed.

  Patent Document 2 discloses a technique related to a software update method for a processor in a system having a plurality of processors. Specifically, Patent Document 2 discloses a technique for reflecting the control state of the active system in the standby system by performing equivalent processing in the active and standby processors, and matching the data between the processors. It is disclosed.

  Patent Document 3 discloses a technique related to a loosely coupled computer system having a host computer and an MSCP device (Multi-System Control Processor: exclusive control device).

JP 2003-157151 A JP2002-49502 JP2011-216059

  However, Patent Document 1 discloses a technique when a failure occurs, that is, a technique for replacing parts without stopping the system when a failure occurs. Japanese Patent Application Laid-Open No. 2004-228561 describes a technique related to updating of processor software.

  In the computer system disclosed in Patent Document 3, in the case of performing an update accompanied by a change in the interface specification between the host computer and the MSCP device, if the technology disclosed in Patent Document 1 and Patent Document 2 described above is applied, the system is Must stop. The reason for this is that, when performing an update with a change in interface specifications, it is necessary to update both the software on the host computer side and the software on the MSCP device side at the same time.

  In view of the above-described problems, the present invention provides an update for both the host computer software and the control device software that is accompanied by a change in the interface specifications between the host computer and the control device without stopping the entire system. It is an object to provide a cluster system or the like that can perform the above.

  In order to achieve the above object, a cluster system according to the present invention includes a plurality of host computers and a control device including a first control unit and a second control unit shared by the host computers. In the cluster system, the host computer includes software for controlling an interface with the first control unit and the second control unit, and the first control unit and the second control unit are configured as the host computer. The host computer updates the software of its own computer one by one in response to completion of the software update of the first control unit, and the second control unit Update all software on the host computer. In response to, characterized by updating the software of the second control unit.

  In order to achieve the above object, a cluster system update method according to the present invention includes a plurality of host computers and a control device including a first control unit and a second control unit shared by the host computers. In the cluster system configured, the host computer includes software that controls an interface with the first control unit and the second control unit, and the first control unit and the second control unit include: Software for controlling an interface with the host computer, and updating the software of the host computer one by one in response to completion of the software update of the first control unit. As the update of To update the software of the control unit of.

  In the cluster system according to the present invention, when updating the software of the host computer and the software of the control device accompanied by the change of the interface specification between the host computer and the control device, both updates are performed without stopping the entire system. It can be carried out.

It is a figure which shows the structure of the cluster system 1 which concerns on the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the cluster system 1 which concerns on 1st Embodiment. It is a flowchart which shows the detail of step S104 of the cluster system 1 which concerns on 1st Embodiment. It is a flowchart which shows the detail of step S108 of the cluster system 1 which concerns on 1st Embodiment. It is a flowchart which shows the detail of step S114 of the cluster system 1 which concerns on 1st Embodiment. It is a figure showing the connection condition of each host 10A-10n and the control apparatus 20 in the middle of the update of the cluster system 1 which concerns on 1st Embodiment.

  Hereinafter, the present invention will be described in detail with reference to the drawings.

<First Embodiment>
First, the cluster system according to the first embodiment of the present invention will be described.

  FIG. 1 is a diagram showing a configuration of a cluster system 1 according to the first embodiment of the present invention. As shown in FIG. 1, the cluster system 1 includes a plurality of host computers (hereinafter referred to as “hosts”) 10 that operate by program control, and a first control unit 22 and a second control unit 24. 20.

  The plurality of hosts 10 are independent computers. Each host 10 is communicably connected to the first control unit 22 and the second control unit 24 (hereinafter, abbreviated as “connection” in the present application). An example of the host 10 is a server. Each of the plurality of hosts 10 includes software (hereinafter referred to as “SW”) that controls a communication interface (hereinafter referred to as “interface”) 14 with each of the first control unit 22 and the second control unit 24. ”). When the host 10 updates the SW 12, the host 10 disconnects from the first control unit 22 and the second control unit 24. The host 10 tries to connect to the first control unit 22 and the second control unit 24 when the update of the SW 12 is completed.

  When the host 10 is connected to both the first control unit 22 and the second control unit 24, the loads on both the first control unit 22 and the second control unit 24 are almost equal. Distribute processing in such a way (load distribution). On the other hand, when the connection is established with only one of the first control unit 22 and the second control unit 24, the host 10 performs processing only with the control unit with which the connection is established.

  The control device 20 includes a first control unit 22 and a second control unit 24, and is a device shared in the cluster system. The first control unit 22 and the second control unit 24 are connected to each other.

  An example of the control device 20 is an exclusive control device. When a conflict occurs due to simultaneous access from multiple processes to a shared resource used by multiple processes, the exclusive control device allows other processes to exclusively use the resources. Consistency is maintained by making it unavailable.

  The 1st control part 22 and the 2nd control part 24 are the same structures, for example. Each of the first control unit 22 and the second control unit 24 independently executes an exclusive control process and backs up the processing results. The first control unit 22 and the second control unit 24 are configured to back up processing results to each other so that when one device stops due to a failure, the operation of the system can be continued by the remaining device. .

The first control unit 22 and the second control unit 24 include SW 26 and SW 30 that control the interface 28 and the interface 32 with the host 10. The first control unit 22 and the second control unit 24 disconnect the connection with each host 10 when updating the SW 26 and SW 30. The first control unit 22 and the second control unit 24 try to connect to each host 10 when the update of the SW 26 and the SW 30 is completed. The first control unit 22 and the second control unit 24 read the revision (version) of the SW 12 of each host 10 and compare it with the compatible revision range information included in the updated SW 26 and SW 30. The first control unit 22 and the second control unit 24 establish a connection with the host 10 when the revision of the host 10 is within the range, and with the host 10 when the revision of the host 10 is out of the range. Next, the operation of the cluster system 1 will be described.

  FIG. 2 is a flowchart showing the operation of the cluster system 1 according to the first embodiment of the present invention.

  As shown in FIG. 2, in step S102, the first control unit 22 updates the SW 26.

  In step S <b> 104, the first control unit 22 tries to connect to each host 10 when the update of the SW 26 is completed. Details of step S104 will be described later.

  In step S106, each host 10 updates SW12.

  In step S <b> 108, the host 10 that has updated the SW 12 tries to connect to the first control unit 22 and the second control unit 24. Details of step S108 will be described later.

  In step S110, when the update of SW 12 of all the hosts 10 is completed, the process proceeds to step S112. If updating of the SWs 12 of all the hosts 10 has not been completed, the process proceeds to step S106.

  In step S112, the second control unit 24 updates the SW 30.

  In step S114, the second control unit 24 tries to connect to each host 10 when the update of the SW 30 is completed. Details of step S114 will be described later.

  FIG. 3 is a flowchart showing details of step S104 of the cluster system 1 according to the first embodiment of the present invention.

  As shown in FIG. 3, in step S <b> 202, the first control unit 22 that has completed the update of the SW 26 tries to connect to the host 10.

  In step S204, the first control unit 22 reads the revision of the SW 12 of the host 10 and compares it with the compatible revision range information included in the updated SW 26.

  In step S206, if the revision of the host 10 is included in the compatible revision range included in the updated SW 26, the first control unit 22 advances the process to step S208. On the other hand, if the revision of the host 10 is not included in the compatible revision range included in the updated SW 26, the first control unit 22 advances the process to S210.

  In step S208, the first control unit 22 establishes a connection with the host 10 including the compatible SW 12 (completes the connection).

  In step S210, the first control unit 22 suspends the connection with the host 10 including the incompatible SW12.

  In step S212, when the first control unit 22 performs the above-described steps S202 to S210 for all the hosts 10, the process proceeds to step S106. On the other hand, the first control unit 22 returns the process to S202 when the operations of S202 to S210 are not performed on all the hosts 10.

  FIG. 4 is a flowchart showing details of step S108 of the cluster system 1 according to the first embodiment of the present invention.

  As shown in FIG. 4, in step S <b> 302, the host 10 that has completed the update of the SW 12 tries to connect to the first control unit 22.

  In step S304, the first control unit 22 recognizes the connection from the host 10 in which the update of the SW 12 is completed, thereby reading the revision of the SW 12 of the host 10, and the compatible revision range information included in the SW 26. Compare with

  In step S306, if the revision of the host 10 is included in the compatible revision range included in the SW 26, the first control unit 22 advances the process to S308. On the other hand, if the revision of the host 10 is not included in the compatible revision range of the SW 26, the first control unit 22 advances the process to S310.

  In step S308, the first control unit 22 establishes a connection with the host 10 including the compatible SW 12 (completes the connection).

  In step S310, the first control unit 22 suspends the connection with the host 10 including the incompatible SW12.

  In step S312, the host 10 that has completed the update of the SW 12 tries to connect with the second control unit 24.

  In step S314, the second control unit 24 recognizes the connection from the host 10 in which the update of the SW 12 is completed, thereby reading the revision of the SW 12 of the host 10, and the compatible revision range information included in the SW 30. Compare with

  In step S316, if the revision of the host 10 is included in the compatible revision range included in the SW 30, the second control unit 24 advances the process to step S318. On the other hand, if the revision of the host 10 is not included in the compatible revision range of the SW 30, the second control unit 24 advances the process to S320.

  In step S318, the second control unit 24 establishes a connection with the host 10 including the compatible SW 12 (completes the connection).

  In step S320, the second control unit 24 suspends the connection with the host 10 including the incompatible SW 12.

  FIG. 5 is a flowchart showing details of step S114 of the cluster system 1 according to the first embodiment of the present invention.

  As illustrated in FIG. 5, in step S <b> 402, the second control unit 24 that has completed the update of the SW 30 tries to connect to the host 10.

  In step S404, the second control unit 24 reads the revision of the SW 12 of the host 10 and compares it with the compatible revision range information included in the updated SW 30.

  In step S406, if the revision of the host 10 is included in the compatible revision range included in the updated SW 30, the second control unit 24 advances the process to step S408. On the other hand, if the revision of the host 10 is not included in the compatible revision range of the updated SW 30, the second control unit 24 advances the process to S410.

  In step S408, the second control unit 24 establishes a connection with the host 10 including the compatible SW 12 (completes the connection).

  In step S410, the second control unit 24 suspends the connection with the host 10 including the incompatible SW12.

  In step S412, the second control unit 24 ends the process when the above-described steps S402 to S410 are performed on all the hosts 10. On the other hand, the second control unit 24 returns the process to S402 when the operations of S402 to S410 are not performed for all the hosts 10.

  Here, a case where information is exchanged between different hosts 10 via the control device 20 will be described. As shown in FIG. 1, when the connection between each host 10 and the first control unit 22 and the second control unit 24 included in the control device 20 is established, information exchange between different hosts is performed as follows. This is performed via one of the first control unit 22 and the second control unit 24.

  On the other hand, FIG. 6 is a diagram showing the connection status between the hosts 10A to 10n and the control device 20 during the update of the cluster system. As shown in FIG. 6, consider a case where the update of the SW 26 of the first control unit 22 and the SW 12A of one host 10A among the plurality of hosts 10A to 10n is completed. At this time, it is assumed that the connection between the first control unit 22 and the host 10A is established. In addition, it is assumed that the connection between the second connection unit 24 and the other hosts 10B to 10n has been established.

  Here, consider a case where the host 10B exchanges information with the host 10A. The host 10B instructs the second control unit 24 to exchange information with the host 10A. However, the second control unit 24 recognizes that there is no connection with the host 10A, and transfers the instruction from the host 10B to the first control unit 22. The first control unit 22 to which the instruction has been transferred recognizes that the connection with the host A has been established, and transfers the instruction to the host 10A. As described above, the host B exchanges information with the host A via the first control unit 22 and the second control unit 24.

  Therefore, when the cluster system 1 according to the first embodiment updates the SW of the host computer and the SW of the control device that perform functional enhancement accompanied by the change of the interface specification between the host computer and the control device, the system Both updates can be performed without stopping the whole. This is because the update is performed in the order of the first control unit 22, each host 10, and the second control unit 24. In addition, when the update is performed, the first control unit 22 and the second control unit 24 try to connect to each host 10 and connect to the host 10 having the compatible SW 12. Each host 10 tries to connect to the first control unit 22 and the second control unit 24 when updating, so that the first control unit 22 and the second control unit This is because the host 10 having the compatible SW 12 is connected.

1 Cluster system 10 Host computer (host)
12, 26, 30 Software (SW)
14, 28, 32 Communication interface (interface)
20 control device 22 first control unit 24 second control unit

Claims (4)

In a cluster system constituted by a plurality of host computers and a control device including a first control unit and a second control unit shared by the host computers,
Each of the host computers includes software for controlling an interface with the first control unit and the second control unit,
The first control unit and the second control unit include software for controlling an interface with each host computer,
Wherein each host computer in response to software update of the first control unit is completed, the software of the host computer and A Ppudeto,
The second control unit updates the software of the second control unit in response to completion of the software update of the plurality of host computers. The software of the first control unit and the second control unit reads the revision of the host computer when trying to connect to the host computer in the plurality of host computers, and is compatible with the software. compared to a revision of the range information, the case revision of the host computer is in a range to complete the connection process with the host computer, if the revision of the host computer is out of range the connection with the host computer The cluster system according to claim 1, wherein the processing is suspended. When each host computer updates the software of its own computer , the host computer disconnects the communication connection with the first control unit and the second control unit, and when the update of the software of its own computer is completed , the update is completed. The cluster system according to claim 1 , wherein the own computer individually tries the connection process . In a cluster system constituted by a plurality of host computers and a control device including a first control unit and a second control unit shared by the host computers,
Each of the host computers includes software for controlling an interface with the first control unit and the second control unit,
When the first control unit and the second control unit include software for controlling an interface with each host computer,
By the cluster system,
In response to the software update of the first control unit is completed, the software of the respective host computer A Ppudeto,
A cluster system update method, comprising: updating software of the second control unit in response to completion of software update of the plurality of host computers.

Images