JP2008003735A - Automatic stop system of information processing system connected to uninterruptible power supply - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To construct an information processing system with less man-hours at a lower cost than conventional systems by connecting a server having a hardware logical partitioning function and a hardware management console for managing the server and a logical patition, to an uninterruptible power supply and automatically stopping the logical partition, the server, and hardware management console, in operation, in this order in response to the occurrence of power suply failure such as a power failure and then stopping power supply from the uninterruptible power supply to the server and the hardware management console. <P>SOLUTION: The hardware management console and the uninterruptible power supply are connected by a network, and a server power automatic control module which automatically stops the logical partition, the server, and the hardware management console in this order in response to reception of a report of power supply failure such as a power failure detected by the uninterruptible power supply through the network is mounted on the hardware management console, and a management table storing information of the hardware management console currently operating is mounted on the uninterruptible power supply. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information processing system protected from an unexpected power supply abnormality such as a power failure by an uninterruptible power supply.

  When shutting down the server, it is necessary to execute the normal shutdown process. Otherwise, in the worst case, the data on the server can be destroyed. The time required for the shutdown process usually takes several minutes or more depending on the configuration of the server. For this reason, even when a power failure such as an unexpected power failure occurs, power must be continuously supplied to the server until the shutdown process is completed. One method for realizing this is to supply power to the server via an uninterruptible power supply. When an uninterruptible power supply detects a power failure such as a power failure, the source of output power to the server or the like is instantly switched to the built-in battery. As a result, the uninterruptible power supply can continue to supply power to the server or the like as long as the built-in battery power continues after the occurrence of power failure such as a power failure. However, since the time during which power can be continuously supplied from the built-in battery is limited, it is necessary to immediately start the server shutdown process when a power failure occurs.

  As a method for automatically starting a server shutdown with a power failure such as a power failure detected by the uninterruptible power supply as a trigger, there is a method disclosed in Patent Document 1, for example. A network communication function is provided in the uninterruptible power supply, and the server and the uninterruptible power supply are connected via a network. A monitoring process is resident on the server side to receive notifications about power supply abnormalities sent from the uninterruptible power supply via the network. The monitoring process automatically shuts down the server when it receives notification of power supply abnormality occurrences. Start processing. This method makes it possible to connect a large number of servers and the uninterruptible power supply more flexibly than the previous method of connecting with an RS232C cable.

JP 2005-18384 A

  In recent years, servers that provide a function of dividing hardware resources such as a processor, memory, and I / O into a plurality of logical partitions have appeared. In each logical partition, different operating systems and application programs can be installed and executed simultaneously. In the enterprise server EP8000 series provided by the present applicant, there are some which can define 100 or more such logical partitions. When a server having a large number of logical partitions and an uninterruptible power supply are connected by a conventional network method, a network adapter must be installed for each logical partition, cables must be connected, and network settings must be made. A monitoring process must be installed and the necessary settings must be made. The number of target logical partitions may be about 10, but if it exceeds several tens, the amount of work and cost become enormous. The first problem to be solved by the present invention is to reduce the amount of work and cost required to construct a system comprising a server having a large number of logical partitions and an uninterruptible power supply as compared with the conventional method.

  In addition, a hardware management console, which is a dedicated console that provides functions for starting, stopping, and monitoring the status of servers and logical partitions, is usually provided together with the server that provides the logical partitioning function for hardware resources as described above. The Since this hardware management console also requires the execution of a predetermined shutdown process when it is stopped, power must be supplied from the uninterruptible power supply as well as the server, and if a power failure occurs, the shutdown process must be performed promptly. . Also, when stopping a server, all logical partitions operating on it must be stopped first. The second problem to be solved by the present invention is to stop the logical partition, stop the server, stop the hardware management console, and stop power supply from the uninterruptible power supply to the server and hardware management console in this order. This is to reduce the amount of work and cost required to construct a system that executes automatically compared to the conventional method.

  In order to solve the above problems, a hardware management console and an uninterruptible power supply are connected via a network, and a server power supply automatic control module is newly mounted on the hardware management console. This module provides a function to automatically stop the logical partition, server, and hardware management console in this order when a power failure notification such as a power failure is received from the uninterruptible power supply via the network. This eliminates the need for having a monitoring process resident on the server side and connecting the uninterruptible power supply and the server via a network as in the prior art. In addition, the uninterruptible power supply is newly equipped with a management table that stores information on the currently operating hardware management console, and the power failure such as a power failure detected by referring to this management table Notify the management console. The uninterruptible power supply refers to the management table, and after confirming that all the hardware management consoles have stopped, stops power supply to the server and the hardware management console.

  By applying the present invention, installation and setting of network functions for each server and installation and setting of a monitoring process, which are necessary in the conventional method when connecting an uninterruptible power supply to a server or the like, become unnecessary. For this reason, especially when a server having a large number of logical partitions and an uninterruptible power supply are connected, the amount of work and cost required for system construction can be greatly reduced. In the method of the present invention, the amount of work and cost are determined by the number of hardware management consoles, regardless of the number of servers and logical partitions, so there are many servers and logical partitions connected per hardware management console. The effect becomes greater. For example, when connecting one server having 100 logical partitions, one hardware management console, and one uninterruptible power supply, the work required for every 100 logical partitions in the conventional method is This method only needs to be performed by one hardware management console, so the amount of work becomes 1/100. In addition, by applying the present invention, the system automatically stops power supply from the logical partition, the server, the hardware management console, and the uninterruptible power supply to the server and the hardware management console in this order. This can be easily realized without requiring complicated setting work in advance.

The purpose of reducing the amount of work and cost of building a system characterized by automatically shutting down servers and logical partitions triggered by power failure such as power failure detected by the uninterruptible power supply It was realized by installing a server power supply automatic control module on the hardware management console that receives power failure notifications such as power outages and automatically stops servers and logical partitions.
(Example 1)
First, Fig. 1 shows the system configuration. There are n logical partitions 1020 to 102n on the server 101, and an operating system and an application program are installed and running. The hardware management console 104 communicates with the service processor 103 of the server 101 via the network cable 113 and provides a function of monitoring the activation, stop, and status of the server 101 and each logical partition 1020 to logical partition 102n. The power supplied from the commercial power supply 109 enters the uninterruptible power supply 106 through the power cable 110 and is supplied to the server 101 and the hardware management console 104 through the power cable 111. On the hardware management console 104, a server power supply automatic providing a function of automatically stopping the server 101 and each logical partition 1020 to logical partition 102n triggered by a power failure such as a power failure detected by the uninterruptible power supply 106 A control module 105 is mounted. A power failure such as a power failure detected by the uninterruptible power supply 106 is notified from the communication module 107 to the server power automatic control module 105 through the network cable 112. The communication module 107 has a management table 108 for storing information of the currently operating hardware management console as means for storing a partner to notify a power supply abnormality such as a power failure.

  Next, the operation of the system will be described with reference to FIGS. FIG. 2 shows a processing flow when the server power supply automatic control module is activated. The server power automatic control module activated in 211 notifies the uninterruptible power supply of the activation of the hardware management console in 212 and waits for a stop request from the uninterruptible power supply in 213. On the other hand, the communication module in the uninterruptible power supply receives the activation notification of the hardware management console at 221 and registers the information of the hardware management console in the management table at 222.

  FIG. 3 shows a processing flow of the uninterruptible power supply when a power failure occurs. When the uninterruptible power supply detects a power failure at 301, the source of power supply to the server or the like is instantaneously switched to the built-in battery at 302, and the following two standby processes are started. The first standby process waits at 313 for the duration of the internal battery. The second process refers to the management table at 323 to obtain a list of active hardware management consoles, issues a stop request to the active hardware management console at 324, and periodically updates the management table at 325. And wait until all hardware management consoles stop. When the battery duration elapses or when the entire hardware management console stops, the power supply to the server or the like is stopped at 306.

  FIG. 4 shows the processing of the server power supply automatic control module on each hardware management console that has received the stop request issued from the uninterruptible power supply at 324 “issue a stop request to all operating hardware management consoles” in FIG. The flow is shown. When the server power supply automatic control module receives a stop request issued from the uninterruptible power supply at 411, it acquires a list of all servers currently operating at 412 and issues a stop command to those servers at 413 (this Details of the part are described in the next paragraph). After that, it waits for all servers to stop at 414, and when all servers are confirmed to stop, notifies the uninterruptible power supply to stop the hardware management console at 415 and issues a stop command to the hardware management console at 416 To do. On the other hand, when the communication module in the uninterruptible power supply receives a hardware management console stop notification at 421, the communication module deletes the information of the hardware management console on the management table at 422.

  Next, 413 “issue a stop command to all active servers” in FIG. 4 will be described in more detail. FIG. 5 is a stop command issue flow executed by the server power automatic control module for each server. In 511, the server power supply automatic control module issues a command for acquiring the statuses of all logical partitions on the server to the service processor in the server. When receiving this at 521, the service processor returns the status of all logical partitions to the server power supply automatic control module at 522. The server power supply automatic control module receives this at 512 and issues a stop command to all logical partitions currently operating at 513. The service processor receives this at 523 and stops the logical partition specified at 524. The server power supply automatic control module periodically acquires the status of all logical partitions on the server in the same manner as 511 and 512 in 514, and waits until all logical partitions are stopped. When all logical partitions are stopped, a command to stop the server is issued to the service processor at 515. The service processor receives this at 525 and stops the server at 526.

  The present invention may be utilized in a system that is protected from an unexpected power failure such as a power failure by an uninterruptible power supply device such as a core system.

1 is a diagram illustrating a system configuration of Embodiment 1. FIG. The figure which shows the server power supply automatic control module starting process flow of Example 1. FIG. The figure which shows the uninterruptible power supply processing flow at the time of the power failure generation of Example 1. FIG. The figure which shows the server power supply automatic control module processing flow at the time of the power failure generation of Example 1. FIG. The figure which shows the server stop process flow at the time of the power failure generation of Example 1. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 ... Server, 1024-102n ... Logical partition, 103 ... Service processor, 104 ... Hardware management console, 105 ... Server power supply automatic control module, 106 ... Uninterruptible power supply, 107 ... Communication module, 108 ... Management table, 109 ... Commercial power, 110 ... power cable, 111 ... power cable, 112 ... network cable, 113 ... network cable.

Claims (3)

  A server having a hardware logical partitioning function, a hardware management console that provides start, stop, and status monitoring functions of the server and the logical partition, and an uninterruptible power supply that supplies power to the server and the hardware management console In the information processing system, the server power supply automatic control module mounted on the hardware management console detects an abnormality in the power input to the uninterruptible power supply such as a power failure via the network, and automatically operates the logical partition. An information processing system characterized in that the server and the hardware management console are stopped in this order, and the power supply to the server and the hardware management console is stopped after the uninterruptible power supply confirms the stop of the hardware management console.   A server having a hardware logical partitioning function, a hardware management console that provides start, stop, and status monitoring functions of the server and the logical partition, and an uninterruptible power supply that supplies power to the server and the hardware management console In an information processing system, the server power automatic control module installed on the hardware management console detects the arrival of a preset time in the timer in the uninterruptible power supply via the network, and automatically operates logic An information processing system characterized by stopping a partition, a server, and a hardware management console in this order, and stopping power supply to the server and the hardware management console after the uninterruptible power supply confirms the stop of the hardware management console. An external power supply control device comprising an information processing system comprising a server, an external power supply control device that provides server start / stop and state monitoring functions, and an uninterruptible power supply that supplies power to the server and the external power supply control device The server power automatic control module installed above detects an abnormality in the power input to the uninterruptible power supply such as a power failure, etc. via the network, and after stopping the operating server automatically, stops the external power control device Then, after the uninterruptible power supply confirms the stop of the external power supply control device, power supply to the server or the external power supply control device is stopped.

Images