JP2007026038A - Path monitoring system, path monitoring method and path monitoring program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a path monitoring system, a path monitoring method and the like capable of effectively avoiding decrease in the performance of a system due to path faults. <P>SOLUTION: A faulty path number detecting part 11C detects the number of faulty paths where faults have occurred from a plurality of paths that connect an active server 1 to a storage 4. According to a limit number of paths that is stored in a path monitoring definition file 11A, a server switch determining part 11B determines whether or not the number of faulty paths detected by the faulty path number detecting part 11C is not less than the limit number of paths. When the server switch determining part 11B determines that the number of faulty paths is not less than the limit number of paths, the path monitoring system operates and a switch request for switching the active server 1 to a standby server 2 is outputted so that the switching is carried out. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a path monitoring system for monitoring a path in which a failure has occurred from a plurality of paths connecting a storage and an active server in order to switch a current server to a standby server when a failure occurs. The present invention relates to a method and a path monitoring program.

  2. Description of the Related Art Conventionally, a storage device multi-bus system technology that distributes access to a storage device to a plurality of access paths is known (see, for example, Patent Document 1). This multi-bus system includes a host computer that operates under program control, a bus control adapter connected to the host computer, a control bus, and a storage device. The host computer is provided with a plurality of path detecting means for detecting a plurality of paths, and the storage device has a unique storage area for storing a unique ID given to the storage device by the host computer.

  In this system, the multi-path detection means uses a single access path among a plurality of access paths between the host computer and the storage device to store the unique ID for the corresponding storage device in the unique ID storage area in the storage device. The storage device is associated with a plurality of access paths by writing the ID.

  In addition, a technique of a line system failure processing method in a hot standby system has been known (see, for example, Patent Document 2). This line fault processing method includes a first processing system comprising a main storage device and a central processing unit, a second processing system comprising a main storage device and a central processing unit, as well as the first processing system, A first communication control device and a second communication control device connectable to any one of the central processing unit of the first processing system and the central processing unit of the second processing system, and the first communication control device described above, The second communication control device is composed of a line switching device for switching the line to either the first processing system central processing device or the second processing system central processing device described above.

  In this line system fault processing method, even if a communication control device failure that does not degrade the processing system performance occurs, if the number of lines more than a certain level degenerates, the processing system is switched and the communication control device is initialized. In addition, the system is relieved from being stopped by reconnection with an external terminal.

JP 2000-330924 A JP-A-3-266133

  However, in the above-described multi-bus system of storage devices, a plurality of access paths to each storage device are automatically detected, and the load of access to the storage device is appropriately distributed to the plurality of access paths. However, the processing is not automatically switched to the standby server due to a path failure connecting between the active server being processed and the storage. Therefore, when a failure occurs in the access path, there is a problem that the system performance is degraded.

  Also, in the line fault processing method in the hot standby system, communication network line switching processing is performed, basically two lines are used, and there are a plurality of paths (for example, more than ten). It is not intended to maintain storage performance.

  The present invention has been made in view of the above-described problems, and it is possible to avoid in advance deterioration in performance due to a failure of a plurality of paths connected in good balance between the active server and the storage. An object of the present invention is to provide a path monitoring system, a path monitoring method, and a path monitoring program that can be used.

  To achieve the above object, according to the present invention, the occurrence of a failure is monitored for a plurality of paths connecting an active current server and a storage, and the current server functions equally when a failure occurs. A function for switching to a standby server is provided. The above-described current server is provided with a current bus monitoring mechanism that monitors the occurrence of a failure in a plurality of paths connecting the current server and the storage. This current bus monitoring mechanism includes a plurality of setting information storage means for preliminarily storing a limit path number indicating the number of faulty paths that is a reference for transferring processing to the standby server, and a plurality of connections between the current server and the storage. Failure path detection means for detecting the number of paths in which a failure has occurred from among the paths, and the limit number of paths in which the number of failure paths detected by the failure path detection means is stored in advance in the setting information storage means described above Path number determining means for determining whether or not the above is true. Furthermore, this current system bus monitoring mechanism operates when the path number determination means described above determines that the number of failed paths is greater than or equal to the limit number of paths, and outputs a request to switch from the current server to the standby server. The switching request output means is provided (claims 1 to 4).

  For this reason, it is possible to effectively detect the failure of a plurality of paths connected between the active server and the storage and to switch to the standby server immediately when the number of failed paths exceeds a predetermined number. It is possible to effectively avoid beforehand a decrease in performance of a system having multiple paths.

  Here, the standby server described above may be provided with a standby bus monitoring mechanism having the same configuration and functioning as the existing bus monitoring mechanism installed in the existing server. 2). In addition, a server switching control mechanism that controls the processing operation performed between the current server and the storage from the current server to the standby server may be provided in the above-described current server. . In this way, since server switching can be performed instantaneously, it is possible to construct a system that is reliable even when a failure path occurs.

  Further, the server switching control mechanism described above operates based on the switching request output from the switching request means of the existing bus monitoring mechanism and switches from the current server to the standby server, and switching of this server. A monitoring operation stop function that operates simultaneously and stops monitoring of the existing bus, and a monitoring operation start function that starts after the server is switched and starts monitoring the path on the side of the new current server; It is good also as a structure provided with. In this way, even if the server is switched, the monitoring of the connection path described above is not interrupted, and a reliable system can be constructed in this respect as well.

  In addition, the path monitoring method according to the present invention monitors the occurrence of a failure for a plurality of paths connecting an active current server and a storage, and functions the existing server equally when a failure occurs. A path monitoring method that functions to switch to a standby server, the fault path detecting step of detecting the number of paths in which a fault has occurred among the plurality of paths connecting the existing server and the storage, A path number determining step for determining whether or not the number of detected faulty paths is equal to or greater than a limit path number indicating the number of faulty paths that is a switching reference value stored in advance, and the number of paths in which the fault has occurred And a switching request output step that operates when it is determined that the number of paths is equal to or more than the limit path and outputs a server switching request.

  For this reason, it is possible to reliably detect the failure of a plurality of paths connected between the active server and the storage, and to immediately switch to the standby server when the number of failed paths exceeds a predetermined number.

  Here, a server switching control process for switching a server from a current server to a standby server based on the server switching request output by the switching request output process described above, and the current bus A monitoring operation stop step for stopping monitoring and a monitoring operation start step that starts after the server is switched and starts monitoring the path on the side that has become the new existing server may be provided. Item 6).

  A path monitoring program that periodically monitors the failure occurrence status of multiple paths connecting the current server and storage, and the number of paths that have failed from the multiple paths connecting the current server and storage Failure path detection function that detects the number of failed paths that determine whether or not the number of detected faulted paths is greater than or equal to a limit path number indicating the number of failed paths that constitute a server switching criterion stored in advance A determination function, a switching request output function that operates when it is determined that the number of failed paths is greater than or equal to the limit number of paths and outputs a switching request from the current server to the standby server is executed on the computer (Claim 7).

  As a result, it is possible to avoid in advance and quickly a decrease in performance due to a failure of a plurality of paths connected between the active server and the storage.

  According to the present invention, the failure of a plurality of paths connected between the active server and the storage can be detected effectively, and when the number of failed paths exceeds the predetermined number, the server is immediately switched to the standby server. Therefore, it is possible to effectively avoid a decrease in performance of a system having a plurality of paths in advance.

  The best mode for carrying out the invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a path monitoring system according to an embodiment of the present invention. The path monitoring system of this embodiment is a system that takes over processing (switches connection) to a standby server when a failure occurs in an active server such as a DB (database) server. FIG. 1 shows a connection relationship among an active server 1 as an active server, a standby server 2 as a standby system, a server switching control mechanism 3, and a storage 4. Here, the storage 4 is connected to the active server via a plurality of paths.

As shown in FIG. 1, the active server 1 is provided with a current path monitoring mechanism 11. Further, FC cards 12 denoted by reference numerals 1 to n represent path connection portions of the active server.
The existing path monitoring mechanism 11 includes a path monitoring definition file 11A, a server switching determination unit 11B, and a fault path number detection unit 11C. Here, the active server 1 includes a CPU (not shown) and a memory (not shown) for storing a path monitoring control program for executing the path monitoring system. Then, the CPU executes the functions of the server switching determination unit 11B and the fault path number detection unit 11C by executing the path monitoring control program.

  The server switching unit discriminating unit (failed path number determining unit and switching requesting unit) 11B first (of the paths connected to the storage 4 from the failed path number detecting unit 11C) (when the failed path number determining unit is activated). Whether the number of paths in which a failure has occurred (hereinafter referred to as "failed path") is periodically acquired and whether or not the number of paths exceeds the limit number described in the setting information defined in the path monitoring definition file 11A Is determined. If the server switching unit determination unit 11B determines that the number of limit paths described in the setting information is greater than or equal to the number, the server switching unit determination unit 11B switches processing to the standby server 2 that is on standby (with the switching request unit functioning). A switch request is output to the server switch control mechanism 3. The limit number of paths indicates the number of paths in which a failure has occurred and constitutes a reference for transferring processing from the active server 1 to the standby server 2.

  The path monitoring definition file 11A as setting information storage means stores setting information for execution by the server switching determination unit 11B. As shown in FIG. 2, the setting information includes an information collection time interval item 11Aa and a limit path number item 11Ab. The information collection time interval item 11Aa describes a time interval for detecting the number of failed paths from a plurality of paths by the failed path number detection unit 11C, for example, 60 seconds. The limit path number item 11Ab is described above. The limit path number is described. When the limit path number described in the limit path number item 11Ab is the default, the limit path number is “(server path number n / 2) +1”.

  In response to a request from the server switching determination unit 11B, the failure path number detection unit 11C as a failure path detection unit determines whether the active server 2 has the active server 2 for each information collection time interval described in the information collection time interval item 11Aa in the setting information. The number of failed paths is detected from 1 to n paths connected to the storage 4, and at the same time, the detected number of failed paths is output to the server switching determination unit 11B.

  The FC cards 12 denoted by reference numerals 1 to n are serial transmission cards, each having one path, and corresponding portions of the FC modules denoted by reference numerals 1 to n mounted in the storage 4. Data and control signals are input / output during

  The standby server 2 takes over the processing of the active server 1 when the active server 1 fails or when the possibility of failure is high, and has the same configuration as the active server 1. That is, the standby server 2 is provided with a standby path monitoring mechanism 21 as shown in FIG. FC cards 22 denoted by reference numerals 1 to n represent terminals for connection with the storage 4. Here, when the standby server 2 takes over the processing from the active server 1, it actually has the position of an active server.

  The standby path monitoring mechanism 21 described above includes a path monitoring definition file 21A, a server switching determination unit 21B, and a failed path number detection unit 21C. Here, the standby server 2 includes a CPU (not shown) and a memory (not shown) for storing the same path monitoring control program as the active server 1 that executes the path monitoring system. Then, the CPU executes the functions of the server switching determination unit 21B and the fault path number detection unit 21C by executing the path monitoring control program.

  The server switching unit discriminating unit 21B periodically obtains from the faulty path number detection unit 21C the faulty path in which a fault has occurred among the paths connected to the storage 4, and is defined in the path monitoring definition file 21A. It is determined whether or not the number of limit paths described in the setting information is exceeded. When it is determined that the number of limit paths described in the setting information is greater than or equal to the limit path number, the server switching unit determination unit 21B switches the process to the standby server that is on standby (actually, the above-described active server 1). The request is output to the server switching control mechanism 3.

  The path monitoring definition file 21A stores setting information to be executed by the server switching determination unit 21B. This setting information is the same as the path monitoring definition file 11A constituting the active server 1.

  The failure path number detection unit (failure path detection unit) 21C stores the active server 2 for each information collection time interval described in the information collection time interval item in the setting information in accordance with a request from the server switching determination unit 21B. 4 detects the number of failed paths from 1 to n paths connected to 4, and outputs the detected number of failed paths to the server switching determination unit 21B.

  The FC cards 22 with reference numerals 1 to n are serial transmission cards, each of which has one path, and corresponding portions of the FC modules with reference numerals 1 to n installed in the storage 4. Data and control signals are input / output during

The server switching control mechanism 3 takes over the processing of the active server 1 to the standby server 2, and includes a server switching control function 3A, a path monitoring operation stop function 31 that also functions as a server operation stop control unit, and a server operation A path monitoring operation start function 32 that also functions as a start unit is provided.
The server operation stop unit (path monitoring operation stop function) 31 stops the operation of the active server 1 (simultaneously or based on a request to transfer processing from the active server 1 to the standby server 2 input from the server switching determination unit 11A. Subsequently, the path monitoring operation is stopped), and a message to that effect is output to the server operation start unit (path monitoring operation start function) 32. The server operation start unit 32 operates based on the information indicating that the active server 1 input from the server operation stop unit 31 has been stopped, and starts the operation of the standby server 2 (at the same time or subsequently, the current operation system becomes active) The path monitoring operation on the standby server 2 side is started).

  The storage 4 is a storage device and includes a disk 41 and an FC module 42. The disk 41 stores data and the like, and the FC module 42 has 1 to n connector portions. Each connector portion includes predetermined FC cards 1 to n of the active server 1 and a standby server. 2 to 1-N predetermined FC cards.

  Next, the operation of the path monitoring system of this embodiment will be described with reference to FIG.

  When the power of the active server 1 is turned on, the server switching determination unit 11A constituting the path monitoring mechanism 11 reads the setting information stored in the path monitoring definition file 11A (step 100). Thereafter, the server switching determination unit 11B determines whether or not an information collection time interval described in the read setting information, for example, 60 seconds has elapsed (step 101).

  If the server switching determination unit 11B determines that the information collection time has not elapsed (step 101; no), the server switching determination unit 11B continues this process until the information collection time has elapsed, while the information collection time has elapsed. If the determination is made (step 101; Yes), a request for detecting the number of faulty paths among n paths connected between the active server 1 and the storage 4 is output to the fault path number detection unit 11C. To do.

  Next, the failure path number detection unit 11C outputs a test signal to the 1-N FC modules of the storage 4 via the 1-n FC cards 12 and outputs an output result of a response signal to the test signal, for example, When there is no response signal, the case where the response signal is not correct even though there is a response signal is treated as a faulty path, and the number of faulty fault paths out of the n paths is detected (step 102). Thereafter, the failure path number detection unit 11C outputs the detected number of failure paths to the server switching determination unit 11B.

  When the number of failure paths is input from the failure path number detection unit 11C, the server switching determination unit 11B determines whether or not the limit path number is described in the limit path number item 11Ab of the setting information read out earlier. Determination is made (step 103). When it is determined that the limit path number item is not described in the limit path number item, that is, when the limit path number item is the default (step 103; No), the server switching determination unit 11B sets the limit path number to “(all Pass n / 2) +1 ”(step 104), and the process proceeds to step 105. On the other hand, when it is determined that the limit path number is described in the limit path number item 11Ab (step 103; Yes), the server switching determination unit 11B determines the described number as the limit path number (step 105). .

  When the limit number of paths is determined, the server switching determination unit 11B determines whether or not the number of failure paths input from the failure path number detection unit 11C is equal to or greater than the determined limit path number (step 106). As a result, if the server switching determination unit 11B determines that the number of faulty paths is less than the determined limit number of paths (step 106; no), the server switching determination unit 11B starts counting the information collection time again, and the process proceeds to step 101. And the same processing as described above is continued. On the other hand, when the server switching determination unit 11B determines that the number of failed paths is equal to or greater than the determined limit number of paths (step 106; Yes), a path failure occurs between the active server 1 and the storage 4. In order to prevent a path failure from occurring, the switching control function 3A is activated to switch the active server 1 stop request in order to operate the standby server 2 instead of the active server 1. The data is output to the server operation stop control unit 31 of the control mechanism 3 (step 107).

  When the server switching control unit 31 receives a request to stop the active server 1 from the server switching determination unit 11B, the server switching control function 3A immediately operates to stop the operation of the active server 1 and stop the path monitoring operation. The fact that the active server 1 has been stopped is output to the server operation start control unit 32 (step 108).

  Thereafter, the server operation start control unit 32 outputs an operation request for the standby server 2 to operate the standby server 2 and start a path monitoring operation, and causes the standby server 2 to execute the same processing as that of the active server 1. Thereby, the switching process from the active server 1 to the standby server 2 is completed (steps 109 and 110).

  According to this embodiment, the server switching determination unit 11B determines whether or not the number of failed paths detected by the failed path number detection unit 3 is greater than or equal to the limit path number, and a failure has occurred. When it is determined that the number of paths is equal to or greater than the limit number of paths, a request to switch processing from the active server 1 to the standby server 2 is output, so that performance degradation due to a path failure between the active server 1 and the storage The fear of this can be reliably avoided in advance.

It is a block diagram which shows the structure of embodiment which concerns on this invention. It is explanatory drawing which shows the example of the setting information memorize | stored in the path monitoring definition file in FIG. It is a flowchart which shows operation | movement of embodiment which concerns on this invention.

Explanation of symbols

1 Active server (current server)
2 Standby server (standby server)
3 Server switching control mechanism 4 Storage 11 Current system path monitoring mechanism 11A Path monitoring definition file (setting information storage means)
11B Server switching determination unit (failure path number determination means, switching request means)
11C Failure path number detection unit (failure path detection means)
12 FC card 21 standby path monitoring mechanism 22 FC card 21A path monitoring definition file 21B server switching determination unit (failure path number determination means, switching request means)
21C Failure path number detection unit 31 Server operation stop control unit (path monitoring operation stop function)
32 Server operation start control unit (path monitoring operation start function)
41 disc 42 FC module

Claims (7)

Path monitoring with the ability to monitor the failure of multiple paths that connect the active current server and storage, and to switch the current server to a standby server that functions equally when a failure occurs In the system,
The current server is provided with a current bus monitoring mechanism for monitoring the failure occurrence of a plurality of paths connecting the current server and the storage,
This existing bus monitoring mechanism is configured to store in advance a limit path number indicating the number of failure paths, which is a reference for transferring processing to the standby server;
A failure path detection means for detecting the number of paths in which a failure has occurred from among the plurality of paths connecting the current server and the storage;
A path number determination means for determining whether or not the number of the failure path detected by the failure path detection means is greater than or equal to a limit path number stored in advance in the setting information storage means;
A switching request output unit that operates when it is determined by the path number determination unit that the number of failed paths is greater than or equal to the limit path number, and outputs a request to switch from the current server to the standby server;
A path monitoring system characterized by comprising: The path monitoring system according to claim 1, wherein
A path monitoring system, wherein a standby bus monitoring mechanism having the same configuration and functioning as the current bus monitoring mechanism installed in the current server is added to the standby server. In the path monitoring system according to claim 1 or 2,
A path monitoring system, characterized in that a server switching control mechanism for switching a processing operation performed between the current server and the storage from the current server to the standby server is provided in the current server. The path monitoring system according to claim 3, wherein
The server switching control mechanism operates based on a switching request output from the switching request means of the existing bus monitoring mechanism and switches from the current server to the standby server, and operates simultaneously with the server switching. A monitoring operation stop function for stopping monitoring of the existing bus, and a monitoring operation start function for starting monitoring of a path on the side of the new existing server that is operated after the server is switched. A path monitoring system characterized by Monitors the failure of multiple paths that connect the active current server and storage, and also switches the current server to a standby server that functions equally when a failure occurs. A method,
A failure path detection step of detecting the number of paths in which a failure has occurred from among the plurality of paths connecting the current server and the storage;
A path number determination step for determining whether or not the number of detected faulty paths is equal to or greater than a limit path number indicating the number of faulty paths that are switching reference values stored in advance;
A switching request output step that operates when it is determined that the number of paths in which the failure has occurred is greater than or equal to the limit number of paths, and outputs a server switching request;
A path monitoring method comprising: The path monitoring method according to claim 5, wherein
A server switching control step of switching the server from the current server to the standby server based on the switching request from the current server to the standby server;
A monitoring operation stop process that operates simultaneously with the server switching operation and stops monitoring the existing bus, and starts monitoring the path on the side that has become the new existing server that operates after the server switching. A path monitoring method comprising: a monitoring operation start step. A path monitoring program that periodically monitors the failure occurrence status of multiple paths connecting the current server and storage.
A failure path detection function for detecting the number of paths in which a failure has occurred from among the plurality of paths connecting the current server and the storage;
A failure path number determination function for determining whether or not the number of detected failure paths is greater than or equal to a limit path number indicating the number of failure occurrence paths that constitute a server switching criterion stored in advance;
Causing the computer to execute a switching request output function that operates when it is determined that the number of failed paths is greater than or equal to the limit number of paths, and outputs a switching request from the current server to the standby server. A featured path monitoring program.

Images