JP3117666B2 - Cold standby type dual system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cold standby type dual system in which only one server starts up at the same time, and more particularly to efficient operation of a standby server.

[0002]

2. Description of the Related Art A server constituting a dual system can take three status modes depending on its status: business, standby, and failure. When the server that is running and performing normal tasks is in the task mode, the server in the task mode fails for some reason and can no longer continue the task, and the server in the task mode fails. Then, the server that is in an operable standby state to take over the business on behalf of the server is set to the standby mode. As shown in FIG. 10, each of the servers 1 and 2 is equipped with server management devices 3 and 4 for managing the servers 1 and 2, and the state modes are managed by the server management devices 3 and 4. ing. Each of the server management devices 3 and 4 includes a CPU, a power supply, and the like separately from the servers 1 and 2.
Is configured to be able to operate independently even if the operation is stopped. Each of the server management devices 3 and 4 is connected by a dedicated line 5, and by exchanging information not only with itself but also with a server management device in another server via the dedicated line 5, the status mode of the other server is also changed. I know. The dual system is managed by the server management devices 3 and 4 monitoring the status mode of the entire system so that both servers 1 and 2 do not enter the business mode.

A dual system is usually used for mission-critical business and POS
Since the system is used for tasks that are difficult to stop when the system is stopped, the data on the shared disk 6 is used so that the data read and written by the active server can be continuously accessed from the server after switching. Management is essential. Particularly, in a cold standby type dual system in which only one server starts up at the same time, a logical switch 8 realized by the firmware of the disk system 7 is provided between each of the servers 1 and 2 and the shared disk 6 to provide a system. The configuration is such that access to the shared disk 6 can be performed only from the server in the business mode at the time of startup. As described above, by simultaneously accepting access from only one server, management of system files such as a paging file is facilitated.

Referring to FIG. 10, the operation of the conventional cold standby dual system when the business server 1 operating in the normal operation mode fails and switches to the standby server 2 will be described.

When the cold standby type dual system is started, the shared disk 6 stores the business server 1 in the business mode.
Connected only to The business server 1 has a shared disk 6
Is started by loading the OS stored in the OS. On the other hand, since the standby server 2 is not connected to the shared disk 6, the standby server 2 stands by without starting the system.

When the business server 1 fails for some reason, the server management unit 3 detects the failure, switches itself to the failure mode, and notifies the standby server 2 of the failure.
Is notified to the server management device 4. The server management device 4
When recognizing that the business server 1 has failed, the system switches the standby mode to the business mode and reboots the system.

At this time, the logical switch 8 switches the connection destination of the shared disk 6 to the standby server 2 in response to a connection request from the standby server 2 in the business mode. The standby server 2 is started by loading the OS stored in the accessible shared disk 6 into the memory. On the other hand, the failed business server 1 does not naturally start up, but is not connected to the shared disk 6 even after the repair is completed, and thus shifts from the failure mode to the standby mode in a state where the system has not started up. Will be.

[0008] The client 9 that can access the server of the dual system can specify the logical network address common to the servers, and thereby the business server 1
Alternatively, the user can access the standby server 2. That is, the client 9 can access without being conscious of the service provider.

[0009] As described above, the dual system is a key system even if one server goes down when one server is providing business to a client and the other server is simply put on standby. The main purpose is to be able to provide services continuously without stopping business. Thus, the server does not need to have a local disk.

[0010]

However, the cost required for each server constituting the dual system is not inexpensive at all, so if only one of the servers is kept on standby, the cost efficiency is considered. It is hard to say that this is a typical operation. The server in the standby mode only needs to be able to take over the execution of the business on behalf of the failed server when the server in the business mode fails, and there is no problem if the server is used for other purposes while waiting. It can be said that it does not come.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and a purpose thereof is to provide a dual system capable of effectively utilizing a standby server without causing any trouble. It is to provide a heavy system.

[0012]

In order to achieve the above object, a cold standby type dual system according to a first aspect of the present invention comprises two units, a business system and a standby system.
A cold standby dual system including a server having a boot mechanism that scans devices in a certain order until a boot device is searched for at the time of system boot, and a shared disk that allows only one of the servers to access at the same time. In the above, the standby server has a local system disk connected to a position scanned after the shared disk, and starts up from the local system disk when standby as a dual system.

[0013] In the cold standby dual system according to a second aspect of the present invention, in the first aspect, the standby server performs a reboot after receiving a command from the outside when the business server fails. It has an activation management means.

In the cold standby type dual system according to the third invention, in the first invention, when the standby server is performing a business on behalf of the business server, the standby system server may execute the business server. After the restoration, the system has a system activation management means for rebooting both of the servers after receiving an external instruction.

In the cold standby type dual system according to a fourth aspect, in the first aspect, when the standby server is performing a job on behalf of the business server, the standby server may be used for the business server. It has a system activation management means for automatically detecting the restoration and rebooting both servers.

According to a fifth aspect of the present invention, in the cold standby dual system according to the first aspect, the service server is connected to a slot scanned after the slot to which the shared disk is connected. It has a local system disk, and when it stands by as a dual system after restoration, it starts up from the business local system disk.

[0017]

Preferred embodiments of the present invention will be described below with reference to the drawings. Note that components that may be the same as those in the related art are given the same reference numerals.

Embodiment 1 FIG. 1 is an overall configuration diagram showing Embodiment 1 of a cold standby type duplex system according to the present invention. The dual system has two servers 10 and 20 that execute business, one of which is operated as a business system and the other is operated as a standby system. Each of the servers 10 and 20 is equipped with server management devices 3 and 4 for managing the servers 10 and 20, and the state modes are managed by the server management devices 3 and 4. Each of the server management devices 3 and 4 includes a CPU and a power supply device separately from the servers 10 and 20, and is configured to be able to operate independently even when the servers 10 and 20 are stopped. I have. Each of the server management devices 3 and 4 is connected by a dedicated line 5, and by exchanging information not only with itself but also with a server management device in another server through the dedicated line 5, the status mode of the other server is also changed. I know. The dual system is
The server management devices 3 and 4 monitor the status mode of the entire system so that both servers 10 and 20 are managed so as not to enter the business mode.

A shared disk 6 storing an OS and the like is connected to each of the servers 10 and 20. In the dual system according to the present embodiment, each server 10,
A logical switch 8 implemented by the firmware of the disk system 7 is provided between the shared disk 6 and the shared disk 6 so that the shared disk 6 can be accessed only by the server in the operation mode when the system is started. . That is, only one of the servers can access the shared disk 6 at the same time. The standby server 20 has a locally bootable OS.
Is connected.

FIG. 2 is a block diagram of the standby server 20 according to the present embodiment. A CPU 23, a memory 24, a BIO
A bus controller 27 connecting the S25 and the PCI bus 26 is connected. The PCI bus 26 has two S
The CSI boards 28 and 29 are connected to each other. Among them, the SCSI board 28 inserted in the slot # 0 stores one shared disk 6 in the S # inserted in the slot # 1.
The CSI board 29 connects one local disk 21 to each. Note that the boards, devices, and the like that constitute the standby server 20 can be composed of boards similar to those of a general computer.

In a general computer, at the time of system boot, a connected device (usually a disk) is scanned to find a boot sector, and a system (OS) is started from the device. If there is no boot sector in the scanned device, go to another device for scanning. This boot mechanism is generally realized by the BIOS, but the scanning order of the devices is predetermined. Normally, the PCI bus with the smallest number (ID) starts with the smallest number, the slot number with the same PCI bus has the smallest number, and the SCSI adapter with the same SCSI adapter has the smallest number.
Scan is performed in ascending order of ID. That is, BIOS
Will scan the devices connected to each connector in a certain order according to the above rules until a boot device is found. Of course, it is possible to change the order of the devices to be scanned, but in any case, there is no change in scanning the boot devices in a set fixed order.

The servers 10 and 20 according to the present embodiment also have the boot mechanism described above. Based on the above boot time specification, when starting up the standby server 20 shown in FIG.
In this example, one PCI bus 2 is used.
Only 6 are connected. And the PCI bus 26
At first, the system on the shared disk 6 connected to the slot # 0 having the smaller slot number is to be started. If not possible, the position scanned after the shared disk 6, that is, the next smaller slot #
The system operates to start up the system on the local disk 21 connected to 1.

The business server 10 in the present embodiment has a configuration in which the SCSI board 29 and the local disk 21 are removed from the configuration of the standby server 20 shown in FIG.

A characteristic of this embodiment is that the standby server 20 is scanned after the slot # 0 to which the shared disk 6 is connected when the system boots.
1 is connected to a system disk that can be started up locally. Thus, when the business server 10 is operating normally, the local disk 2
Standby server 2 by starting system 1
While 0 is waiting as a dual system, it will be available for another purpose.

Next, the operation of the dual system according to the present embodiment will be described with reference to the flowchart shown in FIG.

First, when the servers 10 and 20 constituting the dual system start up, the logical switch 8 connects the shared disk 6 to the business server 10 in the business mode. Thus, the business server 10 starts up by loading the OS stored in the shared disk 6. On the other hand, the standby server 20 starts the system boot process even in the standby mode. At this time, the standby server 20 attempts to start up the system from the shared disk 6 connected to the slot # 0 first as described above (step 101). However, since it is not connected to the shared disk 6, the OS stored on the shared disk 6 cannot be loaded. Therefore, the BIOS 25 searches for the next boot device, that is, the boot sector of the local disk 21 connected to the slot # 1. Since the standby server 20 can always access the local disk 21, the system is started from the local disk 21 (step 102).

As described above, according to the present embodiment, when the business server 10 has started up normally, the standby server 20 can start up the system on the local disk 21. During normal operation, the standby server 20 can be used for a task different from the task.

If the business server 10 fails for some reason, the server management device 3 detects the failure, switches itself to the failure mode, and notifies the server management device 4 of the standby server 20 of the failure. Notice. When recognizing that the business server 10 has failed, the server management device 4 switches the standby mode to the business mode and automatically reboots the system.

At this time, the logical switch 8 switches the connection destination of the shared disk 6 to the standby server 20 in response to a connection request from the standby server 20 in the business mode.
The standby server 20 first attempts to start up the system from the shared disk 6 in the same manner as described above (step 1).
01). In this case, since the standby server 20 is connected to the shared disk 6, the system can be started from the shared disk 6 (step 103). That is, even if the standby server 20 performs the same processing operation as in the standby mode, the standby server 20 can start up the system of the shared disk 6 this time. Can be demonstrated.

As described above, according to the present embodiment, the local disk 21 is connected to the slot # 1 scanned after the slot # 0 to which the shared disk 6 is connected at the time of system boot in the standby server 20. Therefore, the function as a standby system is exhibited by starting up from the shared disk 6 in the business mode, and the system can be used as another business system in the standby mode. For this reason, the standby server 20 during standby
Can be used effectively.

For example, the standby server 20 started from the local disk 21 is used for incidental work. Unlike the core business provided by the dual system, the auxiliary business is a business that can be interrupted, so there is no special problem even if it is forcibly stopped to reboot as a dual system . In other words, if the task does not cause a problem even if the task is interrupted, the standby server 20
Can be used, and it is more efficient to effectively use the waiting server 20 than to be interrupted. As described above, even if the standby server 20 is used for ancillary work, the cold standby type is still maintained from the viewpoint of the core work, that is, the work originally performed by the dual system.

Embodiment 2 FIG. As described in the first embodiment, when the business server 10 fails, the server management device 4 of the standby server 20 automatically reboots after switching the state mode from the standby mode to the business mode. . However, in the standby server 20, a system for another job has been started, and it may be unfavorable to forcibly bring down the local system, for example, even if it is an incidental job, requiring some post-processing.

Therefore, in the present embodiment, as shown in FIG. 4, the system startup management in which the standby server 20 performs a reboot after waiting for an external instruction when the business server 10 fails. It is characterized in that a unit 30 is provided. The operation in the present embodiment having this configuration will be described with reference to FIG.

When the business server 10 fails for some reason, the server management device 4 of the standby server 20 switches the standby mode to the business mode and automatically reboots when recognizing that the business server 10 has failed. In the present embodiment, the operation of the server management device 4 is suppressed and an external instruction is awaited (step 111). Then, when post-processing and the like of the incidental work are completed and an instruction for permitting the reboot of the standby server 20 is received, the state mode of the standby server 20 is switched from the standby mode to the business mode (step 112), and the embodiment is performed. 1 is executed (step 113). However, here
Since the standby server 20 is in the business mode, the server is booted from the shared disk 6.

As described above, according to the present embodiment, the standby server 20 is not automatically rebooted after the failure of the business server 10 is detected, but is instead waited for an external instruction (manual operation). can do.

It should be noted that the operation of the server management device 4 is suppressed.
A failure of the business server 10 may be detected and the state mode may be updated and before the reboot. Further, the conventional server management device 4 may be improved, or another means different from the server management device 4 may be provided.

Embodiment 3 In the present embodiment, when the standby server 20 is performing the business in place of the business server 10, the business can be executed again by the business server 10 because the business server 10 has been repaired. It is characterized by doing so. The system configuration in the present embodiment is the same as FIG. However, the system activation management unit 30 in the present embodiment
When receiving instructions from outside, both servers 10,
20 was rebooted. This business server 10
The process from the failure during the execution of the business to the return to the state of performing the business again after the repair will be described with reference to the flowcharts shown in FIGS.

In FIG. 6, when a failure occurs for some reason while the business server 10 is performing a business, the server management device 3 detects the failure (step 12).
1) Switch itself to failure mode (step 122) and bring down the system (step 123).
Note that a failure can be detected by using a watchdog timer or the like.

In FIG. 7, the system activation management unit 30
Upon recognizing from the server management device 3 that the business server 10 has failed (step 131), the system switches the standby mode to the business mode and automatically reboots the system (step 132). At this time, as described above, the logical switch 8 connects the shared disk 6 to the standby server 20 in accordance with the switching of the operation mode. Thereafter, the system activation management unit 30
Waits for an instruction to return to the original state (step 13
3). Then, when the system activation management unit 30 receives an instruction to return to the original state after the business server 10 is restored (step 133), the system startup management unit 30 instructs the server management apparatus 4 to change the standby server 20 from the business mode to the standby mode. Is switched to (step 134). Further, the server management apparatus 3 of the business server 10 is notified of the state mode switching and boot instruction via the server management apparatus 4 (step 13).
5). Thus, the system activation management unit 30 can cause the server management device 3 to switch from the standby mode to the business mode and to boot the business server 10. The status mode of the business server 10 has been changed from the failure mode to the standby mode during the operation check test after the restoration (steps 124 and 125). On the other hand, the standby server 2
0, the boot process shown in FIG. 3 of the first embodiment is executed (step 136). As a result, the business server 10 in the business mode starts the system from the shared disk 6, and the standby server 20 starts the system from the local disk 21 as described above.

As described above, according to the present embodiment, after the business server 10 is repaired, the original state, that is, the business server 10 is allowed to execute the business after waiting for an external instruction.
The standby server 20 can be returned to a state where it is used for another purpose.

Embodiment 4 FIG. In the present embodiment, as in the third embodiment, the business server 10 is restored and the business is executed by the business server 10 again. However, the business server 10 is automatically executed instead of an external instruction. It is characterized in that it is possible to detect the restoration of the server 10 and reboot. The system configuration in the present embodiment is the same as FIG. However, the system activation management unit 30 according to the present embodiment automatically detects the restoration of the business server 10 when the standby server 20 is performing a business on behalf of the business server 10, and detects both servers 1.
Reboot 0,20. FIG. 8 shows a flowchart of the processing in the standby server 20 in the present embodiment, which is basically the same as FIG. 7 except that the processing shown in step 133 is different. The business server 1
The processing at 0 may be the same as in FIG. Accordingly, only the processing related to step 137 corresponding to step 133 will be described in detail. After the standby server 20 in the business mode is rebooted (step 132), the standby server 20 operates as the business system, but the system startup management unit 30 Monitors the business mode of the business server 10 in parallel with the processing (step 137).

When the business server 10 recovers, the server management device 3 switches the business mode from the failure mode to the standby mode (step 124). When the system activation management unit 30 detects that the business server 10 has been switched to the standby mode (step 137), the standby server 2
0 is booted (steps 134 and 136), and the business server 10 is booted by the server management apparatus 3 by notifying the switching of the business mode and the boot instruction (steps 135 and 125). At this time, the logical switch 8 connects the shared disk 6 to the business server 10 in response to the switching of the business mode. Therefore, the business server 10 starts up the system from the shared disk 6 and the standby server 20 The system will be started from 21.

As described above, according to this embodiment, when the business server 10 fails, the business system is automatically switched, and when the repair of the business server 10 is detected, the business system is automatically switched. To return to the original state.

Embodiment 5 FIG. In each of the above embodiments, in the initial state, one server 10 is fixedly set to the business system and the other server 20 is fixedly set to the standby system.
0 is connected to the local system disk based on the rules described above. In the present embodiment, FIG.
Is characterized in that a local disk (business local system disk) 31 is also connected to the business server 10. The connection position of the local disk 31 in the business server 10 is connected to a position scanned after the position where the shared disk 6 is connected, just like the standby server 20 shown in FIG.

In such a configuration, when the business server 10 in the business mode boots, the shared disk 6 is connected to the business server 10, so the business system is started from the shared disk 6. If the business server 10 breaks down, the standby server 20 starts up the system from the shared disk 6 in place of the business server 10. Thereafter, when the business server 10 is repaired, the status mode is switched from the failure mode to the standby mode. Here, when the business server 10 is booted, since the shared disk 6 is connected to the standby server 20, the local disk 31 does not start up from the shared disk 6.
Will rise from.

According to the present embodiment, the business server 10
Also, by connecting the local disks 31 in a predetermined order, the local disks 31 can be used for another purpose when the business server 10 is on standby. The processing functions described in the above embodiments can be realized in the business server 10 by providing the business server 10 with the same configuration as the standby server 20 in addition to the system activation management unit. it can.

[0047]

According to the present invention, the local system disk is connected to the position scanned after the shared disk, so that the business server is operating normally and the standby server is connected to the dual system. When the server is on standby, the standby server can be used for another purpose by starting up from the local system disk. As a result, the standby server at the time of standby can be used effectively, so that efficient operation can be achieved.

Further, instead of automatically rebooting the standby server after detecting the failure of the business server, the standby server can be rebooted after receiving an external instruction.

Further, when the standby server is performing a business on behalf of the business server, after the business server is restored, it can be returned to its original state after waiting for an external instruction.

Further, when the standby server is performing a business on behalf of the business server, it can be automatically returned to the original state by detecting that the business server has been repaired.

Also, by connecting the local system disk to a position where the business server is scanned after the shared disk, the standby server is operating normally and the business server stands by as a dual system. When you are
By starting the business server from the local system disk, it can be used for another purpose.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram showing Embodiment 1 of a cold standby type duplex system according to the present invention.

FIG. 2 is a block diagram of a standby server according to the first embodiment.

FIG. 3 is a flowchart showing processing of a standby server according to the first embodiment.

FIG. 4 is an overall configuration diagram showing Embodiment 2 of a cold standby type duplex system according to the present invention.

FIG. 5 is a flowchart showing processing of a standby server according to the second embodiment.

FIG. 6 is a flowchart showing processing of a business server according to the third embodiment.

FIG. 7 is a flowchart showing processing of a standby server according to the third embodiment.

FIG. 8 is a flowchart showing processing of a standby server according to the fourth embodiment.

FIG. 9 is an overall configuration diagram showing Embodiment 5 of a cold standby type duplex system according to the present invention.

FIG. 10 is an overall configuration diagram of a conventional dual system.

[Explanation of symbols]

3, 4 server management device, 5 dedicated line, 6 shared disk, 7 disk system, 8 logical switch, 9 client, 10 business server, 20 standby server, 21, 31 local disk, 22 system bus, 23 CPU, 24 Memory, 25 BIOS, 2
6 PCI bus, 27 bus controller, 28, 29
SCSI board, 30 System startup management unit.

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06F 9/445 G06F 11/16-11/20 G06F 13/00 G06F 15/16-15/177

Claims (5)

(57) [Claims] 1. A server comprising a business system and a standby system, and having a boot mechanism that scans devices in a certain order until a boot device is searched at the time of system boot, and at the same time, one of the servers A shared disk that only allows access to the standby system; and wherein the standby server has a local system disk connected to a position scanned after the shared disk, and A cold standby type duplex system, wherein the system is started up from the local system disk when waiting. 2. The cold standby type secondary server according to claim 1, wherein said standby server has a system startup management unit for performing a reboot after receiving an external instruction when said business server fails. Heavy system. 3. When the standby server is performing a job on behalf of the business server, after the business server is repaired, a system activation for rebooting both servers after waiting for an external instruction. 2. The cold standby type duplex system according to claim 1, further comprising management means. 4. A system activation management means for automatically detecting restoration of the business server and rebooting both servers when the standby server is performing business on behalf of the business server. 2. The cold standby type dual system according to claim 1, wherein: 5. The service server has a service local system disk connected to a slot scanned after the slot to which the shared disk is connected, and is in standby as a dual system after restoration. 2. The cold standby type dual system according to claim 1, wherein the system is started up from the business local system disk.