JP2005055995A - Storage control method and server system with redundancy function - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize a redundancy function which is simple in structure and low in cost, the redundancy function executing recovery from a malfunction by switching an active system in a short period of time, when the malfunction occurs, and not stopping the services of a system as a whole due to maintenance or the like. <P>SOLUTION: N sets of active system servers and one standby system server are prepared, and a hard disk is subjected to redundancy for mirroring. When the active system is stopped due to maintenance or the like, the mirroring operation is interrupted, and the services of the active system are maintained while storing in one hard disk updated address information to which data are written along with the services, and the data are copied to a hard disk of a standby system from the other hard disk. After the copying is finished, the hard disk of active system and the hard disk of standby system are accorded with each other based on the updated address information to switch the standby system and the active system. When the malfunction occurs, the data are unconditionally copied to the standby system from the active system. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a storage control method for a server system, and more particularly to a server system having a redundancy function, in which an active system and a standby system are prepared, and the server system that operates the system without stopping. The present invention relates to a storage control method for a server system suitable for use.
[0002]
[Prior art]
In recent years, server systems have become a key technology in the information processing society, and high reliability has been demanded. As an example of redundancy technology in general server systems for ensuring reliability, multiple servers that satisfy the functions required for a single server are collected and operated as if they were a single system. In a cluster server system, such as a cluster server system, even if one server goes down, other servers in the cluster take over necessary resources and resume business in a short time. As a data takeover means in the cluster server system, there is a shared disk type, which realizes the takeover of data by sharing the extended disk array device between two servers.
[0003]
For example, in Patent Document 1 below, the state of each server can be stored in a shared disk so that a plurality of servers can be managed.
[0004]
As another redundancy technology, there is a fault tolerant server system in which a main hardware component is completely duplicated from a processor and a memory to a PCI module and a storage in one server.
[0005]
[Patent Document 1]
Japanese Patent Laid-Open No. 10-247911
[0006]
[Problems to be solved by the invention]
The system designer performs system design in consideration of the risk, availability, and cost of the system. In particular, taking over data assets in a server system when a failure occurs is a key point of a redundant system.
[0007]
Data takeover by sharing data between servers performed by the shared disk method in the cluster server system is non-stop and does not damage data assets. However, since the shared disk is generally expensive, the cost of system construction tends to increase. In addition, since data is shared, there is a system design restriction that the OS of each server must be the same. In addition, special care must be taken not to affect the use of shared data when performing maintenance such as updating the data in the disk by stopping the server system for a certain period of time for maintenance. There are also maintenance restrictions.
[0008]
On the other hand, in a fault-tolerant server system, complete hardware duplication is performed and high-reliability operation is guaranteed, but the cost of system construction becomes enormous.
[0009]
The present invention has been made to solve the above problems, and its purpose is related to storage control of a server system. An operation system and a standby system of a server system having a redundancy function are prepared, and a normal server stop is performed. In the required maintenance work, it is possible to continue the service by switching the active system, and in the event of a failure, the redundancy function that recovers the failure by switching the active system in a short time can be realized with a simple configuration at low cost. It is to provide a server system that can be used.
[0010]
[Means for Solving the Problems]
To achieve the above object, in the storage control method of the present invention, N active servers and one standby server are prepared, and the hard disks of the active server and the standby server are made redundant and mirrored. Keep it.
[0011]
When the active server is stopped for maintenance or the like, the redundancy operation of the active server is interrupted, and the service of the active server is continued with one of the mirroring hard disks.
[0012]
At this time, the update address of the hard disk for data writing accompanying the service of the active server is stored.
[0013]
On the other hand, data is copied from the other hard disk of the mirrored hard disk of the active server to the hard disk of the standby server.
[0014]
Then, when copying is completed and there is no data write accompanying the service of the active server, the data written with the service of the active server is changed to the standby system based on the stored update address. Copy to server.
[0015]
When copying is complete and all hard disk data on the active server and standby server match, switch the standby server to the active server, switch the active server to the standby server, and stop the active server Do maintenance, etc.
[0016]
When a failure occurs, data is unconditionally copied from the active server to the standby server.
[0017]
Each active server and standby server are installed in the same chassis, and hard disks are connected by a high-speed dedicated bus so that copying can be performed and relay of copy control to other servers can be switched. .
[0018]
As described above, according to the configuration of the server system of the present invention, the data in the HDD of the active server is completely copied in each state at the time of stoppage due to maintenance, etc., or when a failure occurs. A low-cost redundant configuration with only one standby server can be realized for the active server.
[0019]
Further, at the time of stoppage due to maintenance or the like, the redundancy operation of one HDD among the two or more redundant HDDs is stopped and used for data copy for server operation switching. Therefore, the active system and the standby system can be switched at high speed without stopping the active service. In addition, the HDD with the highest failure frequency can be continuously operated in the event of a failure due to the redundant configuration of the HDD. In the event of a failure other than the HDD, switching from the active server to the standby server is realized after the copy is completed. .
[0020]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment according to the present invention will be described with reference to FIGS. 1 to 6.
[0021]
[Configuration of server system with redundancy function]
First, the configuration of a server system having a redundancy function according to an embodiment of the present invention will be described with reference to FIG.
FIG. 1 is a configuration diagram of a server system having a redundancy function according to an embodiment of the present invention.
[0022]
As shown in FIG. 1, the server system of this embodiment is generally composed of a plurality of server units 2, a spare server unit 3 provided for backup, a management unit 4, and a LAN switch unit 5. .
[0023]
The number of server units 2 is determined at the time of system design in consideration of factors such as system reliability, availability, and performance. The spare server unit 3 is a server that is provisionally provided for the operation of the system. A plurality of spare server units 3 may be provided, but the number is usually smaller than the number of server units 2 that are operated.
[0024]
The management unit 4 is a part that manages operations that are related to each other, and is connected to the monitoring / control management unit 20 of each server unit 2 and spare server unit 3. The LAN switch unit 5 is connected to an external LAN 8 and relays the network with each server unit 2 and spare server unit 3.
[0025]
A maintenance terminal 6 and a client group 7 are connected to the LAN 8. The maintenance terminal 6 performs system maintenance and various settings, and the client group 7 uses a service provided by a server.
[0026]
The server unit 2 includes a monitoring / control management unit 20, a CPU 21, a memory 22, a LAN / IF 23, a device control unit 24, a disk controller unit 25, and HDDs 28 and 29. This configuration is the same for the spare server unit 3.
[0027]
The monitoring / control management unit 20 is a part that controls the server unit 2 by controlling the device control unit 24 and the disk controller unit 25 in response to monitoring of a server failure and a command from the management unit 4.
[0028]
The device control unit 24 is a part that controls each device of the server.
[0029]
The disk controller unit 25 includes a synchronization control unit 26 and a disk control unit 27, and is a part that controls the HDD.
[0030]
The synchronization control unit 26 of the disk controller unit 25 is a part that performs control for synchronizing data of the HDDs 28 and 29 of the server unit 2 and the HDDs 38 and 39 of the spare server unit 3.
[0031]
The disk control unit 27 of the disk controller unit 25 directly controls the HDDs 28 and 29, and receives an instruction from the HDD redundancy operation and the synchronization control unit 26 to synchronize data between the server unit 2 and the spare server unit 3. I / O control.
[0032]
The synchronization control unit 26 includes an update address memory unit 261 and a disk control monitoring unit 262. The update address memory unit 261 is a part that stores a write address when data on the hard disk is updated. The disk control monitoring unit 262 is a part that monitors accesses associated with hard disk services. These are components necessary to keep the data in the HDD of the server unit 2 and the HDD of the spare server unit 3 synchronized, and detailed operations will be described later.
[0033]
[Operation of server system with redundancy function]
Next, the operation of the server system having a redundancy function according to an embodiment of the present invention will be described with reference to FIGS.
(I) Operation at scheduled stop
First, the operation of the server system when the server system is stopped for maintenance / maintenance work will be described with reference to FIG.
[0034]
Hereinafter, when the maintenance person intends to perform maintenance / maintenance work on the server system, stopping the active system is referred to as “scheduled stop”.
[0035]
Further, considering the case where the server unit 2 is operated as an active system and the spare server unit 3 is operated as a standby system, they are referred to as an active server unit 2 and a standby system spare server unit 3, respectively.
FIG. 2 is a timing chart for explaining the system operation when the scheduled stop is performed.
[0036]
In the active server unit 2, in normal operation, the HDD 28 and the HDD 29 are made redundant, and read / write access is performed by the operation of mirroring (RAID 1). The operation of one standby standby server unit 3 is stopped.
[0037]
Here, it is assumed that a scheduled stop of the active server unit 2 is instructed from the maintenance terminal 6 connected to the LAN 8.
[0038]
This instruction is transmitted to the monitoring / control management unit 20 of the active server unit 2 and the monitoring / control management unit 30 of the standby standby server unit 3 by the management unit 4, and the disk controller unit 25 and the disk controller unit 35. The following operation is started by controlling.
[0039]
The disk control unit 27 of the disk controller unit 25 continues to perform read / write access of the active server unit 2 to one HDD 28 mirrored as shown in FIG.
[0040]
The disk control unit 27 of the disk controller unit 25 and the disk control unit 37 of the disk controller unit 35 are the same as both the HDD 38 and the HDD 39 of the standby spare server unit 3 with respect to the other mirrored HDD 29. Copy the data.
[0041]
Since the read / write access of the active server unit 2 is performed by the HDD 28, the operation of the active system continues at this stage. The update address memory unit 261 of the synchronization control unit 26 of the disk controller unit 25 stores the update address information 260a of the update data 1000 written in the HDD 28 after stopping the redundancy operation. At this time, over the data update of the same address, it is overwritten so that only the final data update information becomes valid.
[0042]
When the copy operation from one HDD 29 is completed, the monitoring / control management unit 30 of the standby standby server unit 3 activates the OS and application software for preparation for operation via the device control unit 34.
[0043]
Further, the disk control monitoring unit 262 of the synchronization control unit 26 and the disk control monitoring unit 362 of the synchronization control unit 36 stop after the redundancy operation of the HDD 28 is stopped based on the update address information 260a in the update address memory unit 261. The update data 1000 is copied to the HDD 29 of the same active server unit 2 and the HDDs 38 and 39 of the standby spare server unit 3.
[0044]
Here, even when the HDD 28 is copying the update data 1000 to another HDD, if there is a read / write access for an active service request, the update address information 260a in the update address memory unit 261 is newly added. Is remembered.
[0045]
In a situation where a large amount of data update due to an active service request continues for a long time and the copy operation continues for a long time, the copy operation by another HDD of the update data 1000 may not be completed easily.
[0046]
In such a case, a time instruction given in advance by the maintenance terminal 6 is given, and a service stop request is notified to the management unit 4 via the monitoring / control management unit 20 according to the conditions. Then, the management unit 4 instructs the LAN switch unit 5 to perform blocking control on the corresponding port, thereby avoiding a situation in which the update data copy operation continues for a long time.
[0047]
The monitoring / control management unit 20 uses the disk controller unit 25 to start the redundancy operation by capturing the moment when the copy of the update data 1000 from the HDD 28 is finished and the contents of all the HDDs are the same. A mirroring operation for the HDD 29 is started.
[0048]
In addition, the monitoring / control management unit 20 and the monitoring / control management unit 30 synchronize with each other via the management unit 4 and start the redundancy operation of the standby standby server unit 3 by the disk controller unit 35. The mirroring operation for the HDD 39 is started.
[0049]
That is, at this moment, all the HDDs of the active server unit 2 and the standby spare server unit 3 perform the mirroring operation with the same contents.
[0050]
Thereafter, the management unit 4 gives an instruction to the monitoring / control management unit 20 and the monitoring / control management unit 30 to switch the active system to the standby standby server unit 3.
[0051]
Thereafter, maintenance / maintenance work is performed on the server unit 2, but the update address memory unit 261 of the synchronization control unit 26 stores the update address information 260b of the data 1001 written in the HDD during the maintenance / maintenance work. Remember.
[0052]
Further, although the spare server unit 3 becomes a new active system, the update address memory unit 361 of the synchronization control unit 36 stores the update address information 260c of the data 1002 written in the HDD generated during this alternative operation period. To do.
[0053]
The update address information 260b of the data 1001 written in the HDD at the time of maintenance / maintenance work in these server units 2, and the update address information of the data 1002 written in the HDD generated during the alternative operation of the spare server unit 3 260c is used to recover HDD data during the following recovery procedure.
(II) Restoration operation after maintenance (Part 1)
Next, the operation of the server system when the maintenance / maintenance work is completed and the server system is restored will be described with reference to FIG.
FIG. 3 is a timing chart for explaining the system operation when the server system is restored (part 1).
[0054]
At the time of maintenance / maintenance work, the spare server unit 3 is the new active spare server unit 3 and the server unit 2 is the new standby server unit 2.
[0055]
In the new active spare server unit 3, the HDD 38 and the HDD 39 are made redundant, and read / write access is performed by a mirroring operation.
[0056]
Here, it is assumed that the maintenance / maintenance work of the new standby server unit 2 is finished and the maintenance terminal 6 connected to the LAN 8 gives an instruction to restore the server unit 2 after the maintenance / maintenance work.
[0057]
This instruction is transmitted by the management unit 4 to the monitoring / control management unit 20 of the new standby server unit 22 and the monitoring / control management unit 30 of the new standby system server unit 3, and the disk controller unit 25 and the disk controller. The following operation is started by controlling the unit 35.
[0058]
When instructed, the mirroring operation in the new active spare server unit 3 is stopped, and based on the updated address information 260c of the data 1002 written in the HDD generated during the alternative operation period of the spare server unit 3, Data is copied as additional data from the HDD 39 to the HDD 28 and the HDD 29 of the new standby server unit 2. In this specification, a copy operation in which new data is added to HDD data that has been subjected to maintenance / maintenance work is referred to as “merge copy”.
[0059]
On the other hand, the read / write access of the new active spare server unit 3 is continued to the other HDD 38 of the mirroring. Thus, the new active spare server unit 3 can continue the service without stopping.
[0060]
In addition, the update address memory unit 361 of the synchronization control unit 36 stores the address of the read / write access update data 1003 in the HDD 38 as the update address information 260d after the mirroring operation is stopped.
[0061]
At this stage, there are two types of update address information. For example, the update address information can be distinguished by attaching a flag to a table storing the update address information. Here, the update address information 260d of the read / write access update data in the HDD 38 during the recovery is distinguished as “update address information with flag”.
[0062]
When the merge copy based on the update address information 260c of the data 1002 written in the HDD generated during the alternative operation for the HDD 28 and HDD 29 of the new standby server unit 2 is completed, the new standby server unit 2 The monitoring / control management unit 20 activates the OS and application software via the device control unit 24 to prepare for operation.
[0063]
Then, based on the update address information with flag 260d, copying is performed from the HDD 38 to the HDD 28 and the HDD 29 of the new standby server unit 2.
[0064]
As a result, data that has been newly accessed to the new active standby server unit 3 during the recovery operation is reflected in the HDD 28 and HDD 29 of the new standby server unit 2.
[0065]
Further, in a situation where a large amount of data update due to a service request for a new operation system continues for a long time and the copy operation continues for a long time, the copy operation by another HDD of the update data 1003 may not be completed easily.
[0066]
In such a case, a time instruction given in advance by the maintenance terminal 6 is given, and a service stop request is notified to the management unit 4 via the monitoring / control management unit 30 according to the conditions. Then, the management unit 4 instructs the LAN switch unit 5 to perform blocking control on the corresponding port, thereby avoiding a situation where the copy operation of the update data 1003 continues for a long time.
[0067]
This is the same as the operation of temporarily suspending the service of the active server 2 at (I).
[0068]
Then, when the copy of the update data 1003 from the HDD 38 of the update data is completed and the contents of all the HDDs become the same, the monitoring / control management unit 20 performs the redundancy operation by the disk controller unit 25. The mirroring operation for the HDD 28 and the HDD 29 is started.
[0069]
Thereafter, the management unit 4 gives an instruction to the monitoring / control management unit 20 and the monitoring / control management unit 30 to switch the operation system to the server unit 2. Thus, the recovery operation ends, the server unit 2 returns to normal operation, and the spare server unit 3 returns to the standby system.
(III) Restoration operation after maintenance (Part 2)
Next, another operation of the server system when the maintenance / maintenance work is completed and the server system is restored will be described with reference to FIG.
FIG. 4 is a timing chart for explaining the system operation when the server system is restored (part 2).
[0070]
When this recovery operation is performed, the update address information of the data written in the HDD generated during the replacement operation period of the spare server unit 3 is previously stored, for example, because the replacement server period of the spare server unit 3 is long. This is the case when a certain number is exceeded.
[0071]
In this case, since it is not efficient to use the updated address information 260c of the data 1002 written in the HDD generated during the alternative operation period, the contents of the HDD of the new active spare server unit 3 are temporarily updated. A batch copy is made to the HDD of the standby server unit 2.
[0072]
Also in this case, as in (II), in the new active spare server unit 3, the HDD 38 and the HDD 39 are made redundant, and read / write access is made by the mirroring operation. Similarly to (II), it is assumed that the maintenance / maintenance work of the new standby server unit 2 is finished and the maintenance terminal 6 connected to the LAN 8 gives an instruction to restore the server unit 2 after the maintenance / maintenance work. .
[0073]
It is determined whether or not the update address information of data written in the HDD generated during the alternative operation period of the spare server unit 3 exceeds a predetermined number. When it does not exceed a certain number, it is restored in the same manner as the operation (II). When it exceeds a certain number, the restoration operation is performed as follows.
[0074]
First, the contents of the HDD 39 of the new active standby server unit 3 are collectively copied to the HDD 29 of the new standby server unit 2.
[0075]
During the batch copying, storing the update address for read / write access of the new active spare server unit 3 as flagged update address information 260d is the same as the operation (II).
[0076]
When the batch copy from the HDD 39 to the HDD 29 of the new standby server unit 2 is completed, the update address information 260b of the data 1001 written to the HDD during the maintenance operation described in (I) is then used. Data is merged and copied to the HDD 29. As a result, the content of the HDD 29 is a combination of the data content maintained by the new standby server unit 2 and the data updated by the read / write access in the service of the new active standby server unit 3.
[0077]
Thereafter, the entire contents of the HDD 29 are copied back to the HDD 28 so that the contents of the HDD 28 and the HDD 29 are the same.
[0078]
Thereafter, the operation of copying the contents of the read / write access update data 1003 during batch copy and switching the new standby server unit 2 to the active system based on the flagged update address 260d is the same as in (II). It is.
(IV) Operation at the time of failure
Next, the operation at the time of failure of the server system will be described with reference to FIG.
FIG. 5 is a timing chart for explaining the system operation when the server system fails.
[0079]
Here, a description will be given of a case where a failure other than the HDD occurs when the HDD of the active server unit 2 is redundant and the mirroring operation is performed. The operation when a failure occurs in the HDD is the same as the operation of recovering the system by normal RAID.
[0080]
For example, when the operational server unit 2 detects a failure such as a circuit abnormality or an increase in operating temperature, the monitoring / control management unit 20 gives an instruction to the disk controller unit 25 and the operational server unit 2 is mirrored. Copying is started from one HDD 29 to the HDD 38 and HDD 39 of the standby spare server unit 3.
[0081]
When the copying is completed, the monitoring / control management unit 30 of the standby standby server unit 3 activates the OS and application software to prepare for operation via the device control unit 34.
[0082]
Then, the standby spare server unit 3 is switched to the active system, and the mirroring operation is started by the HDD 38 and the HDD 39.
[0083]
[High-speed dedicated bus for storage control between servers]
Next, a high-speed dedicated bus for storage control between servers will be described with reference to FIG.
FIG. 6 is a configuration diagram focusing on the part that performs storage control between servers.
[0084]
At the beginning of the embodiment of the invention, the entire configuration of the server system having a redundancy function has been described with reference to FIG. 1, but here, a configuration in which a high-speed dedicated bus is provided as means for performing storage control between servers will be described. .
[0085]
Each server of the server system of the present embodiment can be implemented as a blade server (Blade Server). Here, the blade server refers to a server-dedicated machine configured by mounting necessary elements as a computer on a single board and connecting the necessary numbers.
[0086]
Here, as shown in FIG. 6, the server unit A 10, the server unit B 11, and the spare server unit 12 are connected via the management unit 9. Each of these server units is assumed to be mounted in one chassis as a blade server.
[0087]
Each server unit is connected by a high-speed dedicated bus by a disk control switching unit, and can receive data and control commands. The disk control switching unit of each server unit switches the HDD connection in response to an instruction from the device control unit.
[0088]
As an interface of the high-speed dedicated bus, a small computer system interface (SCSI), an AT attachment (ATA), a serial ATA, a fiber channel (Fibre Channel), or the like can be used.
[0089]
Here, as described in the procedure (I), the server unit B schedules the server unit A to remain operating.
[0090]
At this time, the HDD 104 of the server unit A receives the instruction from the device control unit 107 and continues the operation service. The data in one HDD 105 is copied to the HDD 124 and HDD 125 of the spare server unit 12.
[0091]
At this time, the disk control switching unit 106 is connected from the HDD 105 to the disk control switching unit 116 of the server unit B11 via the dedicated high-speed bus A13 so that copying can be performed. The disk control switching unit 116 of the server unit B11 is connected to the spare server unit 12 so as to relay copy control by the disk control switching unit 126 of the spare server unit 12 through the dedicated high-speed bus B14. Then, data is copied from the HDD 105 of the server unit A to the HDD 124 and the HDD 125 of the spare server unit 12 via the server unit B11.
[0092]
The disk control switching unit 116 of the server unit B11, which is the active system, connects the HDD 114 and the HDD 115 so as to perform a mirroring operation.
[0093]
The storage control is performed by each monitoring / control management unit exchanging information with other monitoring / control management units via the management unit 7. In FIG. 6, the disk control switching is schematically shown as a mechanical switch. However, in actuality, the disk control switching is performed electrically.
[0094]
As described above, in this embodiment, each server unit has a dedicated high-speed bus that connects adjacent servers in the same chassis, and is not subject to a copy operation associated with switching from the active system to the standby system. And bus relay control. As a result, even when all server units in the same chassis are connected, the pattern arrangement on the backplane in the chassis is facilitated, and a high-speed HDD control bus having a speed of several tens to several hundreds of MHz or more is provided. Electrical relaying is performed to enable long-distance transmission to a standby server when multiple servers are mounted in the same chassis.
[0095]
【The invention's effect】
According to the present invention, with regard to storage control of a server system, an operation system and a standby system of a server system having a redundancy function are prepared, and in maintenance work requiring a normal server stop, the service is continued by switching the operation system. Enables a redundant function that enables failure recovery by switching the operating system in a short time when a failure occurs, and can be realized with a server system with a simple configuration at low cost
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a server system having a redundancy function according to an embodiment of the present invention.
FIG. 2 is a timing chart for explaining a system operation when a scheduled stop is performed.
FIG. 3 is a timing chart for explaining the system operation when the server system is restored (part 1);
FIG. 4 is a timing chart for explaining the system operation when the server system is restored (part 2);
FIG. 5 is a timing chart for explaining the system operation when the server system fails.
FIG. 6 is a configuration diagram with emphasis on a portion that performs storage control between servers.
[Explanation of symbols]
1 ... Server system
2 ... Server part
3, 12 ... spare server part
4,9… Management Department
5 ... LAN switch
6. Maintenance terminal
7 ... Client County
8 ... LAN
10 ... Server A
11 ... Server part B
13 ... Exclusive highway bus A
14 ... Exclusive express bus B
20, 30, 101, 111, 121 ... monitoring / control management section
21, 31 ... CPU
22, 32 ... Memory
23, 33 ... LAN / IF
24, 34, 107, 117, 127 ... Device control unit
25, 35, 102, 112, 122... Disk controller section
26, 36 ... synchronization control unit
27, 37, 103, 113, 123... Disk control unit
28, 29, 38, 39, 104, 105, 114, 115, 124, 125 ... HDD
261, 361 ... Update address memory section
262, 362 ... Disk control monitoring unit
106, 116, 126... Disk control switching unit
260a-260d ... Update address information
1000 to 1003 ... Various data written to the HDD

Claims (12)

In a storage control method for a computer system having an active system and a standby system,
The operational system is a system having redundant means for redundantly writing the same data to two or more hard disks,
When the operational system is scheduled to stop,
(1) A procedure for interrupting the redundancy operation of the active system and continuing the service of the active system on one of the two or more hard disks;
(2) a procedure of copying from the other hard disk of the two or more hard disks of the active system to the hard disk of the standby system;
(3) In the procedure of (1), a procedure for storing the update address of the hard disk for data writing accompanying the service of the active system;
(4) When the copy of the procedure of (2) is completed and there is no data write accompanying the service of the operation system, the active system is based on the update address of the procedure of (3). A procedure for copying data written with the service to the standby system;
(5) When the copying of the procedure of (4) is completed and all the hard disk data of the active system and the standby system match, the standby system is switched to the active system, and the active system A computer system storage control method comprising: a procedure for switching a system to a standby system and performing a scheduled stop of the active system. The new operation system switched from the standby system is also a system having redundancy means for redundantly writing the same data to two or more hard disks,
When recovering a new standby system that has been switched from the active system,
(10) A procedure for storing an update address of a hard disk for data writing accompanying a service after switching to a new operation system;
(11) A procedure for interrupting the redundancy operation of the new active system and continuing the service of the new active system on one of the two or more hard disks;
(12) Merge copy from the other hard disk of the two or more hard disks of the new active system to the hard disk of the new standby system based on the update address of the procedure of (10);
(13) A procedure for storing an update address of a hard disk for data writing accompanying a service of the new active system during the merge copy of the procedure of (12),
(14) When the merge copy of the procedure of (12) is completed and there is no data write accompanying the service of the new operation system, the new copy is performed based on the update address of the procedure of (13). A procedure for copying the data written with the service of the active system to the new standby system;
(15) When the copying of the procedure of (14) is completed and the hard disk data of the new active system and the new standby system all match, the new standby system is switched to the active system, 2. The storage control method for a computer system according to claim 1, further comprising a procedure for switching the new active system to a standby system and performing a recovery operation of the new standby system. The new operation system switched from the standby system is also a system having redundancy means for redundantly writing the same data to two or more hard disks,
When recovering a new standby system that has been switched from the active system,
(20) storing a hard disk update address of data updated during maintenance of the new standby system;
(21) a procedure for storing the update address of the hard disk for data writing accompanying the service after switching to the new active system;
(22) a procedure for determining whether or not the number of update addresses in the procedure of (21) exceeds a predetermined number;
(24) A procedure for interrupting the redundant operation of the new active system and continuing the service of the new active system on one of the two or more hard disks;
(25) When the number of update addresses exceeds a predetermined number in the procedure of (22), the hard disk of the new standby system is switched from the other one of the two or more hard disks of the new active system. To one of the steps,
(26) a procedure for storing an update address of a hard disk for data writing accompanying the service of the new active system during the batch copying of the procedure of (25);
(27) Merge copying the data of the other hard disk to the hard disk collectively copied of the new standby system based on the update address updated at the time of maintenance of the procedure of (20);
(28) A procedure for batch copying the data of one hard disk merge-copied in the procedure of (27) to the other hard disk of the new standby system;
(29) When the merge copy of the procedure of (28) is completed and there is no data write accompanying the service of the new operation system, the new copy is performed based on the update address of the procedure of (26). A procedure for copying the data written with the service of the active system to the new standby system;
(30) When the copy of the procedure of (29) is completed and the hard disk data of the new active system and the new standby system all match, the new standby system is switched to the active system, 2. The storage control method for a computer system according to claim 1, further comprising a procedure for switching the new active system to a standby system and performing a recovery operation of the new standby system. 2. The process of stopping the service of the active system when the copy operation is stopped for a long time by the service of the active system during copying of the procedure of (4). A storage control method for the described computer system. The copy processing of (14) includes a process of stopping the service of the new operation system when the copy operation is stopped for a long time by the service of the new operation system. Item 3. A storage control method for a computer system according to Item 2. The copy processing of (29) includes a process of stopping the service of the new operation system when the copy operation is stopped for a long time by the service of the new operation system. Item 4. A storage control method for a computer system according to Item 3. In a server system having a redundancy function having an active server and a standby server,
The operational server has redundant means for redundantly writing the same data to two or more hard disks,
Further, the operational server
Means for interrupting the redundancy operation of the active server and continuing the service of the active server with one of the two or more hard disks when the active server is scheduled to stop;
Means for copying from the other hard disk of the two or more hard disks of the active server to the hard disk of the standby server;
Means for storing the update address of the hard disk for data writing accompanying the service of the active server;
When the copying to the hard disk of the standby server is completed and there is no data writing associated with the service of the active server, writing is performed with the service of the active server based on the stored update address. Means for copying the received data to the standby server,
When copying to the hard disk of the standby server is completed and the hard disk data of the active server and the standby server all match, the standby server is switched to the active server and the active server is waited A server system having a redundancy function, characterized by switching to an active server and performing a scheduled stop of the active server. The new active server switched from the standby server also has redundant means for redundantly writing the same data to two or more hard disks,
Furthermore, the new operational server
Means for storing the update address of the hard disk for data writing accompanying the service after switching to the new active server when performing the recovery operation of the new standby server switched from the active server;
Means for interrupting the redundancy operation of the new active server and continuing the service of the new active server on one of the two or more hard disks;
Means for performing a merge copy from the other hard disk of the two or more hard disks of the new active server to the hard disk of the new standby server based on the update address;
Means for storing, as a flagged update address, an update address of a hard disk for data writing accompanying the service of the new active server during the merge copy;
When the merge copy is completed and there is no data writing associated with the service of the new active server, the data written with the service of the new active server is determined based on the flagged update address. And means for copying to the hard disk of the new standby server,
When copying to the hard disk of the new standby server is completed and the hard disk data of the new active server and the new standby server all match, the new standby server is switched to the active server, 8. The server system having a redundancy function according to claim 7, wherein the new active server is switched to a standby server and the recovery operation of the new standby server is performed. The new active server switched from the standby server is also a system having redundant means for redundantly writing the same data to two or more hard disks,
The new operational server is
Means for storing the update address of the hard disk for data writing accompanying the service after switching to the new active server when performing the recovery operation of the new standby server switched from the active server;
Means for determining whether the number of the update addresses exceeds a predetermined number;
Means for interrupting the redundancy operation of the new active server and continuing the service of the new active server with one of the two or more hard disks;
Means for batch copying from the other hard disk of the two or more hard disks of the new active server to one of the hard disks of the new standby server when the number of the update addresses exceeds a predetermined number;
Means for storing, as a flagged update address, the update address of the hard disk for data writing accompanying the service of the new active server during batch copying;
The new standby server is:
Means for storing a hard disk update address of data updated during maintenance of the new standby server as a maintenance update address;
Means for performing a merge copy on the hard disk collectively copied from the new active server, based on the maintenance update address updated during the maintenance, on the data of the other hard disk;
Means for collectively copying the data of one of the hard disks merged and copied to the other hard disk of the new standby server;
Furthermore, the new operational server
When the merge copy performed at the new standby server is completed and there is no data write accompanying the service of the new active server, the service of the new active server is based on the flagged update address. Means for copying the data written along with the new standby server,
When copying to the new standby server is completed and the hard disk data of the new active server and the new standby server all match, the new standby server is switched to the active server, and the new operation 8. The server system having a redundancy function according to claim 7, wherein a recovery operation of the new standby server is performed by switching the standby server to a standby server. 8. The server system having a redundancy function according to claim 7, wherein there are N active servers (N ≧ 1) and one standby server. When a failure other than the hard disk of the active server occurs, means for copying from the hard disk of the active server where the failure has occurred to the hard disk of the standby server,
8. The server system having a redundancy function according to claim 7, wherein when the copying is completed, the standby server is switched to an active server. The active server and the standby server are mounted on the same chassis,
The hard disk of the adjacent active server and the hard disk of the standby server are connected to the hard disk of the other adjacent active server and the hard disk of the standby server via a high-speed dedicated bus,
When copying is required between the hard disk of the active server and the hard disk of the standby server,
8. The copy control operation between each active server and standby server is determined, and control is performed by switching whether to perform copying or to perform copy control relay using the high-speed dedicated bus. A server system having the redundancy function according to claim 9.

Images