Data replication method

• US 20090240974A1
• Filed: 08/06/2008
• Published: 09/24/2009
• Est. Priority Date: 03/18/2008
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Hara, Norihiro, Hamada, Shinji, Baba, Tsunehiko, Hanai, Tomohiro, Kushi, Kenji
• Granted Patent: US 7,865,763 B2
• Time in Patent Office: Days
• US Class Current: 714/4
• CPC Class Codes: G06F 11/2028 eliminating a faulty processor or activating a spare
• CPC Class Codes: G06F 11/2041 with more than one idle spare processing component
• CPC Class Codes: G06F 11/2048 where the redundant components share neither address space nor persistent storage
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups

Abstract

Provided is a data replication method capable of reducing the number of communication times when a processing result of an active system is replicated to a standby system. The data replication method, in which a first computer receives a first message containing a first processing request, and a plurality of second computers replicates the first message, includes the steps of: sending, by a third computer, the first message to the first computer and the second computers; sending, by each of the second computers, a message receive notification of the first message to the first computer; sending, by the first computer, after reception of the message receive notification from the second computers, the message receive notification of the first message to the third computer; and sending, by the first computer, a notification indicating that the first processing request becomes executable by the first computer to the second computers.

Claims

1. A data replication method used in a data replication system comprising:

a first computer equipped with a processor and a memory to execute an application; and

a second computer equipped with a processor and a memory and capable of taking over from an active system,

the first computer receiving a first message containing a first processing request,

a plurality of the second computers replicating the first message,

the method comprising the steps of;

sending, by a third computer, the first message to the first computer and the plurality of the second computers;

sending, by each of the second computers, a message receive notification of the first message to the first computer;

waiting, by the first computer, for the message receive notifications from all the second computers; and

receiving, by the first computer, the message receive notifications from the plurality of second computers, wherein

the step of receiving the message receive notifications from the plurality of the second computers comprises the steps of;

sending, by the first computer, after reception of the message receive notifications from the plurality of the second computers, the message receive notification of the first message to the third computer;

validating, by the first computer, after the reception of the message receive notifications from the plurality of the second computers, the first processing request to be executable by the first computer; and

notifying, by the first computer, after the reception of the message receive notifications from the plurality of the second computers, the second computers of validity of the first message.

2. The data replication method according to claim 1, further comprising the step of sending, by the third computer which has received the message receive notification of the first message, a second message containing a second processing request to the first computer and the plurality of the second computers, wherein

the step of notifying, by the first computer, the plurality of the second computers of the validity of the first message sends together the second message containing the second processing request received from the third computer, and the notification indicating the validity of the first message.

3. The data replication method according to claim 1, further comprising the steps of:

detecting, by the second computers, an occurrence of a fault in the first computer; and

taking over, by the second computers, when the fault occurs in the first computer, from the first computer, wherein

the step of taking over from the first computer when the fault occurs in the first computer includes the steps of;

setting one of the plurality of the second computers as a failover destination computer which takes over from the first computer, and controlling the failover destination computer to make an inquiry about message reception states to the other second computers;

collating, by the failover destination computer, the message reception states obtained as a result of making the inquiry with a message reception state held by the failover destination computer;

setting, by the failover destination computer, when a result of the collation shows that an inconsistent state is present in the message reception states between the failover destination computer and the other second computers, reception states of all the second computers to the identical reception state; and

taking over, by the failover destination computer, after setting of the reception states of all the second computers to the identical reception state, processing from the first computer where the fault has occurred.

4. The data replication method according to claim 3, wherein

the step of setting, by the failover destination computer, the reception states of all the second computers to the identical reception state includes:

a first determination step of determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message;

a first consistency guarantee step of instructing, by the failover destination computer, when a result of the determination shows the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, the second computers to delete the message;

a second determination step of determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message, and whether at least one of the second computers has received the message;

a second consistency guarantee step of replicating, when a result of the determination shows that the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, and that the at least one of the second computers has received the message, the message from the second computer which has received the message to all the second computers; and

a step of selecting at least one of the first consistency guarantee step and the second consistency guarantee step for each message.

5. The data replication method according to claim 4, wherein the step of selecting the at least one of the first consistency guarantee step and the second consistency guarantee step selects, by using at least one of processing contents of the message, an identifier of the third computer which has sent the message, time when the message is sent, and a resending timing of the message as a criterion, the first consistency guarantee step when the criterion is satisfied, and the second consistency guarantee step when the criterion is not satisfied.

6. The data replication method according to claim 5, wherein the criterion is set by a user of one of the first computer and the second computer.

7. The data replication method according to claim 1, wherein the step of notifying, by the first computer, the second computers of the validity of the first processing request controls the first computer to send together, when sending a third message to the second computers, the third message and the notification of the validity of the first processing request.

8. The data replication method according to claim 7, further comprising the step of updating, by the first computer, data in the memory by executing a message containing a third processing request, and notifying, by the first computer, the second computers of a result of the updating, wherein

the third message is a notification of the result of the updating.

9. The data replication method according to claim 8, further comprising the steps of:

detecting, by the second computers, an occurrence of a fault in the first computer; and

taking over, by the second computers, when the fault occurs in the first computer, from the first computer, wherein

the step of taking over from the first computer when the fault occurs in the first computer includes the steps of;

setting one of the plurality of the second computers as a failover destination computer which takes over from the first computer, and controlling the failover destination computer to make an inquiry about message reception states to the other second computers;

collating, by the failover destination computer, the message reception states obtained as a result of making the inquiry with a message reception state held by the failover destination computer;

setting, by the failover destination computer, when a result of the collation shows that an inconsistent state is present in the message reception states between the failover destination computer and the other second computers, reception states of all the second computers to the identical reception state; and

taking over, by the failover destination computer, after setting of the reception states of all the second computers to the identical reception state, processing from the first computer where the fault has occurred.

10. The data replication method according to claim 9, wherein

the step of setting, by the failover destination computer, the reception states of all the second computers to the identical reception state includes:

a first determination step of determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message;

a first consistency guarantee step of instructing, by the failover destination computer, when a result of the determination shows the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, the second computers to delete the message;

a second determination step of determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message, and whether at least one of the second computers has received the message;

a second consistency guarantee step of replicating, when a result of the determination shows that the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, and that the at least one of the second computers has received the message, the message from the second computer which has received the message to all the second computers; and

a step of selecting at least one of the first consistency guarantee step and the second consistency guarantee step for each message.

11. The data replication method according to claim 10, wherein the step of selecting the at least one of the first consistency guarantee step and the second consistency guarantee step selects, by using at least one of processing contents of the message, an identifier of the third computer which has sent the message, time when the message is sent, and a resending timing of the message as a criterion, the first consistency guarantee step when the criterion is satisfied, and the second consistency guarantee step when the criterion is not satisfied.

12. The data replication method according to claim 1, wherein one of the first computer and the second computer executes a process of the application only in the memory.

13. A data replication system, comprising:

a first computer equipped with a processor and a memory to execute an application; and

a second computer equipped with a processor and a memory and capable of taking over from an active system,

the first computer and a plurality of the second computers replicating a first message containing a first processing request sent from a third computer, wherein;

the third computer includes a third control unit for sending the first message to the first computer and the plurality of the second computers;

each of the second computers includes a second control unit for sending a message receive notification of the first message to the first computer; and

the first computer includes a first control unit for waiting for the message receive notifications from all the second computers, the first control unit including;

a first communication unit for sending, after reception of the message receive notifications from the plurality of the second computers, the message receive notification of the first message to the third computer;

a fix unit for validating, after the reception of the message receive notifications from the plurality of the second computers, the first processing request to be executable by the first computer; and

a second communication unit for notifying, after the reception of the message receive notifications from the plurality of the second computers, the second computers of validity of the first message.

14. The data replication system according to claim 13, wherein:

the third control unit of the third computer sends, after reception of the message receive notification of the first message, a second message containing a second processing request to the first computer and the second computer; and

the second communication unit sends together the second message containing the second processing request received from the third computer and the notification indicating the validity of the first message.

15. The data replication system according to claim 13, wherein the second communication unit sends together, when the first computer is controlled to send a third message to the second computers, the third message and the notification indicating the validity of the first message.

16. The data replication system according to claim 15, wherein:

the first computer includes a third communication unit for updating data in the memory by executing a message containing a third processing request, and for sending a result of the updating to the second computer; and

the third message is a notification of the result of the updating.

17. The data replication system according to claim 13, wherein

the second computer further includes:

a fault detection unit for detecting an occurrence of a fault in the first computer; and

a cluster control unit for taking over, when the fault detection unit detects the fault in the first computer, from the first computer, the cluster control unit being configured to;

set one of the plurality of the second computers as a failover destination computer which takes over from the first computer, and control the failover destination computer to make an inquiry about message reception states to the other second computers;

control the failover destination computer to collate the message reception states obtained as a result of making the inquiry with a message reception state held by the failover destination computer;

control the failover destination computer to set by a consistency guarantee unit of the cluster control unit, when a result of the collation shows that an inconsistent state is present in the message reception states between the failover destination computer and the other second computers, reception states of the second computers to the identical reception state; and

take over, after setting of the reception states of the second computer to the identical reception state, processing from the first computer where the fault has occurred.

18. The data replication system according to claim 17, wherein the consistency guarantee unit includes:

a first determination unit for determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message;

a first consistency guarantee unit for controlling, when a result of the determination shows the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, the failover destination computer to instruct the second computers to delete the message;

a second determination unit for determining whether the message reception states of the inconsistent state are reception states before at least one of the second computers receives the message, and whether at least one of the second computers has received the message;

a second consistency guarantee unit for replicating, when a result of the determination shows that the message reception states of the inconsistent state are the reception states before the at least one of the second computers receives the message, and that the at least one of the second computers has received the message, the message from the second computer which has received the message to all the second computers; and

a selection unit for selecting at least one of the first consistency guarantee unit and the second consistency guarantee unit for each message.

19. The data replication system according to claim 18, wherein the selection unit selects, by using at least one of processing contents of the message, an identifier of the third computer which has sent the message, time when the message is sent, and a resending timing of the message as a criterion, the first consistency guarantee unit when the criterion is satisfied, and the second consistency guarantee unit when the criterion is not satisfied.

20. The data replication system according to claim 13, wherein one of the first computer and the second computer executes a process of the application only in the memory.