CN103875229A

CN103875229A - Asynchronous replication method, device and system

Info

Publication number: CN103875229A
Application number: CN201380003469.4A
Authority: CN
Inventors: 项锐; 张伦
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2013-12-02
Filing date: 2013-12-02
Publication date: 2014-06-18
Anticipated expiration: 2033-12-02
Also published as: CN103875229B; WO2015081473A1

Abstract

The embodiment of the invention provides an asynchronous replication method, a device and a system. The method comprises that a main data center receives a write request that is emitted by a user device and comprises to-be-written data; object metadata and object data corresponding to the to-be-written data are generated and stored, wherein the object metadata comprises an object identity and the object data is equal to the to-be-written data; a first synchronous message is sent to a backup data center and comprises the object metadata; a synchronous request emitted by the backup data center is received and comprises the object identity; and a second synchronous message is sent to the backup data center according to the synchronous message and comprises the object data corresponding to the object identity contained in the synchronous message. The method overcomes the problem that time of object access in the backup data center by a user is uncontrollable.

Description

Asynchronous replication method, equipment and system

Technical field

The embodiment of the present invention relates to memory technology, relates in particular to asynchronous replication method, equipment and system between a kind of many data centers based on object storage.

Background technology

Object stores service, is generally applied to large-scale data storage service, as services such as internet mass content, backups.To liking the elementary cell of object storage, comprise object metadata and object data.(Multiple Data Center is called for short: MDC) provide data Replica strategy and data redundancy strategy flexibly, ensure the continuity of service operation in many data centers.The synchronous asynchronous replication mode that conventionally adopts of data between data center: data are write successfully and returned successfully to user at master data center (DC1), and then background synchronization is to standby data center (DC2).

In prior art, object data is when between data center, asynchronous replication is synchronous, DC1 is stored in local storage engines according to the Raw Data Generation object metadata of user's write request and object data two parts, object data is original data to be written, object metadata comprises the various attributes of object data: object name (Object Name), object identity (Object ID), the Object Creation time (Object Create Time), object modification time Object Modify Time), the information such as object data verification and MD5, if wherein object data verification and MD5 change, the Object Creation time is also ensued variation, generate task according to user's write request original record, this task comprises object name, object identity, task creation time (Task Create Time), task modification time (Task Modify Time), authority information etc., and task is passed to DC2 one by one, carry out in order more all user's write requests one time at DC2.DC1 carries out object synchronization when the DC2 according to task, first object data is synchronized to DC2, then object metadata is synchronized to DC2, and after object data and metadata are all successfully transferred to DC2, user could conduct interviews to the object on DC2.In synchronizing cycle, if user has carried out repeatedly write request to same target, DC1 is also correspondingly repeatedly synchronized to DC2 by object.

Based on above-mentioned prior art, object asynchronous replication between data center, when object data larger, as 5 terabytes (Tera Byte, be called for short: TB) time, DC1 is larger while carrying out object data propagation delay time to DC2, makes user see that at DC2 the time of object is uncontrollable.

Summary of the invention

The embodiment of the present invention provides a kind of asynchronous replication method, equipment and system, while carrying out object asynchronous replication between the Yu Bei data center of master data center to solve in prior art lock in time uncontrollable and invalid synchronous problem.

First aspect, the embodiment of the present invention provides a kind of asynchronous replication method, comprising:

Master data center receives the write request that subscriber equipment sends, and comprises data to be written in described write request;

Described master data center generates object metadata and the object data corresponding with described data to be written, and stores described object data and described object metadata, and described object metadata comprises object identity, and described object data is identical with described data to be written;

In described master data, mind-set sends the first synchronization message for data center, in described the first synchronization message, comprises described object metadata;

Described master data center receives the synchronization request that described standby data center sends, and comprises described object identity in described synchronization request;

Described master data center sends the second synchronization message according to described synchronization request to described standby data center, comprises object data corresponding to object identity comprising with described synchronization request in described the second synchronization message.

In the possible implementation of the first of first aspect, in described master data, mind-set sends the first synchronization message for data center, before comprising described object metadata, also comprises in described the first synchronization message:

Described master data center generates the synchronous task corresponding with described object metadata, comprises described object identity, task creation time and task modification time in described synchronous task;

Described master data center, according to the task modification time of each synchronous task, merges the synchronous task that comprises same object mark;

In described master data, mind-set sends the first synchronization message for data center, comprising:

Described master data center is according to described task creation time and synchronizing cycle determining whether to start to carry out synchronous task;

Carry out synchronous task if start, described master data center sends the first synchronization message according to the synchronous task after merging to described standby data center.

In conjunction with the possible implementation of the first of first aspect, in the possible implementation of the second of first aspect, described master data center, according to the task modification time of each synchronous task, merges the synchronous task that comprises same object mark, comprising:

Described master data center obtains the each synchronous task that comprises same object mark;

Described master data center relatively comprises the task modification time of each synchronous task of same object mark;

Described master data center obtains the synchronous task after merging according to the up-to-date synchronous task of task modification time in each synchronous task, and deletes the older synchronous task of other task modification times.

In conjunction with the first or the possible implementation of the second of first aspect, in the third possible implementation of first aspect, described master data center, also comprises according to the synchronous task after merging after described standby data center sends the first synchronization message:

Described master data center receives the first sync response that described standby data center sends, and described the first sync response comprises described object identity;

If described the first sync response is success response, described synchronous task is deleted at described master data center.

In conjunction with first to the 3rd any possible implementation of first aspect or first aspect, in the 4th kind of possible implementation of first aspect, before in described master data, mind-set sends the first synchronization message for data center, also comprise:

Determine whether described object metadata exists;

If exist, carry out mind-set in described master data and send for data center the step of the first synchronization message;

If do not exist, in described master data, described in mind-set, standby data center sends the deletion object task that comprises described object identity and task modification time, so that object identical with described object identity and before task modification time is deleted by described standby data center.

In conjunction with first to the 4th any possible implementation of first aspect or first aspect, in the 5th kind of possible implementation of first aspect, described master data center receives the synchronization request that described standby data center sends, after comprising described object identity in described synchronization request, described master data center sends the second synchronization message according to described synchronization request to described standby data center, before comprising the object data that the object identity that comprises with described synchronization request is corresponding in described the second synchronization message, also comprise:

Described synchronization request also comprises the object modification time;

Described master data center determines according to described object identity whether native object metadata exists;

If exist, described master data center judges that whether the described object modification time is identical;

If identical, read described object data according to described native object metadata;

Otherwise described in mind-set, the transmission of standby data center comprises the non-existent sync response of object in described master data.

In conjunction with first to the 5th any possible implementation of first aspect or first aspect, in the 6th kind of possible implementation of first aspect, also comprise: described master data center receives the request of repairing of reading that standby data center sends, described in read to comprise object identity and object modification time in the request of repairing;

Described in the basis of described master data center, read the request of repairing and send the second synchronization message to described standby data center, in described the second synchronization message, comprise with described and read to repair the object identity that comprises of request and described object data corresponding to object modification time, so that the object data that described standby data center comprises described the second synchronization message returns to user, preserve object data that described the second synchronization message comprises to standby data center, and described in mark, the state of native object metadata is that data pass.

In conjunction with first to the 6th any possible implementation of first aspect or first aspect, in the 7th kind of possible implementation of first aspect, described object metadata is transmitted on first passage, and described data to be written are transmitted on second channel.

Second aspect, the embodiment of the present invention provides a kind of asynchronous replication method, comprising:

Standby data center receives the first synchronization message that master data center sends, and in described the first synchronization message, comprises object metadata;

Described standby data center sends synchronization request to described master data center, comprise described object identity in described synchronization request;

Described standby data center receives the second synchronization message that described master data center sends, and comprises object data corresponding to object identity comprising with described synchronization request in described the second synchronization message.

In the possible implementation of the first of second aspect, standby data center receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, described standby data center sends synchronization request to described master data center, before comprising described object identity in described synchronization request, also comprise:

Described object metadata comprises described object identity, check sum, Object Creation time and object modification time;

The object identity that described standby data center comprises according to described object metadata determines whether native object metadata exists;

If do not exist, directly store described object metadata and be the native object metadata after synchronous, and the state that identifies described native object metadata after synchronous is that data do not pass;

If exist, the check sum that the check sum that the more described object metadata of described standby data center comprises and described native object meta-data pack contain;

If described check sum is identical, the state of the native object metadata after native object metadata described in the described object modification time synchronized comprising according to described object metadata for data center, and mark is synchronous is that data pass;

If described check sum difference, native object metadata described in the Object Creation time synchronized that described standby data center comprises according to described object metadata, and the state that identifies described native object metadata after synchronous is that data do not pass;

Described standby data center sends the first sync response to described master data center.

In conjunction with the possible implementation of the first of second aspect, in the possible implementation of the second of second aspect, if described check sum is identical, native object metadata described in the object modification time synchronized that described standby data center comprises according to described object metadata, and the state of the native object metadata of mark after synchronous is that data pass, comprising:

Described standby data center determines that the object modification time whether object modification time that described object metadata comprises contain than described native object meta-data pack is old;

If so, described standby data center directly returns to the first synchronous success response to described master data center;

Otherwise described standby data center upgrades described native object metadata according to described object metadata, generates the data synchronous task that comprises object identity, object modification time, and returns to the first sync response to described master data center;

If described check sum difference, native object metadata described in the Object Creation time synchronized that described standby data center comprises according to described object metadata, and the state that identifies described native object metadata after synchronous is that data do not pass, and comprising:

Described standby data center determines that the Object Creation time whether Object Creation time that described object metadata comprises contain than described native object meta-data pack is new;

If so, described standby data center upgrades described native object metadata according to described object metadata, and returns to the first sync response to described master data center;

Otherwise described standby data center directly returns to the first synchronous success response to described master data center.

In conjunction with the possible implementation of the second of second aspect, in the third possible implementation of second aspect, described standby data center sends synchronization request to described master data center, comprise described object identity in described synchronization request, comprising:

The data synchronous task that comprises object identity, object modification time described in described standby data center basis sends synchronization request to described master data center, comprise described object identity in described synchronization request.

In conjunction with the third possible implementation of second aspect, in the 4th kind of possible implementation of second aspect, the data synchronous task that comprises object identity, object modification time described in described standby data center basis sends synchronization request to described master data center, before comprising described object identity in described synchronization request, also comprise:

The object identity that described standby data center comprises according to described data synchronous task determines whether native object metadata exists;

If do not exist, cancel and delete described data synchronous task;

If exist, determine that whether the object modification time of described native object metadata is identical with the object modification time that described data synchronous task comprises;

If not identical, cancel and delete described data synchronous task;

If identical, determine whether the state of described native object metadata is that data do not pass;

If not data do not pass, cancel and delete described data synchronous task;

If data do not pass, described data synchronous task is divided into N segment data synchronous task by described standby data center, and N is positive integer; Described standby data center generates N synchronization request according to N segment data synchronous task respectively, and the each segment data synchronous task of mark is synchronous for not completing.

In conjunction with first to the 4th any possible implementation of second aspect or second aspect, in the 5th kind of possible implementation of second aspect, described standby data center receives the second synchronization message that described master data center sends, in described the second synchronization message, comprise described object data, comprising:

Described standby data center, receives the second synchronization message that described master data center sends according to each synchronization request;

If the second synchronization message comprises described object data, store described object data, and the data synchronous task of the corresponding section of mark is synchronous for completing;

If the second synchronization message does not comprise described object data, again send synchronization request according to the data synchronous task of corresponding section;

If N segment data synchronous task is all labeled as data and passes, finish described data synchronous task, the state of revising described local metadata is that data pass, and deletes described data synchronous task.

In conjunction with first to the 5th any possible implementation of second aspect or second aspect, in the 6th kind of possible implementation of second aspect, described standby data center is after described master data center sends synchronization request, described standby data center also comprises before receiving the second synchronization message of described master data center transmission:

Described standby data center receives the sync response at described master data center;

Do not exist if described sync response is object, cancel and delete described data synchronous task.

In conjunction with first to the 6th any possible implementation of second aspect or second aspect, in the 7th kind of possible implementation of second aspect, also comprise:

Described standby data center receives the deletion object task that comprises described object identity and object modification time that described master data center sends;

The object identity that described standby data center comprises according to described deletion object task determines whether native object metadata exists;

If exist, the object modification time whether the object modification time that more described deletion object task comprises contains than described native object meta-data pack is old;

If not, delete described native object metadata and the object data corresponding with described native object metadata, and send deletion response to described master data center;

Otherwise cancellation task, and send deletion success response to described master data center.

In conjunction with first to the 7th any possible implementation of second aspect or second aspect, in the 8th kind of possible implementation of second aspect, also comprise:

Described standby data center receives the read request of subscriber equipment, and described user's read request packet is containing object identity;

Described standby data center determines corresponding native object metadata according to described object identity, and determines according to the state of described native object metadata whether object data has passed and/or whether damaged;

If described object data does not pass or damages, described standby data center sends the request of repairing of reading to described master data center, describedly read the request of repairing and comprise object identity and object modification time, so that described master data center according to described in read the request of repairing and send the second synchronization message, described the second synchronization message comprise with described read to repair the object identity that comprises of request and object data corresponding to object modification time;

Described standby data center, the object data that described the second synchronization message is comprised returns to described subscriber equipment;

Described standby data center, the state of native object metadata described in object data that described the second synchronization message comprises mark of storing is that data pass.

In conjunction with first to the 8th any possible implementation of second aspect or second aspect, in the 9th kind of possible implementation of second aspect, also comprise:

Described object metadata is transmitted on first passage, and described object data transmits on second channel.

The third aspect, the embodiment of the present invention provides a kind of master data center, comprising:

Memory module, the write request sending for receiving subscriber equipment, comprises data to be written in described write request; And

Generate object metadata and the object data corresponding with described data to be written, and store described object data and described object metadata, described object metadata comprises object identity, and described object data is identical with described data to be written;

The first synchronized transmission module, for sending the first synchronization message to standby data center, comprises described object metadata in described the first synchronization message;

The second synchronized transmission module, the synchronization request sending for receiving described standby data center, comprises described object identity in described synchronization request; And

Send the second synchronization message according to described synchronization request to described standby data center, in described the second synchronization message, comprise object data corresponding to object identity comprising with described synchronization request.

In the possible implementation of the first of the third aspect, described the first synchronized transmission module, specifically for:

Sending the first synchronization message to standby data center, before comprising described object metadata in described the first synchronization message, generate the synchronous task corresponding with described object metadata, in described synchronous task, comprise described object identity, task creation time and task modification time;

According to the task modification time of each synchronous task, the synchronous task that comprises same object mark is merged;

According to described task creation time and synchronizing cycle determining whether to start to carry out synchronous task;

Carry out synchronous task if start, send the first synchronization message according to the synchronous task after merging to described standby data center.

In conjunction with the possible implementation of the first of the third aspect, in the possible implementation of the second of the third aspect, described the first synchronized transmission module, specifically for:

Obtain the each synchronous task that comprises same object mark;

Relatively comprise the task modification time of each synchronous task of same object mark;

Obtain the synchronous task after merging according to the up-to-date synchronous task of task modification time in each synchronous task, and delete the older synchronous task of other task modification times.

In conjunction with the first or the possible implementation of the second of the third aspect, in the third possible implementation of the third aspect, described the first synchronized transmission module according to the synchronous task after merging after described standby data center sends the first synchronization message, also for:

Receive the first sync response that described standby data center sends, described the first sync response comprises described object identity;

If described the first sync response is success response, delete described synchronous task.

In conjunction with first to the 3rd any possible implementation of the third aspect or the third aspect, before described the first synchronized transmission module sends the first synchronization message to standby data center, also for:

Determine whether described object metadata exists;

If exist, carry out the step that sends the first synchronization message to standby data center;

If do not exist, sends to described standby data center the deletion object task that comprises described object identity and task modification time so that the deletion of described standby data center is identical with described object identity and with task modification time before object.

In conjunction with first to the 4th any possible implementation of the third aspect or the third aspect, in the 5th kind of possible implementation of the third aspect, described the second synchronized transmission module receives the synchronization request that described standby data center sends, after comprising described object identity in described synchronization request, described the second synchronized transmission module sends the second synchronization message according to described synchronization request to described standby data center, before comprising the object data that the object identity that comprises with described synchronization request is corresponding in described the second synchronization message, also for:

Described synchronization request also comprises the object modification time;

Determine according to described object identity whether native object metadata exists;

If exist, judge that whether the described object modification time is identical;

Otherwise, send and comprise the non-existent sync response of object to described standby data center.

In conjunction with first to the 5th any possible implementation of the third aspect or the third aspect, in the 6th kind of possible implementation of the third aspect, described the second synchronized transmission module, also for:

Receive the request of repairing of reading that standby data center sends, described in read to comprise object identity and object modification time in the request of repairing; And

Read the request of repairing and send the second synchronization message to described standby data center according to described, in described the second synchronization message, comprise with described and read to repair the object identity that comprises of request and described object data corresponding to object modification time, so that the object data that described standby data center comprises described the second synchronization message returns to user, preserve object data that described the second synchronization message comprises to standby data center, and described in mark, the state of native object metadata is that data pass.

In conjunction with first to the 6th any possible implementation of the third aspect or the third aspect, in the 7th kind of possible implementation of the third aspect,

Described object metadata is transmitted on first passage, and described data to be written are transmitted on second channel.

Fourth aspect, the embodiment of the present invention provides a kind of standby data center, comprising:

The first synchronous receiver module, the first synchronization message sending for receiving master data center, comprises object metadata in described the first synchronization message;

The second synchronous receiver module, for send synchronization request to described master data center, comprises described object identity in described synchronization request; And

Receive the second synchronization message that described master data center sends, in described the second synchronization message, comprise object data corresponding to object identity comprising with described synchronization request.

In the possible implementation of the first of fourth aspect, the described first synchronous receiver module receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, the described second synchronous receiver module sends synchronization request to described master data center, before comprising described object identity in described synchronization request, the described first synchronous receiver module, also for:

The object identity comprising according to described object metadata determines whether native object metadata exists;

If described check sum is identical, native object metadata described in the object modification time synchronized comprising according to described object metadata, and the state of the native object metadata of mark after synchronous is that data pass;

If described check sum difference, native object metadata described in the Object Creation time synchronized comprising according to described object metadata, and the state that identifies described native object metadata after synchronous is that data do not pass;

Send the first sync response to described master data center.

In conjunction with the possible implementation of the first of fourth aspect, in the possible implementation of the second of fourth aspect, the described first synchronous receiver module, specifically for:

If described check sum is identical, determine that the object modification time whether object modification time that described object metadata comprises contain than described native object meta-data pack is old;

If so, directly return to the first synchronous success response to described master data center;

Otherwise, upgrade described native object metadata according to described object metadata, generate the data synchronous task that comprises object identity, object modification time, and return to the first sync response to described master data center;

If described check sum difference, determines that the Object Creation time whether Object Creation time that described object metadata comprises contain than described native object meta-data pack is new;

If so, upgrade described native object metadata according to described object metadata, and return to the first sync response to described master data center;

Otherwise, directly return to the first synchronous success response to described master data center.

In conjunction with the possible implementation of the second of fourth aspect, in the third possible implementation of fourth aspect, the described second synchronous receiver module, specifically for:

Send synchronization request according to the described data synchronous task that comprises object identity, object modification time to described master data center, in described synchronization request, comprise described object identity.

In conjunction with the third possible implementation of fourth aspect, in the 4th kind of possible implementation of fourth aspect, the described second synchronous receiver module, send synchronization request to described master data center according to the described data synchronous task that comprises object identity, object modification time, before comprising described object identity in described synchronization request, also for:

The object identity comprising according to described data synchronous task determines whether native object metadata exists;

If do not exist, cancel and delete described data synchronous task;

If exist, determine that whether the object modification time of described native object metadata is identical with the object modification time comprising in described data synchronous task;

If not identical, cancel and delete described data synchronous task;

If not data do not pass, cancel and delete described data synchronous task;

In conjunction with first to the 4th any possible implementation of fourth aspect or fourth aspect, the described second synchronous receiver module, specifically for:

Receive the second synchronization message that described master data center sends according to each synchronization request;

In conjunction with first to the 5th any possible implementation of fourth aspect or fourth aspect, in the 6th kind of possible implementation of fourth aspect, the described second synchronous receiver module is after described master data center sends synchronization request, receive described master data center send the second synchronization message before, also for:

Receive the sync response at described master data center;

In conjunction with first to the 6th any possible implementation of fourth aspect or fourth aspect, in the 7th kind of possible implementation of fourth aspect, the described first synchronous receiver module, also for:

Receive the deletion object task that comprises described object identity and object modification time that described master data center sends;

The object identity comprising according to described deletion object task determines whether native object metadata exists;

If exist, whether the object modification time that more described deletion object task comprises is older than the object modification time of described native object metadata;

In conjunction with first to the 7th any possible implementation of fourth aspect or fourth aspect, in the 8th kind of possible implementation of fourth aspect, the described second synchronous receiver module, also for:

Receive the read request of subscriber equipment, described user's read request packet is containing object identity;

Determine corresponding native object metadata according to described object identity, and determine according to the state of described native object metadata whether object data has passed and/or whether damaged;

If described object data does not pass or damages, send the request of repairing of reading to described master data center, describedly read the request of repairing and comprise object identity and object modification time, so that described master data center according to described in read the request of repairing and send the second synchronization message, described the second synchronization message comprise with described read to repair the object identity that comprises of request and object data corresponding to object modification time;

The object data that described the second synchronization message is comprised returns to described subscriber equipment;

Storing the state of native object metadata described in object data that described the second synchronization message comprises mark is that data pass.

In conjunction with first to the 8th any possible implementation of fourth aspect or fourth aspect, in the 9th kind of possible implementation of fourth aspect, described object metadata is transmitted on first passage, and described object data transmits on second channel.

The 5th aspect, the embodiment of the present invention provides a kind of master data center, comprise first processor and process with described first the first memory being connected, in wherein said first memory, store batch processing code, described first processor calls described program code, for carrying out the asynchronous replication method described in first of first aspect or first aspect to the 6th any possible implementation.

The 6th aspect, the embodiment of the present invention provides a kind of standby data center, comprise the second processor and process with described second the second memory being connected, in wherein said second memory, store batch processing code, described the second processor calls described program code, for carrying out the asynchronous replication method described in first of second aspect or second aspect to the 8th any possible implementation.

The 7th aspect, the embodiment of the present invention provides a kind of asynchronous replication system, comprising: the master data center described in the 5th aspect, and standby data center described in the 6th aspect.

Based on such scheme, asynchronous replication method, equipment and system that the embodiment of the present invention provides, by master data center, object metadata and object data are separately synchronized to standby data center, make object metadata can be synchronized to rapidly standby data center, solve user and carried out uncontrollable problem of object accesses time in standby data center, by metadata synchronization task merging, overcome user in synchronizing cycle same object multi-pass operation has easily been caused to the problem that invalid data is synchronous simultaneously.

Brief description of the drawings

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is some embodiments of the present invention, for those of ordinary skill in the art, can also obtain according to these accompanying drawings other accompanying drawing.

Fig. 1 is the flow chart of asynchronous replication embodiment of the method one provided by the invention;

Fig. 2 is the flow chart of asynchronous replication embodiment of the method two provided by the invention;

Fig. 3 is the information interaction figure of asynchronous replication method specific embodiment one provided by the invention;

Fig. 4 is the information interaction figure of asynchronous replication method specific embodiment two provided by the invention;

Fig. 5 is the information interaction figure of asynchronous replication method specific embodiment three provided by the invention;

Fig. 6 is the information interaction figure of asynchronous replication method specific embodiment four provided by the invention;

Fig. 7 is the structural representation of master data center implementation example one provided by the invention;

Fig. 8 is the structural representation of the standby embodiment of data center mono-provided by the invention;

Fig. 9 is the structural representation of master data center implementation example two provided by the invention;

Figure 10 is the structural representation of the standby embodiment of data center bis-provided by the invention;

Figure 11 is the structural representation of asynchronous replication system embodiment provided by the invention.

Embodiment

For making object, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, instead of whole embodiment.

Fig. 1 is the flow chart of asynchronous replication embodiment of the method one provided by the invention, and as shown in Figure 1, the method embodiment, comprising:

S101, master data center receive the write request that subscriber equipment sends, and comprise data to be written in described write request;

S102, described master data center generate object metadata and the object data corresponding with described data to be written, and store described object data and described object metadata, described object metadata comprises object identity, and described object data is identical with described data to be written;

Wherein, described data corresponding objects data division to be written, the community set that described object metadata is object data, comprises object name, object identity, Object Creation time, object modification time, check sum etc.When same target creates first, the object modification time is identical with the Object Creation time, thereafter, for the operation of same target, object data verification and when constant, the Object Creation time is constant, and just the object modification time changes.Repeat no more herein.User's write request can for by described data to be written to object metadata with object data creates or update request; Also can not comprise described data to be written, be that described data to be written are for empty, the object data that corresponding master data center generates is also empty, now, user's write request can be upgating object metadata request or delete object requests, thereby generate corresponding synchronous task and delete object task.

In S103, described master data, mind-set sends the first synchronization message for data center, in described the first synchronization message, comprises described object metadata;

S104, described master data center receive the synchronization request that described standby data center sends, and comprise described object identity in described synchronization request;

S105, described master data center send the second synchronization message according to described synchronization request to described standby data center, comprise data described to be written corresponding to object identity that comprise with described synchronization request in described the second synchronization message.

Preferably, in S103, in described master data, mind-set sends the first synchronization message for data center, before comprising described object metadata, also comprises in described the first synchronization message:

Wherein, described master data center, according to the task modification time of each synchronous task, merges the synchronous task that comprises same object mark, comprising:

It should be noted that, do not exist if the synchronous task corresponding with described object metadata has the synchronous task of same object mark, directly synchronous task corresponding to the described object metadata of storage is the synchronous task after merging; Master data center is according to the task modification time in corresponding synchronous task that obtains of object modification time in object metadata, and task modification time, along with the object modification time changes and changes, is consistent with the object modification time.

Further, in S103, described master data center, also comprises according to the synchronous task after merging after described standby data center sends the first synchronization message:

Further, before in described master data, mind-set sends the first synchronization message for data center, also comprise:

Determine whether described object metadata exists;

If do not exist, in described master data, described in mind-set, standby data center sends the deletion object task that comprises described object identity and task modification time, so that object identical with described object identity and before task modification time is deleted by described standby data center.Wherein, the object before task modification time, refers to that the object modification time comprising in the object metadata of same object mark is new unlike this task modification time.

Further, in S104, described master data center receives the synchronization request that described standby data center sends, after comprising described object identity in described synchronization request, in S105, described master data center sends the second synchronization message according to described synchronization request to described standby data center, before comprising the described object data that the object identity that comprises with described synchronization request is corresponding, also comprises in described the second synchronization message:

Described synchronization request also comprises the object modification time;

Preferably, described master data center receives the request of repairing of reading that standby data center sends, described in read to comprise object identity and object modification time in the request of repairing;

Described in the basis of described master data center, read the request of repairing and send the second synchronization message to described standby data center, in described the second synchronization message, comprise with described and read to repair the object identity that comprises of request and described object data corresponding to object modification time, so that the object data that described standby data center comprises described the second synchronization message returns to user, preserve object data that described the second synchronization message comprises to standby data center, and described in mark, the state of native object metadata is that data pass.Wherein master data center according to described in read master data center in the request of repairing sends process from the second synchronization message to described standby data center and S105 to send the process of the second synchronization message for data center similar to described according to described synchronization request, repeat no more.

Preferably, above-mentioned object metadata can be transmitted on first passage, and above-mentioned object data can transmit on second channel.Wherein, first passage and second channel can be, but not limited to as physics designated lane, can in physical transmission channel bandwidth, divide two virtual designated lanes, for example, can be at 100 Mbytes (MegaByte, be called for short: on light transmission passage M), divide the virtual private passage of 1M as first passage connection object metadata, divide the virtual private passage of 5M or 10M as second channel connection object data, make synchronously not disturbing mutually of object metadata and object data, can more stablize the synchronizing cycle of object metadata fast, make user in standby data center, can have access to sooner object.

Specifically, master data center, first receive the write request from subscriber equipment, this write request can comprise data to be written, this data capacity to be written may be larger, such as 5TB, then generate corresponding object metadata according to user's write request, this object metadata is conventionally smaller less than 1M, and be also that object data stores is in local storage center by this object metadata and data to be written, generate the synchronous task corresponding with described object metadata simultaneously, afterwards this locality is merged to the synchronous task obtaining after merging with each synchronous task of object identity according to task modification time, delete each synchronous task of existing same object identity, the only synchronous task after storage merging, determine afterwards whether current time and the difference of task creation time are greater than synchronizing cycle, thereby according to task creation time and synchronizing cycle determining whether to start to carry out synchronous task, start to carry out synchronous task synchronizing cycle if be greater than, obtain and need synchronous object metadata according to the object identity of synchronous task after merging: if object metadata does not exist, illustrate that this object is deleted, send to standby data center the object task of deleting, comprise object identity and task modification time, the corresponding object of standby data center is eliminated in time, if object metadata exists, send to standby data center the first synchronization message that comprises this object metadata, make the standby data center can be according to the synchronous native object metadata of this object metadata, and send synchronization request according to the native object metadata after synchronous, what the standby data center of finally reception sent comprises object identity and the synchronization request of object modification time, obtain corresponding object data according to object identity, if corresponding object data does not exist, or the object modification time of the object data getting is different from synchronization request, to the standby non-existent sync response of data center's returning an object value, otherwise sends to standby data center the second synchronization message that comprises object data, thereby be also that data to be written are synchronized to in data center and back up by object data.Therefore the time needing to standby data center from master data central synchronous is controlled, make the user can be from standby data center access object as early as possible, and also can accept the reparation request of reading that standby data center sends according to user's read request at master data center, return to corresponding object data according to reading the request of repairing to standby data center, make standby data center that the object data receiving is returned to user, make the user can be in standby data center access object as early as possible.

When specific implementation, master data center can comprise: front-end processor (Foreground Processor), background processor (Background Processor), message queue (Message Queue), storage engines (Storage Engine), this structure is applicable to the application's master data center He Bei data center, does not repeat them here.Wherein, front-end processor, can be used for receiving user's write request, and according to user's write request formation object metadata, and storage object metadata and data to be written are to local storage engines; The synchronous task that formation object metadata is corresponding, the synchronous task that obtains same object identity from message queue carries out the synchronous task task merging obtains merging, and sends synchronous task after this merging to message queue; Receive the first sync response of standby data center, if the first sync response is success response, and in message queue, the modification time of the synchronous task of same object mark is identical with the object modification time of the object metadata in the first synchronization message, delete the synchronous task of same object mark in message queue, otherwise, if the first sync response is failure response or overtime first sync response of not receiving, do not delete the operation of task; Receive the synchronization request of standby data center, and from storage engines, read corresponding object data according to synchronization request, if object data does not exist or is inconsistent, to the standby non-existent sync response of data center's returning an object value, otherwise send to standby data center the second synchronization message that comprises object data; And receive the request of repairing of reading for data center, from storage engines, read corresponding object data according to reading the request of repairing, and to sending for data center the second synchronization message that comprises object data.Background processor, can be used for carrying out according to the task creation time monitoring of synchronizing cycle, in the time that current time and the difference of task creation time are greater than synchronizing cycle, from message queue, synchronous synchronous task is respectively treated in scanning, obtain the object metadata of respectively treating that synchronous synchronous task is corresponding, afterwards to standby data center sends the first synchronization message that comprises corresponding object metadata or comprise object identity and the deletion object task of object modification time to for data center.Preferably, described front-end processor and described background processor can be deployed in same processor, can better carry out master data median plane to customer service with towards the management of the backup traffic of standby data center, taking of adjustresources, increase as required more data synchronization strategy, as first synchronization message of carrying out metadata synchronization can complete as early as possible, and carry out the second synchronous synchronization message of data, can reduce to customer service business busy at master data median plane, and increase in the time that user oriented business is idle.

The clone method that the present embodiment provides, by master data center, object metadata and object data are separately synchronized to standby data center, make object metadata can be synchronized to rapidly standby data center, solve user and carried out uncontrollable problem of object accesses time in standby data center, by metadata synchronization task merging, overcome user in synchronizing cycle same object multi-pass operation has easily been caused to the problem that invalid data is synchronous simultaneously.

Fig. 2 is the flow chart of asynchronous replication embodiment of the method two provided by the invention, and as shown in Figure 2, the method embodiment, comprising:

S201, standby data center receive the first synchronization message that master data center sends, and in described the first synchronization message, comprise object metadata;

S202, described standby data center send synchronization request to described master data center, comprise described object identity in described synchronization request;

S203, described standby data center receive the second synchronization message that described master data center sends, and comprise object data corresponding to object identity comprising with described synchronization request in described the second synchronization message.

Further, in S201, standby data center receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, described in S202, standby data center sends synchronization request to described master data center, before comprising described object identity in described synchronization request, also comprise:

Wherein, if described check sum is identical, the state of the native object metadata after native object metadata described in the described object modification time synchronized comprising according to described object metadata for data center, and mark is synchronous is that data pass, comprising:

Further, in S202, described standby data center sends synchronization request to described master data center, comprise described object identity in described synchronization request, comprising:

The data synchronous task that comprises object identity, task modification time described in described standby data center basis sends synchronization request to described master data center, before comprising described object identity, also comprise in described synchronization request:

If do not exist, cancel and delete described data synchronous task;

If not identical, cancel and delete described data synchronous task;

If not data do not pass, cancel and delete described data synchronous task;

Further, in S203, described standby data center receives the second synchronization message that described master data center sends, and in described the second synchronization message, comprises described object data, comprising:

Further, standby data center described in S202, after described master data center sends synchronization request, and before standby data center described in S203 receives the second synchronization message that described master data center sends, also comprises:

Further, the method embodiment, also comprises:

Preferably, the method embodiment, also comprises:

Described standby data center, the state of local metadata described in object data that described the second synchronization message comprises mark of storing is that data pass.

Preferably, above-mentioned object metadata can be transmitted on first passage, and above-mentioned object metadata can be transmitted on second channel, and detailed process and embodiment illustrated in fig. 1 similar, repeats no more.

It should be noted that, standby data center carries out the deletion of data synchronous task according to the sync response at the master data center receiving: do not exist if the sync response receiving is object, delete the metadata of same object mark, and read the data synchronous task of same object mark, if the task modification time of this data synchronous task does not become, delete, otherwise do not delete; If the sync response receiving is the second synchronization message that comprises data to be written, read the data synchronous task of same object mark, if its task modification time does not become, delete, otherwise, read the object metadata that same object identifies, if the check sum difference that object metadata comprises, delete the data to be written that receive from master data center, and do not delete this data synchronous task.In addition, when standby data center self cancels task, read the data synchronous task of same object mark, if the task modification time of this data synchronous task does not become, delete, otherwise do not delete.

Specifically, standby data center, receive the first synchronization message that master data center sends, according to the synchronous native object metadata of the first synchronization message, and it is synchronous to determine whether to carry out data according to the check sum of object metadata, if desired generate the data synchronous task corresponding with synchronous rear native object metadata, and according to data synchronous task to master data center send synchronization request, thereby obtain and need synchronous object data from master data, the data that complete master data center Dao Bei data center are synchronous.Standby data center, employing multistage is downloaded, and record the state of each section, use the technology of breakpoint transmission, complete synchronous from master data center Dao Bei data center of data: described data synchronous task is divided into N segment data synchronous task (as being that unit divides by 1M) by standby data center, and N is positive integer; Described standby data center generates N synchronization request according to N segment data synchronous task respectively, and the each segment data synchronous task of mark is synchronous for not completing; Receive the second synchronization message that described master data center sends according to each synchronization request, if the second synchronization message comprises described data to be written, the data synchronous task of the corresponding section of mark is synchronous for completing, and store described data to be written, if the second synchronization message does not comprise described data to be written, again send synchronization request according to the data synchronous task of corresponding section.Standby data center is in the time receiving the deletion object task at master data center, first obtain the native object metadata of same object mark, if task modification time is newer or identically delete object metadata and data than object modification time, otherwise does not delete, directly return to deletion success response.Standby data center, in the time receiving user's read request, if now object data does not exist, possible scene is that object metadata has completed synchronous but object data does not complete synchronously, or object data damages, and sends to master data center to read to repair and asks, obtain data from master data center and return to user, and write local storage engines, allow user can fast access to object, simultaneously synchronous or repair local data.

When specific implementation, the front-end processor of standby data center is for carrying out the action mutual with master data center: receive the first synchronization message, synchronous native object metadata, object metadata after storage is synchronous is to local storage engines, according to the object metadata generated data synchronous task after synchronous, send to local message queue, make the background processor of standby data center send synchronization request to master data center according to data synchronous task, receive afterwards the second synchronization message that master data sends, the object data of preserving wherein arrives local storage engines, and delete the data synchronous task of the mark of same object in message queue, the deletion object task that receives the transmission of master data center is carried out object deletion, receive the read request of subscriber equipment, according to the existence of native object data whether or whether damage and send the request of repairing of reading to master data center, and receive the corresponding object data that master data center is returned, this object data is returned to user and is kept at local storage engines.

The method that the present embodiment provides, adopts multistage download and breakpoint transmission technology to initiate data synchronization request to master data center by standby data center and completes the synchronous of object data, improved the synchronizing speed of object data; And adopt and read recovery technique by standby data center, complete synchronous and object data in object metadata and also do not complete synchronous in the situation that, make user can fast access to object.

Below in conjunction with specific embodiment, asynchronous replication method provided by the invention is described.

Fig. 3 is the information interaction figure of asynchronous replication method specific embodiment one provided by the invention, this specific embodiment one is specifying information reciprocal process when task merging is carried out at master data center in embodiment of the method shown in Fig. 1 and between subscriber equipment and standby data center, as shown in Figure 3, master data center comprises front-end processor, storage engines and message queue, and this specific embodiment comprises the following steps:

Step 301, subscriber equipment send write request to the front-end processor at master data center;

Front-end processor formation object metadata and the synchronous task at step 302, master data center;

The front-end processor at step 303, master data center is stored data to be written to local storage engines;

Step 304, local storage engines return successfully to the front-end processor at master data center;

The front-end processor storage object metadata at step 305, master data center is to local storage engines;

Step 306, local storage engines return successfully to the front-end processor at master data center;

The front-end processor at step 307, master data center is obtained same object identity synchronous task from local message queue;

Step 308, local message queue are returned to same object identity synchronous task to the front-end processor at master data center;

The front-end processor at step 309, master data center merges each same object identity synchronous task according to task modification time;

After the front-end processor at step 310, master data center sends and merges, synchronous task is to local message queue;

Step 311, local message queue are returned successfully to the front-end processor at master data center;

The front-end processor at step 312, master data center is returned successfully to subscriber equipment.

Wherein, object metadata in step 302 and synchronous task, can also comprise object name, and the front-end processor of step 312 is returned successfully to subscriber equipment according to object name, object for different users may have same object name, but object identity is unique.In step 308, if local message queue is not returned to the synchronous task of same object mark, front-end processor does not perform step 309, directly sends the synchronous task corresponding with the object metadata generating according to user's write request to local message queue; If local message queue is returned to the synchronous task of same object mark, perform step 309 and carry out data merging, in detailed process similar approach embodiment mono-, merge the process of synchronous task, repeat no more.

In prior art, then sequentially generate synchronous task sends to local message queue to front-end processor reception user's write request one by one, after arriving synchronizing cycle, send to one by one more standby data center and carry out the synchronous of object data and metadata, the object identifying for same object is within synchronizing cycle, may there is multi-pass operation in user, now only need to be according to the last user's write request synchronization object in synchronizing cycle to standby data center.

The method that this specific embodiment one provides, by merging synchronous task, improves in master data mind-set for the synchronous efficiency of data center, and has avoided the invalid synchronous problem causing for the repeatedly user write request of same object in synchronizing cycle.

Fig. 4 is the information interaction figure of asynchronous replication method specific embodiment two provided by the invention, this specific embodiment two is for carrying out the synchronous specifying information reciprocal process of object metadata by the first synchronization message to standby data center in master data center in embodiment of the method shown in Fig. 1, as shown in Figure 4, this specific embodiment comprises the following steps:

Step 401, master data center background processor obtain synchronous task from the queue of master data Central Message;

Step 402, the queue of master data Central Message are returned to synchronous task to master data center background processor;

Step 403, master data center background processor obtain the object metadata of same object mark from master data central store engine;

Step 404, master data central store engine are to master data center background processor returning an object value metadata;

Step 405, master data center background processor compare the consistency of synchronous task and object metadata;

Step 406, master data center background processor send the first synchronization message or delete object task to standby data center;

Step 407, standby data center obtain the standby object metadata of same object mark;

Object is deleted according to the standby object metadata in the synchronous this locality of the first synchronization message or according to deleting object task by step 408, standby data center;

Step 409, standby data center return to sync response or delete object response to master data center background processor;

Step 410, master data center background processor are deleted the synchronous task in the queue of master data Central Message.

Wherein, shown in step 406 and step 410 and Fig. 1 in embodiment of the method one shown in similar step 408～409 of S103 and Fig. 2 in embodiment of the method two S201 similar, repeat no more.In step 405, check the consistency of synchronous task and object metadata, in the S103 of embodiment of the method one shown in Fig. 1, comparison other modification time, determine whether object metadata needs to be synchronized in standby data center by the first synchronization message, and detailed process repeats no more.It should be noted that, in step 401, a kind of possible execution mode is, the background processor at master data center, carry out monitoring synchronizing cycle, such as synchronizing cycle be 6 hours, treat synchronous synchronous task scan message queue in the time of 12 o'clock in, carry out subsequent synchronisation processing if the task creation time for the treatment of synchronous synchronous task in message queue was obtained this synchronous task before 6 o'clock; In step 404, if it is empty, object metadata does not exist returning an object value metadata, performs step 406 transmissions and deletes object task, and the standby data center of band matching condition prompting is carrying out object while deleting, can only delete the task creation time and be the object of certain time period before sweep time.For example being with matching condition is 1 hour before sweep time, in step 408, standby data center is in the time deleting object, the Object Creation time that first will check object whether before 11 o'clock, if the Object Creation time after 11 o'clock, do not delete object, execution step 409, directly return to deletion success message to master data center background processor, if the Object Creation time before 11 o'clock, is deleted object metadata and object data that synchronous task is corresponding.In step 408, standby data center is according to the synchronous local standby object metadata of the first synchronization message, can be to generate according to the object metadata in the first synchronization message, and to replace attribute that new object metadata is relevant to data center be that the attribute of standby data center obtains the object metadata after synchronous, the state of the object metadata after synchronous according to check sum mark is that data have passed or data do not pass, and is saved in afterwards local storage engines and completes the synchronous of object metadata.

The method specific embodiment, sends the first synchronization message synchronization object metadata by mind-set in master data for data center, ensures that object metadata is synchronized to standby data center as early as possible, makes user controlled in the time of standby data center access object.

Fig. 5 is the information interaction figure of asynchronous replication method specific embodiment three provided by the invention, this specific embodiment three is shown in Fig. 2, in embodiment of the method, to send to master data center the specifying information reciprocal process of synchronization request with synchronization object data for data center, as shown in Figure 5, the method specific embodiment, comprises the following steps:

Step 501, standby data center background processor are obtained data synchronous task from standby data center message queue;

Step 502, standby data center message queue are to standby data center background processor return data synchronous task;

Step 503, standby data center background processor are obtained the object metadata of same object mark from standby data center storage engines;

Step 504, standby data center storage engines are to standby data center background processor returning an object value metadata;

Step 505, standby data center background processor check the consistency of data synchronous task and object metadata;

Step 506, standby data center background processor be the segmentation of data synchronous task, the state that mark is each section, and generate N synchronization request;

Step 507, standby data center background processor send synchronization request to master data center;

Step 508, master data center obtain object data;

In step 509, master data, mind-set sends the second synchronization message that comprises object data for data center's background processor;

Step 510, standby data center background processor are upgraded the state of each section, if failure resends the synchronization request of corresponding section;

If the each segment data of step 511 is all successfully obtained, standby data center background processor storage object data are to standby data center storage engines;

Step 512, standby data center background processor amendment object metadata state are that data pass;

Step 513, standby data center background processor are deleted data synchronous task.

Specifically, the background processor of standby data center is integrality and the consistency that ensures data, from local message queue, obtain data synchronous task, and obtain the object metadata of like-identified according to data synchronous task, if object metadata exists and performs step 504, otherwise cancel data synchronous task and perform step 513 deletion data synchronous tasks, if object metadata exists, perform step further 505 by the consistency of task modification time inspection data synchronous task and object metadata, detailed process is with determining whether to send to master data center synchronization request by comparison other modification time in the S202 of embodiment of the method two shown in Fig. 2, specifically comparison procedure is similar repeats no more, if upgrading description object, is updated the object modification time, cancel task execution step 513.If data synchronous task is consistent with object metadata, perform step 505～513, shown in detailed process and Fig. 1 in embodiment of the method one shown in S105 and Fig. 2 in embodiment of the method two S202～S203 similar, repeat no more.

The method specific embodiment, initiates data synchronization request, and utilizes multistage to download and breakpoint transmission technology to master data center by standby data center, improved in master data mind-set for the efficiency of data center's synchronization object data.

Fig. 6 is the information interaction figure of asynchronous replication method specific embodiment four provided by the invention, this specific embodiment four sends to master data center and reads the request of repairing to obtain object data and to return to user's specifying information reciprocal process for standby data center in embodiment of the method shown in Fig. 2 receives user's read request, as shown in Figure 6, the method specific embodiment comprises:

Step 601, subscriber equipment send read request to standby data center front-end processor;

Step 602, standby data center front-end processor are obtained object metadata from standby data center storage engines;

Step 603, standby data center storage engines are to the front-end processor returning an object value metadata success of standby data center;

Step 604, standby data center front-end processor are obtained object data from standby data center storage engines;

Step 605, standby data center storage engines are to standby data center front-end processor returning an object value data failure;

Step 606, standby data center front-end processor, according to object metadata, send the request of repairing of reading to master data center;

Step 607, master data center obtain object data;

In step 608, master data, mind-set is for data center's front-end processor returning an object value data;

Step 609, standby data center front-end processor are to subscriber equipment returning an object value data;

Step 610, standby data center's front-end processor storage object data tagged object metadata state are that data pass;

Step 611, standby data center front-end processor are deleted the data synchronous task in standby data center message queue.

Specifically, subscriber equipment sends read request to standby data center, if object data corresponding to read request also do not complete synchronously or synchronous object data damages, standby data center can send to master data center the request of repairing of reading, execution step 606～610, shown in detailed process and Fig. 2, in embodiment of the method two, in S203, read to repair part similar, repeat no more.It should be noted that, for not completing synchronous object data, in standby data center message queue, may there is the data synchronous task of the execution that is not also able to do in time accordingly, therefore need to perform step 611, from standby data center message queue, obtain the data synchronous task of same object mark, check the consistency of data synchronous task and object metadata, if identical this data synchronous task of deleting of object modification time.

The method specific embodiment, initiate to read the request of repairing to master data center by standby data center and obtain the object data that subscriber equipment need to be accessed, subscriber equipment is not also completed synchronous in the situation that at object data, can fast access object data, the object data having damaged for this locality is repaired the integrality of guarantee native object data simultaneously by reading the request of repairing.

Fig. 7 is the structural representation of master data center implementation example one provided by the invention, and as shown in Figure 7, this master data center 700, comprises memory module 701, the first synchronized transmission modules 702 and the second synchronized transmission module 703.Wherein, memory module 701, the write request sending for receiving subscriber equipment, comprises data to be written in described write request; And

The first synchronized transmission module 702, for sending the first synchronization message to standby data center, comprises described object metadata in described the first synchronization message;

The second synchronized transmission module 703, the synchronization request sending for receiving described standby data center, comprises described object identity in described synchronization request; And

Further, described the first synchronized transmission module 702, specifically for:

Obtain the each synchronous task that comprises same object mark;

Further, described the first synchronized transmission module 702 according to the synchronous task after merging after described standby data center sends the first synchronization message, also for:

Further, before described the first synchronized transmission module 702 sends the first synchronization message to standby data center, also for:

Determine whether described object metadata exists;

Further, the synchronization request that described in described the second synchronized transmission module 703, standby data center sends, after comprising described object identity in described synchronization request, described the second synchronized transmission module sends the second synchronization message according to described synchronization request to described standby data center, before comprising the object data that the object identity that comprises with described synchronization request is corresponding in described the second synchronization message, also for:

Described synchronization request also comprises the object modification time;

Preferably, described the second synchronized transmission module 703, also for:

Preferably, described object metadata is transmitted on first passage, and described data to be written are transmitted on second channel.

This master data center implementation example, can be used to shown in the technical scheme in embodiment of the method shown in execution graph 1 and Fig. 3～Fig. 6 corresponding technical scheme in specific embodiment, and it realizes principle and technique effect is similar, repeats no more.

Fig. 8 is the structural representation of the standby embodiment of data center mono-provided by the invention, and as shown in Figure 8, this comprises the first synchronous receiver module 801 and the second synchronous receiver module 802 for data center 800.Wherein, the first synchronous receiver module 801, the first synchronization message sending for receiving master data center, comprises object metadata in described the first synchronization message;

The second synchronous receiver module 802, for send synchronization request to described master data center, comprises described object identity in described synchronization request; And

Further, the described first synchronous receiver module 801 receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, the described second synchronous receiver module 802 sends synchronization request to described master data center, before comprising described object identity in described synchronization request, the described first synchronous receiver module 801, also for:

Send the first sync response to described master data center.

Further, the described first synchronous receiver module 801, specifically for:

Further, the described second synchronous receiver module 802, specifically for:

Further, the described second synchronous receiver module 802, according to the described data synchronous task that comprises object identity, object modification time to described master data center send synchronization request, before comprising described object identity in described synchronization request, also for:

If do not exist, cancel and delete described data synchronous task;

If not identical, cancel and delete described data synchronous task;

If not data do not pass, cancel and delete described data synchronous task;

Further, the described second synchronous receiver module 802 is sending after synchronization request to described master data center, before receiving the second synchronization message that described master data center sends, also for:

Receive the sync response at described master data center;

Further, the described first synchronous receiver module 801, also for:

Preferably, the described second synchronous receiver module 802, also for: receive the read request of subscriber equipment, described user's read request packet is containing object identity;

Storing the state of native object metadata described in described object data mark is that data pass.

Preferably, described object metadata is transmitted on first passage, and described object data transmits on second channel.

This is for the embodiment of data center, can be used to shown in the technical scheme in embodiment of the method shown in execution graph 2 and Fig. 3～Fig. 6 corresponding technical scheme in specific embodiment, and it realizes principle and technique effect is similar, repeats no more.

Fig. 9 is the structural representation of master data center implementation example two provided by the invention, as shown in Figure 9, this master data center 900, comprise first processor 901 and the first memory 902 being connected with described first processor 901, in wherein said first memory 902, store batch processing code, first processor 901 calls described program code, can be used for carrying out shown in technical scheme in embodiment of the method one as shown in Figure 1 and Fig. 3～Fig. 6 corresponding technical scheme in specific embodiment, it realizes principle and technique effect is similar, repeats no more.

Figure 10 is the structural representation of the standby embodiment of data center bis-provided by the invention, as shown in figure 10, this is for data center 1000, comprise the second processor 1001 and the second memory 1002 being connected with described the second processor 1001, in wherein said second memory 1002, store batch processing code, the second processor 1001 calls described program code, can be used for carrying out shown in technical scheme in embodiment of the method as shown in Figure 2 and Fig. 3～Fig. 6 corresponding technical scheme in specific embodiment, it realizes principle and technique effect is similar, repeats no more.

Figure 11 is the structural representation of asynchronous replication system embodiment provided by the invention, as shown in figure 11, this asynchronous replication system 1100, comprise master data center 900 as shown in Figure 9 and standby data center 1000 as shown in figure 10, can be used for the technical scheme in execution graph 1～embodiment illustrated in fig. 6, it realizes principle and technique effect is similar, repeats no more.

One of ordinary skill in the art will appreciate that: all or part of step that realizes above-mentioned each embodiment of the method can complete by the relevant hardware of program command.Aforesaid program can be stored in a computer read/write memory medium.This program, in the time carrying out, is carried out the step that comprises above-mentioned each embodiment of the method; And aforesaid storage medium comprises: various media that can be program code stored such as ROM, RAM, magnetic disc or CDs.

Finally it should be noted that: above each embodiment, only in order to technical scheme of the present invention to be described, is not intended to limit; Although the present invention is had been described in detail with reference to aforementioned each embodiment, those of ordinary skill in the art is to be understood that: its technical scheme that still can record aforementioned each embodiment is modified, or some or all of technical characterictic is wherein equal to replacement; And these amendments or replacement do not make the essence of appropriate technical solution depart from the scope of various embodiments of the present invention technical scheme.

Claims

1. an asynchronous replication method, is characterized in that, comprising:

2. method according to claim 1, is characterized in that, in described master data, mind-set sends the first synchronization message for data center, before comprising described object metadata, also comprises in described the first synchronization message:

3. method according to claim 2, is characterized in that, described master data center, according to the task modification time of each synchronous task, merges the synchronous task that comprises same object mark, comprising:

4. according to the method in claim 2 or 3, it is characterized in that, described master data center, also comprises according to the synchronous task after merging after described standby data center sends the first synchronization message:

5. according to the method described in claim 1-4 any one, it is characterized in that, before mind-set sends the first synchronization message for data center in described master data, also comprise:

Determine whether described object metadata exists;

6. according to the method described in claim 1-5 any one, it is characterized in that, described master data center receives the synchronization request that described standby data center sends, after comprising described object identity in described synchronization request, described master data center sends the second synchronization message according to described synchronization request to described standby data center, before comprising the object data that the object identity that comprises with described synchronization request is corresponding in described the second synchronization message, also comprise:

Described synchronization request also comprises the object modification time;

7. according to the method described in claim 1-6 any one, it is characterized in that, also comprise:

Described master data center receives the request of repairing of reading that standby data center sends, described in read to comprise object identity and object modification time in the request of repairing;

8. according to the method described in claim 1-7 any one, it is characterized in that, described object metadata is transmitted on first passage, and described object data transmits on second channel.

9. an asynchronous replication method, is characterized in that, comprising:

10. method according to claim 9, it is characterized in that, standby data center receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, described standby data center sends synchronization request to described master data center, before comprising described object identity in described synchronization request, also comprise:

11. methods according to claim 10, it is characterized in that, if described check sum is identical, native object metadata described in the object modification time synchronized that described standby data center comprises according to described object metadata, and the state of the native object metadata of mark after synchronous is that data pass, comprising:

12. methods according to claim 11, is characterized in that, described standby data center sends synchronization request to described master data center, comprise described object identity in described synchronization request, comprising:

13. methods according to claim 12, it is characterized in that, the data synchronous task that comprises object identity, object modification time described in described standby data center basis sends synchronization request to described master data center, before comprising described object identity, also comprise in described synchronization request:

If do not exist, cancel and delete described data synchronous task;

If not identical, cancel and delete described data synchronous task;

If not data do not pass, cancel and delete described data synchronous task;

14. according to the method described in claim 9-13, it is characterized in that, described standby data center receives the second synchronization message that described master data center sends, and in described the second synchronization message, comprises described object data, comprising:

15. according to the method described in claim 9-14, it is characterized in that, described standby data center is after described master data center sends synchronization request, and described standby data center also comprises before receiving the second synchronization message of described master data center transmission:

16. according to the method described in claim 9-15 any one, it is characterized in that, also comprises:

17. according to method described in claim 9-16 any one, it is characterized in that, also comprises:

18. according to the method described in claim 9-17 any one, it is characterized in that,

19. 1 kinds of master data centers, is characterized in that, comprising:

20. master data according to claim 19 centers, is characterized in that, described the first synchronized transmission module, specifically for:

21. according to master data center described in claim 20, it is characterized in that, described the first synchronized transmission module, specifically for:

Obtain the each synchronous task that comprises same object mark;

22. according to the master data center described in claim 20 or 21, it is characterized in that, described the first synchronized transmission module according to the synchronous task after merging after described standby data center sends the first synchronization message, also for:

23. according to the master data center described in claim 19-22 any one, it is characterized in that, before described the first synchronized transmission module sends the first synchronization message to standby data center, also for:

Determine whether described object metadata exists;

24. according to the master data center described in claim 19-23 any one, it is characterized in that, described the second synchronized transmission module receives the synchronization request that described standby data center sends, after comprising described object identity in described synchronization request, described the second synchronized transmission module sends the second synchronization message according to described synchronization request to described standby data center, before comprising the object data that the object identity that comprises with described synchronization request is corresponding in described the second synchronization message, also for:

Described synchronization request also comprises the object modification time;

25. according to the master data center described in claim 19-24 any one, it is characterized in that, described the second synchronized transmission module, also for:

26. according to master data center described in claim 19-25 any one, it is characterized in that,

27. 1 kinds of standby data centers, is characterized in that, comprising:

28. standby data centers according to claim 27, it is characterized in that, the described first synchronous receiver module receives the first synchronization message that master data center sends, after comprising described object metadata in described the first synchronization message, the described second synchronous receiver module sends synchronization request to described master data center, before comprising described object identity in described synchronization request, the described first synchronous receiver module, also for:

Send the first sync response to described master data center.

29. standby data centers according to claim 28, is characterized in that, the described first synchronous receiver module, specifically for:

30. standby data centers according to claim 29, is characterized in that, the described second synchronous receiver module, specifically for:

31. standby data centers according to claim 30, it is characterized in that, the described second synchronous receiver module, send synchronization request to described master data center according to the described data synchronous task that comprises object identity, object modification time, before comprising described object identity in described synchronization request, also for:

If do not exist, cancel and delete described data synchronous task;

If not identical, cancel and delete described data synchronous task;

If not data do not pass, cancel and delete described data synchronous task;

32. according to the standby data center described in claim 27-31, it is characterized in that, and the described second synchronous receiver module, specifically for:

33. according to the standby data center described in claim 27-32, it is characterized in that, the described second synchronous receiver module sends after synchronization request to described master data center, before receiving the second synchronization message that described master data center sends, also for:

Receive the sync response at described master data center;

34. according to the standby data center described in claim 27-33 any one, it is characterized in that, the described first synchronous receiver module, also for:

35. according to the standby data center described in claim 27-34 any one, it is characterized in that, the described second synchronous receiver module, also for:

36. according to the standby data center described in claim 27-35 any one, it is characterized in that,

37. 1 kinds of master data centers, it is characterized in that, comprise first processor and the first memory being connected with described first processor, in wherein said first memory, store batch processing code, described first processor calls described program code, for carrying out the asynchronous replication method as described in claim 1-8 any one.

38. 1 kinds of standby data centers, it is characterized in that, comprise the second processor and the second memory being connected with described the second processor, in wherein said second memory, store batch processing code, described the second processor calls described program code, for carrying out the asynchronous replication method as described in claim 9-18 any one.

39. 1 kinds of asynchronous replication systems, is characterized in that, comprising:

Master data center described in claim 37, and standby data center described in claim 38.