Embodiment
For the purpose, technical scheme and the advantage that make 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, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making under the creative work prerequisite the every other embodiment obtained, belong to the scope of protection of the invention.
Embodiment mono-
The process flow diagram of a kind of asynchronous remote copying method that Fig. 1 provides for the embodiment of the present invention.The present embodiment provides the technical scheme of asynchronous remote copying method to describe from production center angle to the embodiment of the present invention.The storage system that the executive agent of the present embodiment is the production center.As shown in Figure 1, the present embodiment comprises:
Step 11: after the production center receives the write request of main frame, be write request mark week issue according to all issues of current period.
According to all issues of current period, being write request mark week issue, is the residing cycle while for the mark production center, processing this write request.
Step 12: the production center writes local remote copy master LUN by the write request of the week of mark issue, the write request flag sequence number that is described week of mark issue according to the order that writes remote copy master LUN.
The production center, by after the write request of all issues of mark writes local remote copy master LUN, writes the order of remote copy master LUN, for this write request is marked in current period the serial number that is written into remote copy master LUN according to write request.Therefore write request not only has and is marked with all issues corresponding to this write request, also is marked with the serial number of this write request when writing remote copy master LUN.
Step 13: the production center by the week of mark issue and the write request of serial number send to disaster recovery center.
Further, after current period finishes, the production center sends to disaster recovery center by the write request number of the main frame received in current period of all issues of current period and correspondence.Thereby so that disaster recovery center can be according to all issues and corresponding write request number, determine whether to receive all write requests in this week corresponding cycle of issue, and, when determining all write requests that receive this week corresponding cycle of issue, all write requests in this cycle are write to remote copy from LUN.
The embodiment of the present invention, when the storage system of the production center receives the write request of main frame, be all issues of this write request mark, and after write request writes remote copy master LUN, for this write request is marked in current period the serial number that is written into remote copy master LUN.After the write request that the production center will be marked with all issues and serial number sends to disaster recovery center, for the write request disaster recovery center of same period, can be in order number write remote copy from LUN.Thereby a plurality of write requests of having avoided the conflict of in same period reference address arrive the order of disaster recovery centers when inconsistent with the write sequence of the heart aborning, the inconsistent problem of data that the data of storing in the remote copy master LUN of the caused production center and the remote copy of disaster recovery center are stored from LUN.
Embodiment bis-
Provided respectively in the present embodiment the concrete application example of the asynchronous remote copying method shown in Fig. 1 in distributed memory system and non-distributed memory system.
A kind of asynchronous remote copying method that Fig. 2 A provides for the embodiment of the present invention is applied to the process flow diagram of distributed memory system.The structural representation of a kind of production center that Fig. 2 B provides for the embodiment of the present invention.The present embodiment illustrates the technical scheme of above-mentioned asynchronous remote copying method from production center angle.As shown in Fig. 2 B, in the present embodiment, the distributed memory system of the production center comprises at least two production group systems.The write request that the production group system issues for the treatment of the main frame of the production center.Wherein, each production group system comprises the first front-end platform and the first rear end platform, and wherein the first rear end platform comprises the first business processing device and remote copy master LUN.Between the first front-end platform in each production group system, can connect by internal network.As shown in Figure 2 A, the present embodiment comprises:
Step 21: the write request that the main frame of each the production group system reception production center in the production center issues.
Alternatively, in main frame and each production group system, the first front-end platform is by the communication of SCSI agreement, and the first front-end platform of each production group system receives by the SCSI agreement write request that the main frame of the production center issues.
Step 22: each production group system is determined the physical address of this write request according to address mapping table, and whether the physical address space that judges this production group system processing comprises the physical address of this write request, the physical address space that this production group system is processed if judgment result is that comprises the physical address of this write request, enters step 23; Otherwise, ignore this write request or forward this write request to the first front-end platform of other production group system.
After the first front-end platform in each production group system receives the write request that main frame issues, according to the logical address of carrying in write request, determine the physical address of write request by searching address mapping table, preserve the mapping relations of logical address and physical address in distributed memory system in address mapping table.
For example, because the physical address space of the write request that in the production center, each production group system is processed is different (from the physical address space of the write request that in this production center, other production group systems are processed, the first production group system is the write request in 00000000~1FFFFFFF physical address space for the treatment of physical address, and the second production group system is the write request in 20000000~3FFFFFFF physical address space for the treatment of physical address).In the time of in the physical address space that the physical address that the first front-end platform in each production group system is determined write request is not processed in affiliated production group system, ignore this write request or forward this write request to the first front-end platform of other production group system.
Step 23: the physical address space of processing comprises that the production group system of the physical address of this write request obtains all issues of current period, and is write request mark week issue according to all issues of the current period obtained, and enters step 24.
Alternatively, the first front-end platform in described production group system is obtained all issues of current period, and is write request mark week issue according to all issues of the current period obtained.
The scheme that the first front-end platform in wherein production group system is obtained all issues of current period includes but not limited to:
The first front-end platform in scheme one, production group system from the production center one for the setting device that all issues are set, obtain, specifically comprise: when receiving write request, send current period number to described setting device and obtain request, and receive all issues of the current period of described setting device feedback; Perhaps, described setting device periodically sends to all issues of current period the first front-end platform in each production group system in this production center.The first front-end platform periodically receives all issues of the current period of described setting device transmission.
Scheme two selects a production group system as master control production group system in advance from each production group system of the production center, and the first front-end platform in this master control production group system can be referred to as master control the first front-end platform.Wherein master control the first front-end platform is for arranging all issues.
Described be write request mark week the production group system of issue can be master control production group system, can be also non-master control production group system.When the described production group system for all issues of write request mark is master control production group system, all issues of the Set For Current that master control the first front-end platform can arrange self are as all issues of the current period obtained; When the described production group system for all issues of write request mark is non-master control production group system, the mode that non-master control the first front-end platform obtains all issues of current period includes but not limited to following several:
First method, master control the first front-end platform, according to the request of non-master control the first front-end platform in this production center, is handed down to current period number non-master control the first front-end platform sent request.
Second method, master control the first front-end platform periodically sends to all issues of current period each non-master control first front-end platform in this production center.Non-master control the first front-end platform periodically receives all issues of the current period of master control the first front-end platform transmission.Master control the first front-end platform and Ge Fei master control the first front-end platform can adopt internal network to interconnect.
With respect to first method, second method periodically sends all issues by master control the first front-end platform to the first front-end platform in each non-master control production group system, rather than by a plurality of non-master control the first front-end platform frequently to master control the first front-end platform RQ cycle number, thereby alleviated the load of master control the first front-end platform, improved system performance.
Further, master control the first front-end platform is also added up in current period the write request number that receives main frame, after current period finishes, master control production group system sends to disaster recovery center by the write request number of the main frame received in current period of all issues of current period and correspondence.
Step 24, described production group system writes the remote copy master LUN of described production group system by the write request of the week of mark issue.
Alternatively, in described production group system, the first front-end platform is issued to the first business processing device in the first rear end platform in this production group system by the write request of all issues of mark current period, then in this production group system the first business processing device in the first rear end platform by mark week issue write request write remote copy master LUN.
Step 25: described production group system is the write request flag sequence number for the issue of mark week according to the order that writes local remote copy master LUN.
Alternatively, the first business processing device in described production group system, the write request flag sequence number according to the order that writes local remote copy master LUN for the week of mark issue.
Step 26: described production group system by the week of mark issue and the write request of serial number send to disaster recovery center.
Alternatively, the first business processing device in this production group system, by the week of mark issue and the write request of serial number send to disaster recovery center.
Alternatively, above-mentioned the first business processing device can by the week of mark issue and the write request of serial number write buffer memory, by buffer memory, write request is issued to disaster recovery center.
In the embodiment of the present invention, the storage system of the production center is distributed memory system, and wherein each production group system is respectively used to process the write request of different physical address spaces.After receiving write request, when the production group system judges that the physical address space of this production group system processing comprises the physical address of write request, for all issues of this write request mark, and be write request flag sequence number according to the order that this write request is write to remote copy master LUN.The production group system is by after the write request of all issues of mark and serial number sends to disaster recovery center, after disaster recovery center receives the write request that is marked with all issues and serial number, for the write request disaster recovery center of same period, can be in order number write remote copy from LUN.Thereby a plurality of write requests of having avoided the conflict of in same period reference address arrive the order of disaster recovery centers when inconsistent with the write sequence of the heart aborning, the inconsistent problem of data that the data of storing in the remote copy master LUN of the caused production center and the remote copy of disaster recovery center are stored from LUN.
It should be noted that, the asynchronous remote copying method that accompanying drawing 1 provides also can be applied to non-distributed storage system, in the storage system of the production center, only comprises the situation of a production group system in accompanying drawing 2B.Non-distributed memory system also can be regarded as a special case of distributed storage system.A kind of asynchronous remote copying method that Fig. 2 C provides for the embodiment of the present invention is applied to the process flow diagram of non-distributed memory system:
Step 201: the write request that the main frame of the production group system reception production center issues.
The write request that the main frame of the first front-end platform reception production center in production group system alternatively, issues.
Because the storage system of the production center in this example only comprises a production group system, therefore without the determining step of carrying out step 22 in accompanying drawing 2A.
Step 202: the production group system is obtained all issues of current period, and is write request mark week issue according to all issues of the current period obtained.
Alternatively, the first front-end platform in the production group system is obtained all issues of current period, and be write request mark week issue according to all issues of the current period obtained, described all issues can be that the first front-end platform in this production group system arranges, and can be also for the device setting that current period number is set the first front-end platform that periodically is handed down to this production group system in production platform.
Step 203: the production group system writes local remote copy master LUN by the write request of all issues of mark current period.
The first front-end platform in the production group system is issued to the first business processing device in the first rear end platform in the production group system by the write request of all issues of mark current period, then in the production group system the first business processing device in the first rear end platform by mark week issue write request write remote copy master LUN.
Step 204, the production group system is the write request flag sequence number for the issue of mark week according to the order that writes remote copy master LUN.
Alternatively, the first business processing device in the production group system, the write request flag sequence number according to the order that writes local remote copy master LUN for the week of mark issue.
Step 205, the production group system by the week of mark issue and the write request of serial number send to disaster recovery center.
Alternatively, the first business processing device in the production group system, by the week of mark issue and the write request of serial number send to disaster recovery center.Further, the first business processing device in the production group system can by the week of mark issue and the write request of serial number write buffer memory, regularly write request is issued to disaster recovery center by buffer memory.
Embodiment tri-
Corresponding with embodiment mono-, the present embodiment is from the technical scheme of disaster tolerance central angle degree explanation asynchronous remote copying method of the present invention.The process flow diagram of the another kind of asynchronous remote copying method that Fig. 3 provides for the embodiment of the present invention.The storage system that the executive agent of the present embodiment is disaster recovery center.As shown in Figure 3, the present embodiment comprises:
Step 31: disaster recovery center receives the write request that the production center sends, and is marked with all issues and serial number in this write request.
Step 32: disaster recovery center judges whether to receive all write requests of one-period correspondence.
Step 33: disaster recovery center is determining that while having received all write requests of one-period correspondence, all issues and serial number according to mark in write request, write remote copy from LUN by the write request received.
The mode that disaster recovery center judges whether to receive all write requests of one-period correspondence in step 32 includes but not limited to:
Mode one, for the duration in each cycle of the production center, be isometric and the indefinite situation of the number write request that main frame that receive in each cycle is sent.When one-period finishes, the write request number that the main frame received in the described current period of all issues of current period and correspondence is sent in the production center sends to disaster recovery center; All issues that disaster recovery center sends according to the production center and corresponding write request number, and all issues and the serial number of mark in the write request that receives of step 31, judge whether to receive all write requests of one-period correspondence.
Mode two, the situation that is fixing preset value for the indefinite and number write request that main frame that receive in each cycle is sent of the duration in each cycle of the production center.All issues and the serial number of mark in the write request that disaster recovery center receives according to step 31, can judge whether to have received all write requests of one-period correspondence, for example, when sign has the quantity of the write request of same all issues and different order number to arrive described fixing preset value, all write requests that receive the one-period correspondence are described.
The embodiment of the present invention, the write request that disaster recovery center receives is marked with all issues and serial number, for the write request disaster recovery center of same period, can be in order number writes remote copy from LUN.Thereby a plurality of write requests of having avoided the conflict of in same period reference address arrive the order of disaster recovery centers when inconsistent with the write sequence of the heart aborning, the inconsistent problem of data that the data of storing in the remote copy master LUN of the caused production center and the remote copy of disaster recovery center are stored from LUN.
Embodiment tetra-
Provided respectively in the present embodiment the concrete application example of the asynchronous remote copying method shown in Fig. 3 in distributed memory system and non-distributed memory system.
The another kind of asynchronous remote copying method that Fig. 4 A provides for the embodiment of the present invention is applied to the process flow diagram of distributed memory system.The disaster recovery center example structure schematic diagram that Fig. 4 B provides for the embodiment of the present invention.The present embodiment illustrates the technical scheme of this asynchronous remote copying method from disaster tolerance central angle degree.As shown in Figure 4 B, in the present embodiment, the distributed memory system of disaster recovery center comprises at least two disaster tolerance cluster systems.Each disaster tolerance cluster system comprises the second front-end platform and the second rear end platform, and wherein the second rear end platform comprises that the second business processing device and remote copy are from LUN, and the second rear end platform also comprises the far-end buffer memory alternatively.The mode that disaster recovery center adopts the first in embodiment tri-to judge whether to receive all write requests of one-period correspondence in the present embodiment judges whether to receive all write requests of one-period correspondence.As shown in Figure 4 A, the present embodiment comprises:
Step 41: the disaster tolerance cluster system in disaster recovery center receives the write request that send the production center, is marked with all issues and serial number in this write request.This write request is that the identical production group system of physical address space of processing with this disaster tolerance cluster system in the production center sends.
Alternatively, the second front-end platform in the disaster tolerance cluster system receives the write request that the production group system in the production center sends.
That is to say, receive the handled physical address space of disaster tolerance cluster system of write request in disaster recovery center, the physical address space of processing with the production group system that sends this write request in the production center is identical.
Step 42: all issues and the serial number of mark in the write request number of all issues that described disaster tolerance cluster system sends according to the production center and correspondence and the temporary write request received, judge whether to receive all write requests of one-period correspondence, if while judgment result is that all write requests that receive the one-period correspondence, execution step 43, otherwise return to step 41, continue to receive the write request that the production center sends.
Alternatively, after the write request that the first business processing device of the second front-end platform reception production center of this disaster tolerance cluster system sends, the second business processing device in the second rear end platform of disaster tolerance cluster system under this write request is handed down to, the second business processing device can write the write request received in the far-end buffer memory of the second rear end platform.Particularly, according to all issues in write request, the second business processing device writes write request in the queue that in the far-end buffer memory, all issues are corresponding.For example, in the far-end buffer memory, write request is stored with the chain sheet form, chained list of one-period correspondence, and in chained list, write request sorts with serial number.When the second business processing device writes write request according to all issues of write request to chained list in the far-end buffer memory, according to the serial number of mark in write request, write request is inserted to relevant position in chained list.The schematic diagram of the write request chained list of preserving in the far-end buffer memory in the disaster tolerance cluster system that Fig. 4 C provides for the embodiment of the present invention.As shown in Figure 4 C, each week corresponding chained list of issue, in chained list, write request sorts with serial number.
Further, the second front-end platform of this disaster tolerance cluster system also can receive all issues and this week interior write request number received of cycle corresponding to issue that the production center sends, and all issues and the corresponding write request number that receive are sent to the second business processing device in the second rear end platform in affiliated disaster tolerance cluster system.All issues and the serial number of mark in all issues that described the second business processing device sends according to the second front-end platform and corresponding write request number and the temporary write request received, judge whether to receive all write requests of one-period correspondence.Whether the write request number that this week issue that the number of the write request of the same period that particularly, the second business processing device judgement has received number and different order number is sent with the production center is corresponding is identical; If the write request number that this week issue that the number of the write request of the same period received number and different order number is sent with the production center is corresponding is identical, determines and received this week all write requests corresponding to issue; Otherwise, determine and do not receive this week all write requests corresponding to issue.
Step 43: when described disaster tolerance cluster system has received all write requests of one-period correspondence, judge the write request that whether has the reference address conflict in all write requests that this week issue is corresponding.If judgment result is that to be, perform step 44, otherwise perform step 45.
Step 44: if there is the write request of reference address conflict, this disaster tolerance cluster system writes the remote copy of this disaster tolerance cluster system from LUN according to the serial number in write request successively by this week write request corresponding to issue.
Alternatively, the second business processing device of this disaster tolerance cluster system search write request in the queue of far-end buffer memory in queue corresponding to all issues, according to serial number in write request, the write request in queue is write to remote copy from LUN.
Step 45: if there is not the write request of reference address conflict, this disaster tolerance cluster system write request that this week issue is corresponding is concurrent writes remote copy from LUN.
Alternatively, the second business processing device of this disaster tolerance cluster system search write request in the queue of far-end buffer memory in queue corresponding to all issues, by this week issue, corresponding write request was concurrent writes remote copy from LUN.For example, Host1 and Host2 in T2 production center in the cycle issue the A write request of access logic address LBA1 and the B write request of access logic address LBA1 to the first front-end platform first, and the order that in the local remote copy master LUN of the first rear end platform, write request writes is B after first A.When causing the B write request, the reasons such as network delay arrive disaster recovery center prior to the A write request, and while writing in the far-end buffer memory of the second rear end platform of disaster recovery center prior to the A write request, because A write request and B write request all are marked with all issues and serial number, the second rear end platform is according to all issues and serial number, first the A write request is write to the far-end buffer memory, again the B request is write to the far-end buffer memory, and according in the far-end buffer memory, writing remote copy from LUN, thereby the inconsistent problem of data of having avoided remote copy master LUN and remote copy to store from LUN.
In the embodiment of the present invention, the storage system of disaster recovery center is distributed memory system, and each disaster tolerance cluster system is respectively used to process the write request of different physical address spaces.After the disaster tolerance cluster system receives the write request of production group system transmission identical with the physical address space of this disaster tolerance cluster system processing in the production center, judge whether to receive all write requests of one-period correspondence.Determining while having received all write requests, judge again the write request that whether has the physics address conflict in all write requests that this week issue is corresponding, to determine, the write request of this week issue is write to remote copy from LUN successively according to the serial number of mark wherein, or concurrently write remote copy from LUN.
It should be noted that, the another kind of asynchronous remote copying method that accompanying drawing 3 provides also can be applied to non-distributed storage system, in the disaster recovery center storage system, only comprises the situation of a disaster tolerance cluster system in accompanying drawing 4B.Non-distributed memory system also can be regarded as a special case of distributed storage system, its treatment scheme and accompanying drawing 4A are also substantially similar, only in step 41, owing to only having a disaster tolerance cluster system, thereby the write request received without restriction is the write request that the identical production group system of physical address space processed with this disaster tolerance cluster system in the production center sends, need to be only that the write request that send the production center gets final product.
Embodiment five
The distributed memory system structural representation of the production center that Fig. 2 B provides for the embodiment of the present invention.As shown in Fig. 2 B, the production center comprises at least one production group system, and each production group system comprises that first front-end platform 51 and first rear end platform 52, the first rear end platforms comprise the first business processing device 521 and remote copy master LUN522.
The first front-end platform 51, for after receiving the write request of main frame, according to all issues of current period, be write request mark week issue, and by mark week issue write request send under the first business processing device 521 in the first rear end platform 52 in the production group system.
The first business processing device 521, for by the week of mark issue write request write remote copy master LUN, the write request flag sequence number that is described mark week issue according to the order that writes described remote copy master LUN, and by mark week issue and the write request of serial number send to disaster recovery center.
Further, the production center comprises at least two production group systems, it is the storage system of the production center while being distributed memory system: the first front-end platform 51, for all issues according to current period, be also before described write request mark week issue, according to the logical address in the write request received and preserve logical address and the address mapping table of the mapping relations of physical address, determine the physical address of this write request, and in the physical address space processed in affiliated production group system of the physical address of judging this write request.
Further, to comprise at least two production group systems be a master control production group system and at least one non-master control production group system in the described production center;
The first front-end platform 51 in master control production group system, also for all issues of current period are set, and be handed down to each non-master control production group system by all issues of the current period of setting;
The first front-end platform 51 in described master control production group system, also for the cycle labelled notation of current period that self is arranged in described write request;
The first front-end platform 51 in described master control production group system, also for adding up in current period the write request number that receives described main frame; After current period finishes, described the first front-end platform in current period is received to the write request number of described main frame and all issues of current period send to disaster recovery center;
The first front-end platform 51 in non-master control production group system, all issues of the current period also periodically issued for the first front-end platform of obtaining master control production group system.
The first front-end platform 51 in described non-master control production group system, also for the cycle labelled notation of the current period by obtaining in described write request.
The working mechanism of above-mentioned each device, referring to describing in the corresponding embodiment of Fig. 2 A, does not repeat them here.
While comprising at least two production group systems in the heart aborning, in the embodiment of the present invention, the storage system of the production center is distributed memory system, and each production group system is respectively used to process the write request of different physical address spaces.The physical address space of processing comprises the production group system of the physical address of this write request, and all issues that the first front-end platform wherein issues according to master control the first front-end platform are write request mark week issue.The first business processing device of above-mentioned production group system is write request flag sequence number for the order that writes remote copy master LUN according to write request.Disaster recovery center receives write request and is marked with all issues and serial number, for the write request disaster recovery center of same period, can be in order number writes remote copy from LUN.Thereby a plurality of write requests of having avoided the conflict of in same period reference address arrive the order of disaster recovery centers when inconsistent with the write sequence of the heart aborning, the inconsistent problem of data that the data of storing in the remote copy master LUN of the caused production center and the remote copy of disaster recovery center are stored from LUN.
Embodiment six
The structural representation of the disaster recovery center embodiment that Fig. 4 B provides for the embodiment of the present invention.As shown in Figure 4 B, disaster recovery center comprises at least one disaster tolerance cluster system, and the disaster tolerance cluster system comprises that second front-end platform 61 and second rear end platform 62, the second rear end platforms 62 comprise that the second business processing device 621 and remote copy are from LUN622.Wherein,
The second front-end platform 61, the write request sent for receiving the production center, be marked with all issues and serial number in write request.
Further, when disaster recovery center comprises at least two disaster tolerance cluster systems: the second front-end platform 61, the write request sent specifically for receiving production group system identical with the physical address space of the affiliated disaster tolerance cluster system processing of this second front-end platform 61 in the production center.
The second business processing device 621, for judging whether the second front-end platform 61 has received all write requests of one-period correspondence, determining while having received all write requests of one-period correspondence, all issues and serial number according to mark in write request, write remote copy from LUN622 by the write request received.
Alternatively, described disaster tolerance cluster system also comprises far-end buffer memory 623.
The write request that described far-end buffer memory 623 receives for store the second front-end platform 61 in the chained list mode, each week corresponding chained list of issue particularly, the write request in chained list corresponding to each week issue sorts with serial number,
Described the second business processing device 621 judges according to the chained list of storage in far-end buffer memory 623 whether the second front-end platform 61 has received all write requests of one-period correspondence.
The working mechanism of above-mentioned each device, referring to describing in the corresponding embodiment of Fig. 4 A, does not repeat them here.
While comprising at least two disaster tolerance cluster systems in disaster recovery center, in the embodiment of the present invention, the storage system of disaster recovery center is distributed memory system, each disaster tolerance cluster system for the treatment of write request physical address space and other disaster tolerance cluster systems in this disaster recovery center for the treatment of the physical address space of write request different.Receive the handled physical address space of disaster tolerance cluster system of write request in disaster recovery center, the physical address space of processing with the production group system that sends this write request in the production center is identical.When the second business processing device of above-mentioned disaster tolerance cluster system receives the write request of the second front-end platform forwarding, writing of forwarding according to the second front-end platform please be counted and all issues, judges whether to receive all write requests of one-period correspondence.Determining while having received all write requests, judge again the write request that whether has the physics address conflict in all write requests that this week issue is corresponding, to determine, the write request of this week issue is write to remote copy from LUN successively according to the serial number of mark wherein, or concurrently write remote copy from LUN.
Embodiment seven
The structural representation of the second business processing device embodiment in the disaster recovery center that Fig. 5 provides for the embodiment of the present invention.As shown in Figure 5, the second business processing device 621 comprises: number of request judge module 6211, conflict judge module 6212, serial writing module 6213 and concurrent writing module 6214.
The second front-end platform, also the write request number for receiving all issues that the production center sends and receiving in corresponding cycle at this week issue.
Number of request judge module 6211, all issues and the serial number of all issues that send for the production center that receives according to the second front-end platform and write request number and the temporary write request mark received, judge whether to receive all write requests of one-period correspondence.
Conflict judge module 6212, for when number of request judge module 6211 is determined all write requests that receive the one-period correspondence, judge in all write requests that all issues are corresponding the write request that whether has the reference address conflict.
Serial writing module 6213, if judge for conflict judge module 6212 write request that has the reference address conflict, according to the serial number in write request, the write request that all issues is corresponding writes remote copy successively from LUN.
Concurrent writing module 6214, if judge for conflict judge module 6212 write request that does not have reference address conflict, by all issues, corresponding write request is concurrent writes remote copy from LUN.
Embodiment eight
The structural representation of the asynchronous remote copy system that Fig. 6 provides for the embodiment of the present invention.As shown in Figure 6, the production center comprises at least one production group system, and each production group system comprises that first front-end platform 51 and first rear end platform 52, the first rear end platforms 52 comprise the first business processing device 521 and remote copy master LUN522.Disaster recovery center comprises at least one disaster tolerance cluster system, and the disaster tolerance cluster system comprises that second front-end platform 61 and second rear end platform 62, the second rear end platforms 62 comprise that the second business processing device 621 and remote copy are from LUN622.
The working mechanism of above-mentioned each device, referring to describing in Fig. 2 B and the corresponding embodiment of Fig. 4 B, does not repeat them here.
One of ordinary skill in the art will appreciate that: realize that the hardware that all or part of step of said method embodiment can be relevant by programmed instruction completes, aforesaid program can be stored in a computer read/write memory medium, this program, when carrying out, is carried out the step that comprises said method embodiment; 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 embodiment only, in order to technical scheme of the present invention to be described, is not intended to limit; Although with reference to previous embodiment, the present invention is had been described in detail, those of ordinary skill in the art is to be understood that: its technical scheme that still can put down in writing aforementioned each embodiment is modified, or part technical characterictic wherein is equal to replacement; And these modifications or replacement do not make the essence of appropriate technical solution break away from the spirit and scope of various embodiments of the present invention technical scheme.