US20070192555A1 - Remote copy system - Google Patents

Abstract

In a system in which data employed by a computer is stored in a storage system, the storage system transfers this data to another storage system and a copy of the data is maintained in the other storage system. The consistency of the copy is maintained even when data is written, to the storage system by a computer, without having a write time applied. A source storage system, when a write time is applied to a write request, records the write time and applies this write time to the received write data and, when no write time is applied, applies the recorded write time to the received write data and transfers the write data with this write time applied thereto, to a target storage system. The target storage system stores the write data in a logical volume in the target storage system in accordance with the write time.

Description

CROSS-REFERENCES
• This application is a continuation application of U.S. Ser. No. 11/002,105, filed Dec. 3, 2004, which is a continuation application of U.S. Ser. No. 10/796,175, filed Mar. 10, 2004 (now U.S. Pat. No. 7,085,788.
BACKGROUND OF THE INVENTION
• 1. Field of the Invention
• The present invention relates to a storage system that stores data that is employed by a computer and that receives updating of data from a computer, and in particular relates to processing for maintaining copies of data between a plurality of storage systems.
• 2. Description of the Related Art
• In Laid-open European Patent Application No. 0672985, a technique is disclosed whereby the data that is employed by a computer is stored by a storage system and a copy of this data is stored in a separate storage system arranged at a remote location, while reflecting the write sequence of the data. In the processing indicated in Laid-open European Patent Application No. 0672985, the source storage system that has received the write data from the primary host computer reports completion of reception of the write data to the primary host computer only after reception of the write data. After this, the primary host computer reads a copy of the write data from the source storage system. A write time, which is the time at which the write request in respect of the write data was issued, is applied to this write data and, when the write data is read by the primary host computer, the write time is also transferred to the primary host computer. In addition, the primary host computer transfers the write data and the write time to the secondary host computer. After receiving the write data and the write time, the secondary host computer writes information including the write time to a control volume in the storage system on the secondary side and, in addition, writes the write data in the target storage system in the write time sequence, with reference to the write times at which the various items of write data were presented. By writing the write data in the target storage system in the write time sequence, it is possible to maintain consistent data in the target storage system.
• If write data were to be reflected to the target storage system neglecting the write sequence (the operation of storing write data in the target storage system will hereinbelow be referred to as “reflecting” the data), for example in the case of a bank account database, in processing to transfer funds from an account A to an account B, it would not be possible to reproduce the debiting of the account A and the crediting of the account B as a single transaction and it would be possible for example for a period to occur in the target storage system in which the balance of the account B was credited before debiting of the balance of the account A. If, in this case, some fault occurred in the source storage system rendering it unusable prior to debiting the balance of the account A in the target storage system, mismatching data would be left in the target storage system, with the result that incorrect processing would be performed if business were to be subsequently continued using the secondary host computer. Consequently, by storing the write data in the target storage system preserving the write sequence, consistent data can be maintained, making it possible to guarantee correctness of a sequence of related operations in respect of related data.
• U.S. Pat. No. 6,092,066 discloses a technique whereby the data that is used by a computer is stored in a storage system and, by copying the data that is stored in this storage system to a separate storage system arranged at a remote location, the data can be maintained in the separate storage system even if the first storage system has become unusable due to for example a natural disaster or fire.
• U.S. Pat. No. 6,209,002 discloses a technique whereby data employed by a computer is stored in a storage system and, by copying the data that is stored in this storage system to a separate storage system arranged at a remote location, and additionally copying the data that has been received by this separate storage system to a third storage system, a high level of redundancy can be obtained in respect of data.
SUMMARY OF THE INVENTION
• In the technique that is disclosed in Laid-open European Patent Application No. 0672985, consistency of the copy of data stored in the target storage system cannot be maintained unless the host computer applies a write time to the write data, since the write sequence is maintained using the write time applied to the write data by the host computer when the write data from the host computer is reflected to the target storage system. In the case of a so-called mainframe host computer, the write time is applied to the write request, but, in the case of a so-called open system host computer, the write time is not applied to the write request. Consequently, in the technique disclosed in Laid-open European Patent Application No. 0672985, consistency of the copy of the data stored in the target storage system with I/O from an open system host computer cannot be maintained.
• Also in the case of U.S. Pat. No. 6,092,066 and U.S. Pat. No. 6,209,002; there is no disclosure concerning maintenance of consistency of a copy of data stored in a target storage system when the host computers include an open system host computer.
• Accordingly, in a computer system in which data that is employed by computer is stored in a storage system and the data that is stored in this storage system is transferred to a separate storage system so that a copy of the data is also held in this separate storage system, there is herein disclosed a technique for maintaining consistency of the copy of the data stored in the separate storage system (i.e. the target storage system) even in respect of data written to the storage system by a host computer that does not apply a write time to the write data, such as an open system host computer.
• The system comprises a first storage device system having a first logical volume coupled to a computer and in which data received from the computer is stored and a second storage device system coupled to the first storage device system and having a second logical volume in which a copy of data stored in the first logical volume is stored.
• The first storage device system applies time information to the write data received from the computer and sends the write data and this time information to the second storage device system; the second storage device system stores the write data received from the first storage device system in the second logical volume in accordance with the time information applied to this write data.
• In a computer system in which data that is employed by computer is stored in a storage system and the data that is stored in this storage system is transferred to a separate storage system so that a copy of the data is also held in this separate storage system, it is thereby possible to maintain consistency of the copy of the data that is stored in the separate storage system (target storage system), even in the case of data stored in the storage system by a host computer that does not apply the write time to the write data, such as an open system host computer.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a view showing an example of-the layout of a computer system according toembodiment 1;
• FIG. 2 is a diagram showing an example of a logical volume group;
• FIG. 3 is a flow diagram showing an example of processing in the case where a write request is received by a storage device A;
• FIG. 4 is a view showing an example of group management information;
• FIG. 5 is a view showing an example of write data management information for managing write data;
• FIG. 6 is a flow diagram showing an example of transfer processing of write data from the storage device A to a storage device B;
• FIG. 7 is a view showing an example of remote logical volume information of a logical volume;
• FIG. 8 is a view showing an example of arrived write time information;
• FIG. 9 is a flow diagram showing an example of reflection processing of write data in the storage device B;
• FIG. 10 is a flow diagram showing another example of processing in the case where the storage device A has received a write request;
• FIG. 11 is a flow diagram showing another example of processing in the case where the storage device A has received a write request;
• FIG. 12 is a view showing an example of the layout of a computer system according to embodiment 2;
• FIG. 13 is a view showing an example of the layout of a computer system according to embodiment 3;
• FIG. 14 is a flow diagram showing another example of processing in the case where the storage device A in embodiment 3 has received a write request;
• FIG. 15 is a flow diagram showing an example of processing in the case where the management software A gives instructions for deferring processing of a write request in respect of the storage device A and creation of a marker;
• FIG. 16 is a view showing an example of marker number information;
• FIG. 17 is a view showing another example of write data management information;
• FIG. 18 is a flow diagram showing an example of transfer processing of write data from the storage device A in embodiment 3 to the storage device B;
• FIG. 19 is a flow diagram showing an example of reflection processing of write data in the storage device B in embodiment 3;
• FIG. 20 is a flow diagram showing another example of reflection processing of write data in the storage device B in embodiment 3; and
• FIG. 21 is a view showing an example of the layout of a computer system according to embodiment 4.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
• Embodiments of the present invention are described below. However, it should be noted that the present invention is not restricted to the embodiments described below.
Embodiment 1
• FIG. 1 is a view showing an example of the layout of a computer system according to a first embodiment.
• This system comprises a storage device (also referred to as a storage system). A100, a mainframe host computer A (also called MFA)600, an open system host computer A700, a storage device B190, a mainframe host computer B (also referred to as MFB)690 and an open system host computer B790. The storage devices A100 andMFA600 and the opensystem host A700 are respectively connected by I/O paths900. Thestorage device B190 andMFB690 and opensystem host B790 are also respectively connected by I/O paths900. The MFB.690 and opensystem host B790 are normally a standby system. TheMFA600,MFB690 and opensystem host A700 and opensystem host B790 are connected by anetwork920.
• TheMFA600 andMFB690 include anOS610 and application software (APP)620. Also, the opensystem host A700 and opensystem host B790 likewise include anOS710 andAPP720. An I/O request issued from the APP of theMFA600,MFB690, opensystem host A700, or opensystem host B790 through the OS is issued to thestorage device A100 orstorage device B190 through the I/O path900. In this case, software such as a DBMS is included in theAPP620 orAPP720.
• Thestorage device A100 comprises acontrol section200,control memory300 andcache400. Thecontrol section200 comprises a write datareception section A210 and write datatransfer section A220. Thecontrol section200 accesses thecontrol memory300 and performs the following processing, utilizing the information stored in thecontrol memory300. Thecache400 comprises high-speed memory that chiefly stores the read data or write data so that the storage device A can achieve a high I/O processing performance by employing thecache400. It should be noted that, preferably, these components are duplicated and provided with back-up power sources, for purposes of fault resistance and availability.
• Thestorage device B190 also comprises acontrol section200,control memory300 andcache400. Thecontrol section200 comprises a write datareception section B211 and write datareflection instruction section230 and writedata reflection section240. The role of thecontrol memory300 andcache400 is the same as in the description of thestorage device A100 above.
• Thestorage device A100 andstorage device B190 providelogical volumes500 constituting data storage regions in respect of theMFA600, opensystem host A700,MFB690 and opensystem host B790. It is not necessary that a singlelogical volume500 should constitute the single physical device; for example it could be constituted by a set of storage regions dispersed on a plurality of magnetic disc devices. Also, a logical volume may have for example a mirror construction or a construction that has redundancy such as for example a RAID construction, in which parity data is added.
• Thestorage device A100 provides alogical volume500 as described above; however, in the case of theMFA600 and opensystem host A700, the type oflogical volume500 that is provided is different from that provided in the case of thestorage device A100; also, the logical and/or physical interfaces of the I/O paths900 are different. The same applies to thestorage device B190,MFB690 and opensystem host B790. The time of thewrite request630 is included in thewrite request630 from theMFA600 as thewrite time650, but is not included in thewrite request730 from the opensystem host A700.
• Thestorage device A100 and thestorage device B190 are connected bytransfer paths910. As will be described, thestorage device A100 and thestorage device B190 can hold a copy of the content of one logical volume in another logical volume. In this embodiment, a copy of the content of thelogical volume500 of thestorage device A100 is held in thelogical volume500 of thestorage device B190; the content of the updating performed on thelogical volume500 of thestorage device A100 is also stored in thelogical volume500 of thestorage device B190 by being sent to thestorage device B190 through thetransfer path910. As will be described, thestorage device A100 and thestorage device B200 hold management information regarding the copy, indicating the relationship between the logical volumes and maintenance of the copy referred to above is performed by using this management information. The relationship between the logical volumes and the relationship of the logical volume groups, to be described, is set by the user in accordance with the user's needs.
• In this embodiment, the relationships between the logical volumes are grouped.FIG. 2 is a diagram showing an example of a group of logical volumes. The broken lines indicate the copy relationship between thelogical volumes500 or between the logical volume groups i.e. the correspondence relationship of the source and target. In this embodiment, the sequence of write data in thestorage device A100 and reflection in thestorage device B190 are managed in units of logical volume groups comprising a plurality of such logical volumes and allocation of the necessary resources for processing as described above is also performed in units of logical volume groups.
• If these are performed for each of the individual logical volumes, the large number of items to be managed makes the management process complicated and there is also a possibility of the resources required for this processing being increased, due to the large number of items to be processed. On the other hand, if the entirestorage device A100 is treated as a unit, detailed management can no longer be performed. In particular, since demands such as performance in regard to thelogical volumes500 differ greatly between a mainframe host and an open system host, it is desirable to arrange for example for manual control operations from the user in regard to processing and setting such as of tuning conditions to be accepted separately, by arranging for such hosts to perform processing separately, divided into respective groups. By setting up logical volume groups in this way, flexible copy processing management can be provided in response to the requirements of users or businesses.
• Next, processing of writing of data onto eachlogical volume500, transfer of data to astorage device B190 and processing for reflection of data in thestorage device B190 will be described for the case where thelogical volumes500 that are used by theMFA600 and the opensystem host A700 are arranged to belong to different logical volume groups. By means of these processes, reflection to a copy is performed in write sequence between the various logical volumes of thestorage device A100 and, regarding consistency between copies, it is arranged that mutual consistency can always be maintained between the mainframe host data and open system host data.
• FIG. 3 is a view showing the processing that is performed in the case where a write request is received from theMFA600 or opensystem host A700 in respect of a logical volume500 (logical volume500 constituting the source) where a copy of thelogical volume500 is being created. The write datareception section A210 receives a write request from theMFA600 or open system host A700 (step1000). If thewrite time650 is included in the write request that is received (step1001), the write datareception section A210 stores the write data in the cache400 (step1002) and creates (step1003) writedata management information330 by applying (assigning) a sequential number to the write data. The write datareception section A210 then records thewrite time650 in the writedata management information330. Also, when the sequential number is applied, the write datareception section A210 obtains the sequential number from thegroup management information310 of the logical volume group to which the logical volume that is being written belongs and records a value obtained by adding 1 thereto in the writedata management information330 as the sequential number of the write data, and records this new sequential number in thegroup management information310.
• FIG. 4 is a view showing an example ofgroup management information310 of the various logical volume groups. The group ID is the ID for identifying a logical volume group in thestorage device A100. The sequential numbers are numbers that are continuously given to write data in respect of a logical volume belonging to the logical volume group in question. Numbers successively increased by 1 in each case are applied to such write data, the initial value being for example 0. The logical volume number is the number of the logical volume that belongs to the logical volume group in question. The logical volume number is the ID of the logical volume belonging to the logical volume group in question in thestorage device A100. The remote storage device ID has a logical volume group that is paired with the logical volume group in question and is an ID (e.g. serial number) that specifies the storage device (in this embodiment, the storage device B190) where a copy of the content of the logical volume belonging to the logical volume group in question is stored. The remote group ID is an ID that specifies the logical volume group that is paired with the logical volume group in question in the remote storage device (storage device B190) i.e. the logical volume group to which the logical volume500 (also called the remote logical volume) belongs in which a copy of the content of the logical volume belonging to the logical volume group in question is stored.
• FIG. 5 is a view showing an example of writedata management information330 for managing the various write data. The logical volume ID is the ID of the logical volume in which the write data is stored. The write address is the write start address of the write data in question in the aforesaid logical volume. The write data length is the length of the write data in question. The write data pointer is the storage start address of the write data in question in thecache400. The sequential numbers are numbers that are continuously given to write data in the logical volume group and to which the logical volume belongs in which the write data is written. The write time will be discussed below. The “transfer required” bit is a bit that indicates whether or not the write data in question needs to be transferred to the storage device B and is set to ON when writedata management information330 is created by receipt of write data by the write datareception section A210. The writedata management information330 is managed in the form of a list for example for each logical volume group.
• Returning toFIG. 3, instep1004, the write datareception section A210 records thewrite time650 as thewrite time information340 in thecontrol memory300.
• If, instep1001, no write time is included in the write request, the write datareception section A210 stores the write data in the cache400 (step1005) and obtains from the write time information340 a write time, which it applies (assigns) to the write data, and creates write data management information330 (step1006) by applying a sequential number obtained from thegroup management information310. At this time, the write datareception section A210 then records the time at which thewrite time information340 was recorded, as the write time of the writedata management information300, and finds a sequential number by the same procedure as in the case ofstep1003 described above and records this sequential number in the writedata management information300.
• Finally, instep1007, completion of writing is reported to theMFA600 or to the opensystem host A700. The aforesaid processing does not include the time-consuming processing of physically writing the write data that is stored in thecache400 to the recording medium of thelogical volume500 or of transferring the write data to thestorage device B190; this processing is performed subsequently in asynchronous fashion, with an appropriate timing. Consequently, the time required until reporting of completion of writing after receiving the write request by the write datareception section A210 need only be a short time, so rapid response to theMFA600 or opensystem host A700 can be achieved.
• FIG. 6 is a view showing an example of transfer processing of write data to thestorage device B190 from thestorage device A100. The write datatransfer section A220 finds (step1100) the information relating to the write data that is transferred to thestorage device B190 by referring to the list of the writedata management information330 to find the write data that needs to be transferred and, in addition, referring to the writedata management information330,group management information310 and remotelogical volume information320. This information includes the write address acquired from the writedata management information330, the write data length, the sequential number, the write time, the remote storage device ID acquired from the remotelogical volume information320, the remote logical volume number, and the remote group number obtained from thegroup management information310 using the logical volume ID.
• FIG. 7 is a view showing an example of the remotelogical volume information320 of the various logical volumes. The logical volume ID is the ID of the logical volume on the source side (logical volume500 included in thestorage device A100 in embodiment 1). The remote storage device ID is an ID (for example a serial number) specifying the storage device (storage device B190 in embodiment 1) having the logical volume (also called the remote logical volume) in which is stored a copy of the data stored by the logical volume in question that is paired with the logical volume in question. The remote logical volume ID is an ID that specifies the remote logical volume (i.e. thelogical volume500 on the target side, where a copy of the data that was stored in the logical volume is stored) in the remote storage device (storage device B190 in embodiment 1).
• Next, returning toFIG. 6, the write datatransfer section A220 transfers (step1101) to thestorage device B190 the write data and the information found instep1100. The write datareception section B211 of the storage device B stores (step1102) the received write data and information in thecache400 and creates (step1103) writedata management information330 from the received information. The items of the writedata management information330 of thestorage device B190 are the same as the items of the writedata management information330 of thestorage device A100. The content of the writedata management information330 of thestorage device B190 differs from that of the writedata management information330 of thestorage device A100 in that the logical volume ID is the ID of thelogical volume500 on the target side where the copy is stored and the write data pointer is the storage start address of the write data in thecache400 of thestorage device B190 and the “transfer needed” bit is normally OFF, but is otherwise the same.
• Thestorage device B190 also hasgroup management information310, but the items thereof are the same as in the case of thestorage device A100. Regarding the content of thegroup management information310, the group ID is an ID that specifies the logical volume group to which thelogical volume500 on the side of the target where the copy is stored belongs, the remote storage device ID is the ID of the storage device (storage device A100 in the case of embodiment 1) constituting the source and the remote group ID is an ID that specifies the logical volume group to which the remote logical volume (i.e. thelogical volume500 constituting the source) belongs in the remote storage device (storage device A100 in embodiment 1). Thestorage device B190 also has remotelogical volume information320, but the items thereof are the same as in the case of thestorage device A100 and, regarding its content, the logical volume ID is an ID that specifies thelogical volume500 where the copy is stored, the remote storage device ID is an ID that specifies the ID of the storage device (storage device A100) constituting the source and the remote logical volume ID is an ID that specifies the remote logical volume (logical volume500 constituting the source) in the remote storage device (storage device A100).
• Returning toFIG. 6, next, the write datareception section B211 updates the arrived write time information350 (step1104).
• FIG. 8 is a view showing an example of arrivedwrite time information350 of the various groups. The group ID is an ID that specifies the logical volume group in thestorage device B190. The latest write time of the arrived write data is the latest time closest to the current time, of the write times applied to the write data received by the write data reception section ofB211, in respect of the logical volume groups of thestorage device B190. However, if it appears, from the sequential number order, that some of the write data has not yet arrived (some of the sequence of write data is missing), the latest time of the write time applied to these items of write data is recorded as the arrived write data time information, taking the continuous time comparison range in the order of the sequential numbers as being up to the final write data (write data immediately preceding the missing data).
• In transfer of the write data between the write datatransfer section A220 and the write datareception section B211, a plurality of items of write data may be simultaneously transferred in parallel. The write data is therefore not necessarily received in the write datareception section B211 in the order of the sequential numbers but, as will be described, the write data is reflected in the order of the sequential numbers to each of the logical volume groups (i.e. it is stored in the logical volumes of the storage device B190), so the write data is reflected to the copy in the order of updating (i.e. in the order of writing of the write data in the storage device A100).
• Returning once more toFIG. 6, finally, the write datareception section B211 reports completion of reception of the write data to the write data transfer section A220 (step1105). The write datatransfer section A220 of thestorage device A100 that has received this write data turns the “transfer required” bit of the writedata management information330 OFF in respect of the write data corresponding to the report of completion of reception of write data. At this time, thestorage device A100 may discard from the cache the arrived write data that was held for transfer to thestorage device B190.
• FIG. 9 it is a view showing an example of the reflection processing of write data in the storage device B190 (i.e. the processing of storage of the write data to the logical volume).
• The write data reflectioninstruction section B230 checks the arrivedwrite time information350 of all the logical volume groups of thestorage device B190 and finds, of these, the earliest time (step1200). The write data reflectioninstruction section B230 gives instructions (or permission) (step1201) to the write datareflection section B240 for reflection to these logical volumes of the write data whose write time is previous to the time that was thus found. When the writedata reflection section240 receives these instructions (or permission), by referring to the writedata management information330 andgroup management information310, it reflects the write data in the designated time range (i.e. the write data whose write time is previous to the time found in step1200), in the order of the write times, or, if these write times are the same, in the order of the sequential numbers in the various logical volume groups, in respect of thelogical volume500 in which the copy is stored (i.e. the write data is stored in the logical volume on the target side) (step1202). After completion of reflection of all of the write data in the range specified instep1202, the write datareflection section B240 reports completion of the instructed processing (step1203) to the write datareflection instruction section230. The storage device B may discard the reflected write data from thecache400.
• By means of the above processing fromstep1200 to step1203, one of cycle of reflection processing is completed. The write data reflectioninstruction section B230 and the write datareflection section B240 repeat the above cycle in order to reflect the write data transferred from the storage device A continuously.
• By means of the above processing, a copy of the updated data of thestorage device B190 is stored maintaining the order between updating of data by the mainframe host and updating of data by the open system host. Regarding data consistency between the copies, mutual consistency can be maintained between the data of the mainframe host and the data of the open system host.
• Specifically, thestorage device A100 utilizes thewrite time650 contained in thewrite request630 received from the mainframe host and applies a write time also to the write data received from the open system host and, furthermore, manages the received write data using both the write times and the sequential numbers. The targetstorage device B190 designates the write data that is capable of being reflected (i.e. that is capable of storage in a logical volume on the target side) using the sequential numbers and the write times and stores the designated write data in a logical volume on the target side. As a result, even if buffering and/or transferring are provided in parallel mid-way, write order is maintained between the data written from the mainframe host and the data written from the open system host, so copy data can be stored in a logical volume of thestorage device B190 on the target side.
• Also, even if some fault occurs in for example thestorage device A100, so that previously updated write data does not reach thestorage device B190, since the sequential numbers will not be continuous in respect of the write data of write times subsequent to the write time of the write data that failed to arrive, reflection thereof will not be allowed. Gaps of updating of data cannot therefore occur in the target sidestorage device B190 and consistency between the sourcestorage device A100 and targetstorage device B190 is ensured. As a result, even if a fault occurs in the sourcestorage device A100, business can be continued using the content of thelogical volume500 of thestorage device B190, which is matched with theMFB690 and/or opensystem host B790.
• Also, since, in the above processing, write times are applied to all of the write data received by thestorage device A100, irrespective of whether the host that employs the data is a mainframe host or open system host, it is possible to ascertain information such as up to which write time the write data in any desiredlogical volume500 has been transferred from thestorage device A100 to thestorage device B190 or has arrived at thestorage device B190 or has been reflected at the storage device B190 (i.e. has been stored in a logical volume).
• It should be noted that, in order to lighten the processing load in theabove step1202, the write data in the designated time range may be stored in thelogical volume500 that stores the copy in sequential number order in the various logical volume groups, neglecting the write time order. In this case, consistency between the copies (i.e. between the logical volumes of thestorage device B190 on the target side) is maintained by the timing of the reports of completion of processing instep1203. If it is desired to hold consistent data of the period between a report of completion of processing and the next report of completion of processing, a snapshot of thelogical volume500 in which the copy is stored may be acquired with the timing of the report of completion of processing. The technique disclosed in for example U.S. Pat. No. 6,658,434 may be employed as a method of acquiring such a snapshot. In this method, the storage content of a logical volume500 (source volume) in which is stored the data whereof a snapshot is to be acquired is copied to another logical volume500 (target volume) of thestorage device B190, so that the updated content is reflected also to the target volume when the source of volume is updated. However, in this embodiment, once the snapshot of the source volume has been stored in the target volume, the content of the target volume is frozen and verified by stopping reflection at that time.
• Also in the transfer processing of the above write data, it was assumed that, initially, the write datatransfer section A220 transfers the write data in respect of the write datareception section B211; however, it would be possible for the write datareception section B211 to initially issue a write data transfer request in respect of the writedata transfer section220 and for the write datatransfer section A220 to transfer the write data in respect of the write datareception section B211 after having received this request. By employing write data transfer requests, the pace of transfer of write data can be adjusted in accordance with for example the processing condition or load of thestorage device B190 or the amount of write data that has been accumulated.
• Also, in the above processing, it was assumed that the location of storage of the write data was thecache400; however, by preparing a separatelogical volume500 for write data storage, the write data could be stored in thislogical volume500. In general, alogical volume500 of large volume may be prepared in respect of thecache400, so this makes it possible for more write data to be accumulated.
• Also, in the above processing, it was assumed that thewrite time information340 was updated by thewrite time650 of reception from the mainframe host; however, it may be arranged for thestorage device A100 to possess an internal clock and to constantly update thewrite time information340 by reference to this clock. In this case,FIG. 10 shows an example of the processing that is executed when a write request in respect of a logical volume500 (logical volume500 constituting the source) where thestorage device A100 creates a copy is received from theMFA600 or opensystem host A700. This processing is processing corresponding to the processing shown inFIG. 3.
• The write datareception section A210 receives (step1300) a write request from theMFA600 or opensystem host A700. The write datareception section A210 stores (step1301) the write data in thecache400 and applies a write time to the write data by referring to thewrite time information340 that is constantly updated in accordance with the clock provided in thestorage device A100, and creates (step1302) writedata management information330 by applying a sequential number to the write data, by referring to thegroup management information310. Finally, completion of writing is reported to theMFA600 or open system host A700 (step1303).
• Also, in the above processing, a time is used in thewrite time information340 or the write time of the writedata management information300 or the arrivedwrite time information350; however; the time that is employed for this purpose need not necessarily be of the form of years, months, days, hours, minutes, seconds, milliseconds, microseconds, nanoseconds or a total of an ordinary time and instead a sequential number could be employed. In particular,FIG. 11 shows an example of the processing when thestorage device A100 has received a write request in respect of the logical volume500 (logical volume500 constituting the source), where the copy is created, from theMFA600 or opensystem host A700, in a case where thestorage device A100 itself updates thewrite time information340. This processing is processing corresponding toFIG. 3 orFIG. 10. It should be noted that, inFIG. 11, the initial value of thewrite time information340 may for example be 0 and numbers successively incremented by 1 may be applied to the write data as shown below as the write times.
• The write datareception section A210 receives a write request (step1400) from theMFA600 or opensystem host A700. The write datareception section A210 stores the write data in the cache400 (step1401), reads the number from thewrite time information340 and applies to the write data (step1402) as the write time the value obtained by incrementing this by 1. Then the write datareception section A210 records the value after incrementing by 1 as thewrite time information340, thereby updating the write time information340 (step1403). The write datareception section A210 also creates the write data management information330 (step1405) by applying a sequential number to the write data (step1404) by referring to thegroup management information310. The write datareception section A210 finally reports completion of writing (step1406) to theMFA600 or opensystem host A700.
• When a sequential number is employed as the write time in this manner, in thestorage device B190, instead of the write datareception section B211 being arranged to update the arrivedwrite time information350 using the write time applied to the write data received and the write data reflectioninstruction section B230 being arranged to designate the range of write data capable being stored in a logical volume of the storage device B by checking the arrivedwrite time information350 of the various logical volume groups, it may be arranged for the writedata reflection section240 to reflect (i.e. store) the write data arriving at the storage device B by referring to the sequential number recorded at the write time of the writedata management information330 in thelogical volume500 without skipping numbers in the number sequence.
Embodiment 2
• FIG. 12 is a view showing an example of the layout of a computer system according to a second embodiment.
• The differences with respect toembodiment 1 lie in that theMFA600 and opensystem host A700 are connected with thestorage device C180 through an I/O path900 and thestorage device C180 is connected with thestorage device A100 through atransfer path910. In this embodiment, a copy of the data stored in thelogical volume500 of thestorage device C180 is stored in alogical volume500 of thestorage device A100. Further, a copy of the data stored in thelogical volume500 of the storage device A is stored in thelogical volume500 of thestorage device B190 in processing like the processing described inembodiment 1. That is, in this embodiment, a copy of the data stored in thelogical volume500 of thestorage device C180 is stored in thestorage device A100 and thestorage device B190. In order to implement such processing, thestorage device C180 is provided with the various items of information and a construction like that of thestorage device A100 described inembodiment 1.
• When thestorage device C180 has received awrite request630 or awrite request730 for thelogical volume500 from theMFA600 or opensystem host A700, it stores the receivedwrite data640 or writedata740 in a logical volume in thestorage device C180 and transfers this to the write datareception section A210 of thestorage device A100. At this point, in contrast to the processing described inembodiment 1, thestorage device C180 sends notification of completion of writing to theMFA600 or opensystem host A700 after waiting for notification of completion of reception from the write datareception section A210, and thestorage device C180 is thereby able to guarantee that a copy of thewrite data640 or writedata740 that was written thereto is present in thestorage device A100. In this way, if for example due to the occurrence of some fault in thestorage device C180 or on thetransmission path910, transfer of data to thestorage device A100 has not succeeded, theMFA600 or opensystem host A700 will not deem write data that have not been transferred to thestorage device A100 to have been written but will only deem write data that have been received by thestorage device A100 to have actually been written; a copy as expected by theAPP620 on theMFA600 or theAPP720 on the opensystem host A700 will therefore exist on thestorage device A100. Furthermore, after all of the write data received by thestorage device A100 have been sent to thestorage device B190, a copy as expected will also exist on thestorage device B190, so, at the time where the processing executed by theMFA600 or opensystem host A700 was interrupted, theMFB690 or opensystem host B790 will be able to continue business using data as expected identical with the data that are recognized as having been written by theMFA600 or opensystem host A700.
• As initially indicated inembodiment 1, when thewrite time information340 is updated by thewrite time650 applied to the write data, the write datareception section C212 of thestorage device C100, if awrite time650 is included in the receivedwrite request630, records the write time also in the writedata management information330 and the write datatransfer section C222 also transfers this write time to the write data reception section A210 of thestorage device A100 when performing write data transfer. After receiving the write data and the write time, the write datareception section A210 processes the write data and the write time received from thestorage device C180 by the same method as the processing of thewrite request630 that was received from the mainframe host inembodiment 1; consistency between the copies stored in the logical volumes in thestorage device A100 is thereby maintained and consistency between the write data issued from the mainframe host and the write data issued from the open system host can thereby be maintained.
• In this way, even if, due for example to a large-scale disaster, faults occur in both of thestorage device C180 and thestorage device A100, business can be continued using the consistent content of thelogical volume500 of thestorage device B190, which was matched with theMFB690 and opensystem host B790. As indicated in the final part ofembodiment 1, when thewrite time information340 is updated from thestorage device A100 itself, transfer of the write time from thestorage device C180 is unnecessary, so that, after receiving the write data from thestorage device C180, the write datareception section A210 may perform processing on the write data like the processing ofFIG. 11 indicated in the latter part ofembodiment 1.
• It should be noted that there may be a plurality ofstorage devices C180 that connect to thestorage device A100.
• Also, although not shown, if the mainframe host and open system host are connected by an I/O path with thestorage device A100, the mainframe host or open system host that is connected with the storage device A may continue the business that was being conducted by theMFA600 or opensystem host A700 using the consistent content of alogical volume500 of thestorage device A100 that was matched therewith, in the event that a fault occurs in theMFA600 or opensystem host A700 orstorage device C180.
Embodiment 3
• FIG. 13 is a view showing an example of the construction of a computer system according to Embodiment 3.
• The chief differences with respect toembodiment 1 lie in that there are a plurality of respective storage devices A100 andstorage devices B190, theMFA600 and opensystem host A700 are connected through an I/O path900 respectively with a plurality of storage devices A100, theMFB690 and the opensystem host B790 are connected through an I/O path900 respectively with a plurality ofstorage devices B190, theMFA600 includesmanagement software A800 and theMFB690 includesmanagement software B890. Other differences will be described below.
• Hereinbelow, the processing in respect of writing performed to the variouslogical volumes500, transfer of write data to thestorage device B190 and the processing of reflection of write data in the storage device B190 (i.e. storage of the write data in the logical volume) will be described in respect of thelogical volumes500 employed by theMFA600 and the opensystem host A700. This processing ensures that mutual consistency is maintained between the data of the mainframe host and the data of the open system host in regard to consistency between copies respectively stored in the plurality of logical volumes that are possessed by the plurality ofstorage devices B190.
• FIG. 14 is a view showing an example of the processing when a write request in respect of the logical volume500 (logical volume500 constituting the source) in which a copy is created by thestorage device A100 is received from theMFA600 or opensystem host A700.
• The write datareception section A210 receives (step1500) a write request from theMFA600 or opensystem host A700. The write datareception section A210 stores the write data in the cache400 (step1501) or, as inembodiment 1, creates write data management information330 (step1502) by acquiring a sequential number by referring to thegroup management information310. Finally, the write datareception section A210 reports to theMFA600 or opensystem host A700 completion of writing (step1503). Thegroup management information310 is the same as that in the case ofembodiment 1. The writedata management information330 of this embodiment will be described later.
• FIG. 15 is a view showing an example of the processing when themanagement software A800 gives instructions for deferment of processing of write requests in respect of thestorage device A100 and creation of a marker. As will be described later, consistency is established between the copies stored in the plurality ofstorage devices B190 by subsequently performing synchronization of reflection to the copies, with the timing with which this processing was performed during updating of thelogical volume500 of thestorage device A100.
• First of all, themanagement software A800 gives instructions for deferment of processing of write requests to all of the storage devices A100 (step1600). On receipt of these instructions, the write datareception section A210 defers processing of write requests (step1601) and reports to themanagement software A800 the fact that deferment has been commenced (step1602). After themanagement software A800 has confirmed that commencement of deferment has been reported from all of the storage devices A100 that have been so instructed, processing advances to the following processing (step1603 and step1604).
• Next, themanagement software800 instructs all of the storage devices A100 to create markers (step1605). This instruction includes a marker number as a parameter. The marker number will be described subsequently. On receipt of this instruction, the markercreation section A250 records the received marker number in themarker number information360 shown inFIG. 16 stored in the control memory300 (step1606) and creates (step1607) special write data (hereinbelow called a marker) for information transmission in respect of all of the logical volume groups. A marker is write data in which a marker attribute is set in the writedata management information300.
• FIG. 17 is a view showing an example of writedata management information330 of write data in this embodiment a marker attribute bit and marker number are added to the writedata management information330 ofembodiment 1.
• The marker attribute bit is a bit indicating that the write data in question is a marker and is OFF in the case of ordinary write data but is set to ON in the case of a marker. A marker number as described above is set in the “marker number”. The sequential number in the group is acquired and applied in respect of a marker in the same way as in the case of ordinary write data. Specifically, in marker creation, the markercreation section A250 obtains a sequential number from thegroup management information310 of the group in the same way as in the processing of the write datareception section A210 and records a value obtained by adding 1 thereto in the writedata management information330 as the sequential number of the aforesaid marker, and records the new sequential number in thegroup management information310. When the sequential number has been applied in this way to the marker, it is transferred to thestorage device B190 in the same way as in the case of ordinary write data, but the marker is not reflected to thelogical volume500.
• The marker number is a number for identifying the instruction in response to which the marker was created; when a marker creation instruction is issued by themanagement software A800, for example the initial value thereof is 0 and the marker number is incremented by 1 before being issued. Themanagement software A800 may confirm the current marker number by reading the marker number recorded in themarker number information360.
• Returning toFIG. 15, after the markercreation section A250 has created a marker in respect of all of the logical volume groups, the markercreation section A250 reports completion of marker creation to the management software A800 (step1608). After confirming that completion of marker creation has been reported from all of the designated storage devices A100, themanagement software A800 proceeds to the subsequent processing (step1609, step1610).
• Themanagement software A800 gives instructions (step1611) for cancellation of deferment of processing of write requests to all of thestorage devices A100. On receipt of these instructions, the write datareception section A210 cancels deferment of processing of write requests (step1612) and reports to the management software A800 (step1613) the fact that such deferment has been cancelled.
• FIG. 18 is a view showing an example of transfer processing of write data to astorage device B190 from astorage device A100. This processing is substantially the same as the transfer processing described inFIG. 6 ofembodiment 1, but differs in that no updating of the arrivedwrite time information350 is performed by the write datareception section B211. It should be noted that the writedata management information330 of thestorage device B190 is the same as the write data management information shown inFIG. 17, described above; instep1703, the presence or absence of the marker attribute of the write data and/or the marker number recorded in the writedata management information330.
• FIG. 19 is a view showing an example of the processing of reflection (storage) of write data to a logical volume in thestorage device B190. First of all, themanagement software B890 gives instructions for reflection of the write data, as far as the marker, to thelogical volume500 in which a copy is stored (step1800) to all of thestorage devices B190. After receiving such an instruction, the write datareflection section B240 refers to thewrite data information330 andgroup management information310 and reflects (step1801) the write data as far as the marker, in the sequential number order in each group, to thelogical volume500 in which the copy is stored. Specifically, the write datareflection section B240 continues to store the write data in the logical volume in the order of the sequential numbers, but stops data storage processing on finding write data with the marker attribute (i.e. a marker) and then reports completion of reflection to the management software B890 (step1802). In the aforementioned processing, the write datareflection section B240 checks the marker numbers of the markers that are recorded in the writedata management information330 and thereby ascertains whether the marker number is correct (whether the marker conforms to rules which are the same as the marker number decision rules, described above, for example of being a number whose initial value is 0 and that is incremented by 1 with respect to the previous marker number). If the marker number is not correct, the write datareflection section B240 reports an abnormal situation to themanagement software B890; if the marker number is correct, the write datareflection section B240 records the marker number in themarker number information360 and reports a normal situation. Themanagement software B890 may confirm the current marker number by reading the marker number that is recorded in themarker number information360.
• After confirming that a “normal reflection completed” report has been obtained from all of thestorage devices B190 that had been designated, themanagement software B890 proceeds to the next processing (step1803, step1804).
• Next, themanagement software B890 gives instructions (step1805) for updating of the snapshot of thelogical volume500 that stores the copy to all of thestorage devices B190. After receiving this instruction; the snapshotacquisition section B260 updates (step1806) the snapshot of the content of thelogical volume500. As the method of acquiring such a snapshot, for example the technique disclosed in U.S. Pat. No. 6,658,434 may be employed. It should be noted that, in this embodiment, just as in the case of the method described inembodiment 1, reflection of the write data to the volume that stores the snapshot data is stopped at the time of acquisition of the snapshot, and the content of the volume that stores the snapshot is frozen. After updating the snapshot, the snapshotacquisition section B260 reports completion of snapshot updating to the management software B890 (step1807). After confirming that a report of completion of snapshot updating has been obtained from all of thestorage devices B190 that were designated, themanagement software B890 proceeds to the next processing (step1808, step1809).
• Themanagement software A800 and themanagement software B890 respectively repeat the processing of theaforesaid step1600 to step1613 and ofstep1800 to step1809. In this way, the updating of thestorage device A100 to thelogical volume500 is constantly reflected to thelogical volume500 of thestorage device B190.
• By processing as described above, the data updating by theMFA600 and the opensystem host A700 is stopped and a marker is created with the timing (checkpoint) at which the updating condition is unified between the plurality of storage devices; reflection (i.e. storage) of the updated data to the stored copy data in the plurality of target logical volumes provided in the plurality of targetstorage devices B190 can be synchronized at the time immediately preceding the writing of the marker, so mutual consistency between the various copies can be obtained with the data of the mainframe host and the data of the open system host at the time of this marker. In addition, theMFB690 or opensystem host B790 can continue business using the matched data stored in the snapshot volume, since a copy having mutual consistency is held in the snapshot volume, this snapshot being acquired by reflection of the updated data to the copy data at a time that is synchronized between the plurality of copy data.
• In the above processing, the snapshot was assumed to be updated by thestorage device B190 in response to an instruction from themanagement software B890, but it would be possible to update the snapshot with the timing of synchronization of reflection of the updated data between the copy data of a plurality ofstorage devices B190.FIG. 20 shows an example of the reflection processing of write data to the copy in thestorage devices B190 in this case.
• Themanagement software B890 gives instructions (step1900) for reflection of the write data as far as the marker to the logical volume of500 that stores the copy in all of thestorage devices B190. After receiving such an instruction, the write datareflection section B240 reflects the write data in the same way as in the processing described with reference toFIG. 19 but stops the reflection as soon as it finds a marker and notifies the snapshot acquisition section B260 (step1901). After receiving such notification, the snapshotacquisition section B260 updates the snapshot of the content of thelogical volume500 and notifies the write data reflection section B240 (step1902). After receiving this notification, the write datareflection section B240 reports completion of reflection to the management software B890 (step1903). Themanagement software B890 confirms that a report of completion of snapshot updating has been obtained from all of thestorage devices B190 that were designated and then proceeds to the next processing (step1904, step1905).
• Also, in the aforesaid processing, it was assumed that thestorage device A100 orstorage device B190 reported completion of processing in respect of the various types of instructions from themanagement software A800 ormanagement software B890. However, it would also be possible for completion of the various types of processes by thestorage device A100 orstorage device B190 to be detected by themanagement software A800 ormanagement software B890 by themanagement software A800 ormanagement software B890 periodically making inquiries of thestorage device A100 orstorage device B190 regarding their processing condition in respect of the aforesaid instructions.
• Also, in the above processing, transfer processing of write data from thestorage device A100 to thestorage device B190 is performed continuously, but it would be possible for thestorage device A100 to create a marker and to then stop transfer of write data and, in addition, for thestorage device B190, after detecting reflection processing of the received marker (after reflection of the write data previous to the marker) to stop reflection of the write data i.e. to put the processing by thestorage device A100 andstorage device B190 in a stopped condition (also called a suspended condition). However, thestorage device B190 could perform write data reflection up to the detection of the marker without reference to instructions from themanagement software B890. In this case, the marker creation instruction is equivalent to an instruction to shift to the suspended condition and mutually matched copies are created in thelogical volume500 of thestorage device B190 at the time where all of thestorage devices B190 have shifted to the suspended condition. When restarting the copy processing, the copy processing is recommenced by thestorage device A100 andstorage device B190 in response to an instruction for recommencement of copy processing from themanagement software A800 ormanagement software B890 after acquisition of the snapshot of thelogical volume500. As a result, copies having mutual consistency can be held in data stored by the snapshots, soMFB690 or opensystem host B790 can continue business using the matched data.
• Also, in the processing described above, the various types of instructions, reports and exchange of information between themanagement software A800 ormanagement software B890 andstorage device A100 andstorage device B190 may be executed by way of an I/O path900 or could be executed by way of anetwork920. In the case where instructions for marker creation are given in the form of a write request to thestorage device A100, alogical volume500 that is not subject to the processing deferment of write instructions is provided at thestorage device A100 and the marker creation instructions are given in respect of thislogical volume500.
• In the above processing, thestorage device A100 andstorage device B190 need not be connected in one-to-one relationship and it is not necessary that there should be the same number of devices, so long as the respectivelogical volumes500 and logical volume groups correspond as source and copy.
• Also, in the above construction, it was assumed that themanagement software A800 was present in theMFA600 and themanagement software B890 was present in theMFB690; however, it would be possible for themanagement software A800 andmanagement software B890 to be present in any of theMFA600,MFB690, opensystem host A700, opensystem host B790,storage device A100 orstorage device B190. Also, they could be present in another computer, not shown, connected with thestorage device A100 orstorage device B190.
• In the above processing, it was assumed that the write datareflection section B240 determined the correct marker number, but it would also be possible for the correct marker number to be designated to thestorage device B190 as a parameter of the reflection instructions by the management software B. Also, it could be arranged that when themanagement software A800 gives instructions for deferment of processing of write requests and marker creation to thestorage device A100, a unique marker number is determined and designated to thestorage device A100 and communicated to themanagement software A890 and that thismanagement software B890 then designates this marker number to thestorage device B190.
• In the above processing, the occasion at which themanagement software A800 instructions for deferment of processing of write requests and marker creation to thestorage device A100 may be determined in a manner linked with the processing of theAPP620 orAPP720. For example, synchronization of reflection to the copy may be performed at the checkpoint by giving instructions for deferment of write request processing and marker creation on the occasion of creation of a DBMS checkpoint. Business can therefore be continued by theMFB690 or opensystem host B790 using the data of this condition, by obtaining a snapshot in the condition in which the stored content of the sourcelogical volume500 at the checkpoint has been reflected to the copy in the target logical volume.
• It could also be arranged for theMFA600 or opensystem host A700 to defer issue of a write request to thestorage device A100 or to restart, by linking theOS610 orOS710 with themanagement software A800, instead of themanagement software A800 giving instructions for deferment of processing of write requests and canceling of deferment in respect of thestorage device A100.
• Also, as described inembodiment 1, a logical volume for write data storage that is separate from thecache400 could be prepared and the write data stored in thislogical volume500 for write data storage. Also, in the transfer processing of write data, it would be possible for a write data transfer request to be initially issued in respect of the writedata transfer section220 by the write datareception section B211 and for the write data to be transferred in respect of the write datareception section B211 by the write datatransfer section A220 after receiving this request.
• The processing described in this embodiment could also be implemented even if the write request does not contain a write time.
Embodiment 4
• FIG. 21 is a view showing an example of the layout of a computer system in embodiment 4.
• The difference with respect to Embodiment 3 lies in that theMFA600 and the opensystem host A700 are respectively connected with a plurality ofstorage devices C180 by way of an I/O path900 and the plurality ofstorage devices C180 are connected with a plurality of storage devices A100 by way of atransfer path910. In addition, the plurality ofstorage devices C180 are connected with another computer or device by means of anetwork920. Thestorage device A100 and thestorage device B190 of embodiment 4 have the same construction and function as thestorage device A100 andstorage device B190 in embodiment 3.
• In this embodiment, just as in the case of embodiment 2, a copy of the data stored in thelogical volume500 of thestorage device C180 is stored in thelogical volume500 of thestorage device A100. Specifically, thestorage device C180 comprises the same construction and various types of information as in embodiment 2 and after receiving a write request to thelogical volume500 from theMFA600 or opensystem host A700, thestorage device C180 stores the write data that it has received and transfers this received write data to the write datareception section A210 of thestorage device A100; however, it is then guaranteed that a copy of thewrite data640 or writedata740 that was written by thestorage device C180 exists in thestorage device A100, by sending a write completion notification to theMFA600 or opensystem host A700 after waiting for a notification of completion of reception from the write datareception section A210, in the same way as in embodiment 2.
• In addition, a copy of the data stored in thelogical volume500 of thestorage device C180 is stored in alogical volume500 of thestorage device B190 by the same processing as the processing described in embodiment 3. By processing as described above, as described in embodiment 2, even if for example some fault occurs in thestorage device C180 or in thetransfer path910, causing transfer of data to thestorage device A100 to become impossible, the expected content that was recognized as having been stored in the storage device C186 when processing of theMFA600 or opensystem host A700 was interrupted can still be obtained from thestorage device B190, so theMFB690 or opensystem host B790 can continue business using this data.
• In the above processing, themanagement software A800 gives instructions for deferment of processing of write requests or marker creation or cancellation of deferment of processing of write requests in respect of all of thestorage devices C180 in the same way as in the case of the processing performed in respect of thestorage device A100 in embodiment 3. Just as in the case ofstep1600 of embodiment 3, themanagement software A800 first of all gives instructions for deferment of processing of write requests to all of thestorage devices C180. After receiving these instructions, the write datareception section C212 of thestorage device C180 defers processing of write requests in the same way as in the case of the processing performed by thestorage device A100 instep1601 and step1602 of embodiment 3 and reports commencement of deferment to themanagement software A800. As described above, at this time, write data in respect of which a write completion notification has been given in respect of theMFA600 or opensystem host A700 has already been transferred to thestorage device A100 and thestorage device A100 creates writedata management information300 of this write data. In the same way as in the case ofstep1603 and step1604 of embodiment 3, themanagement software A800 confirms that a report of commencement of deferment has been obtained from all of the designatedstorage devices C180 before proceeding to the following processing.
• Next, themanagement software A800 gives instructions for marker creation to all of thestorage devices C180 in the same way as in thestep1605 of embodiment 3. After receiving such an instruction, thestorage device C180 transmits a marker creation instruction through thepath910 ornetwork920 to thestorage device A100 that stores the copy. After receiving the marker creation instruction, thestorage device A100 creates a marker in the same way as instep1606,step1607 and step1608 of embodiment 3 and reports completion of marker creation to thestorage device C180 through thetransfer path910 ornetwork920. After receiving the report, thestorage device C180 reports completion of marker creation to themanagement software A800. Themanagement software A800 confirms that a report of completion of marker creation has been received from all of the designatedstorage devices C180 in the same way as instep1609 and step1610 of embodiment 3 before proceeding to the next processing.
• Next, themanagement software A800, in the same way as instep1611 of embodiment 3, gives instructions for cancellation of deferment of processing of write requests to all of thestorage devices C180. After receiving these instructions, the write datareception section C212 of thestorage device C180 cancels the write request processing deferment in the same way as the processing that was performed by thestorage device A100 instep1612 and step1613 of embodiment 3 and reports this cancellation of deferment to themanagement software A800.
• Specifically, deferment of processing of write requests and cancellation of deferment are performed by thestorage device C180 and marker creation meanwhile is performed by thestorage device A100 on transmission to thestorage device A100 of an instruction by thestorage device C180. As described above, write data in respect of which completion of writing has been notified to theMFA600 or opensystem host A700 has already been transferred to thestorage device A100 and writedata management information300 of such write data is created in thestorage device A100, so deferment of processing of write requests by thestorage device A100 in embodiment 3 and deferment of processing of write requests by thestorage device C180 in this embodiment are equivalent. Consequently, by performing processing as described above and by performing other processing as described in embodiment 3, in the construction of this embodiment, reflection of updating to the copies can be synchronized at the marker time by stopping data updating by theMFA600 and opensystem host A700 in the same way as in embodiment 3 and creating a marker of the updated condition with unified timing (checkpoint) between the plurality of storage devices; mutual consistency of the respective copies with the mainframe host data and the open system host data can thus be achieved at this time. Furthermore, mutually matched copies are maintained in snapshot volumes by acquiring snapshots at the time of synchronization of reflection and theMFB690 or opensystem host B790 can therefore continue business using matched data.
• In the above processing, it was assumed that themanagement software A800 gave instructions for marker creation to thestorage devices C180 and thestorage devices C180 transmitted these instructions to the storage devices A100; however, it would also be possible for themanagement software A800 to give instructions for marker creation directly to all of the storage devices A100 and for the storage devices A100 to report completion of marker creation to themanagement software800. Specifically, themanagement software A800 first of all gives instructions for deferment of write request processing to all of thestorage devices C180 and themanagement software A800 confirms that reports of commencement of deferment have been received from all of the designatedstorage devices C180 before giving instructions for marker creation to all of the storage devices A180 in the same way as instep1605 of embodiment 3. After having received these instructions, thestorage device A100 creates a marker in the same way as instep1606,step1607 and step1608 of embodiment 3 and reports completion of marker creation to themanagement software800. After confirming that reports of completion of marker creation have been obtained from all of the designated storage devices A100 in the same way as instep1609 and step1610 of embodiment 3, themanagement software A800 may be arranged to give instructions for the cancellation of deferment of write request processing to all of thestorage devices C180.
• Also, it would be possible that thestorage devices C180 are provided with a marker creation section andmarker number information330 and create a marker on receipt of instructions for marker creation from themanagement software A800; the marker, which has been created as write data, is then transferred to thestorage device A100 and completion of marker creation may be arranged to be reported to themanagement software A800 when a report of receipt thereof has been received from the writedata reception section210 of thestorage device A100. In this case, thestorage device A100 treats the received marker as a special type of write data, which is transferred to thestorage device B190 after processing in the same way as ordinary write data except that reflection to the copy is not performed.
• In any case, the above can be implemented irrespective of the number ofstorage devices C180 that are connected with the storage devices A100 and deposit copies on thestorage devices A100.
• Also, although not shown, if a mainframe host and open system host are connected with the storage devices A100 by an I/O path, if for example some fault occurs in theMFA600 or opensystem host A700 orstorage devices C180, the aforesaid mainframe host and open system host can continue business using the content of thelogical volume500 of thestorage device A100 that is matched therewith.

Claims (20)

1. A remote copy system comprising:

a plurality of first storage systems, each comprising a first control section to be coupled to a computer, a first logical volume, and a first storage area for storing data to be transmitted to a second storage system; and

a plurality of second storage systems, each coupled to a first storage system and comprising a second control section, a second logical volume, and a second storage area for storing data received from the first storage system,

wherein a plurality of pairs, each including a first logical volume for storing data received from the computer and a second logical volume for storing replicate data of data in the first logical volume, are configured between the plurality of first storage systems and the plurality of second storage systems,

wherein first control sections of the plurality of first storage systems receive a plurality of write requests from the computer, the first control sections store write data, which is received according to the plurality of write requests and to be stored in first logical volumes of the plurality of first storage systems, and write order information assigned to the write data in first storage areas of the first storage systems, and transmit write completion reports to the computer,

wherein when one of the plurality of first storage systems instructs the rest of the plurality of first storage systems to defer operations for executing write requests, the operations for executing write requests are deferred in the plurality of first storage systems, the one of the plurality of first storage systems sends information related to a marker to the rest of the plurality of first storage systems, and the first control sections of the plurality of first storage systems store markers and the write order information assigned to the markers in the first storage areas of the plurality of first storage systems and restart the operations for executing write requests, and

wherein second control sections of the plurality of second storage systems read the write data with the write order information and the markers with the write order information from the plurality of first storage systems, and control operation such that the write data received from the plurality of first storage systems is stored in second logical volumes of the plurality of second storage systems according to the write order information and the markers.

2. A remote copy system according toclaim 1,

wherein when the second control sections read the write data with the write order information and the markers with the write order information from the plurality of first storage systems, the second control sections send a plurality of transmission requests to the plurality of first storage systems, receive the write data with the write order information and the markers with the write order information from the plurality of first storage systems in response to the plurality of transmission requests, and store the received write data with the write order information and the markers with the write order information in second storage areas of the plurality of second storage systems, and

wherein one of the plurality of second storage systems instructs the rest of the plurality of second storage systems to store write data to second logical volumes, and the second control sections control to store write data from the second storage areas to the second logical volumes according to the write order information assigned to the write data, until reaching the markers.

3. A remote copy system according toclaim 2, wherein the write order information assigned to the write data or the markers in a sequential number indicates a receiving order of the write data from the computer.

4. A remote copy system according toclaim 2,

wherein the information related to the marker includes a number, and the first control sections store the markers each including the number in the first storage areas.

5. A remote copy system according toclaim 4,

wherein the one of the plurality of first storage systems instructs the rest of the plurality of first storage systems to defer operations for executing write requests and send the information related to the marker to the rest of the plurality of first storage systems repeatedly, and

wherein the one of the plurality of first storage systems increases a value of the number included in the information related to the marker upon sending the information related to the marker to the rest of the plurality of first storage systems.

6. A remote copy system according toclaim 4,

wherein when the one of the plurality of second storage systems instructs the rest of the plurality of second storage systems to store write data to the second logical volumes, the one of the plurality of second storage systems designates the number included in the markers, the second control sections control to store the write data from the second storage areas to the second logical volumes according to the write order information assigned to the write data, until reaching the marker including the number designated by the one of the plurality of second storage systems.

7. A remote copy system according toclaim 2,

wherein a pace of transmission of write data from a first storage system to a second storage system is controlled by a transmission request issued from a second control section of the second storage system to the first storage system.

8. A remote copy system according toclaim 7,

wherein the pace of the transmission of write data is controlled based on the amount of write data received by the second storage system.

9. A remote copy system according toclaim 7,

wherein the pace of the transmission of write data is controlled based on a load of the second storage system.

10. A remote copy system according toclaim 2,

wherein each first storage area includes a storage area in a logical volume configured by at least one physical storage device.

11. A remote copy system comprising:

a plurality of first storage systems, each comprising a first control section to be coupled to a computer, a first logical volume, and a first storage area for storing data to be transmitted to a second storage system; and

a plurality of second storage systems, each coupled to a first storage system and comprising a second control section, a second logical volume, and a second storage area for storing data received from the first storage system,

wherein a plurality of pairs, each including a first logical volume for storing data received from the computer and a second logical volume for storing replicate data of data in the first logical volume, are configured between the plurality of first storage systems and the plurality of second storage systems,

wherein when a first control section of a first storage system receives a write request from the computer, the first control section stores write data, which is to be stored in a first logical volume included in a pair according to the write request, and write order information assigned to the write data in a first storage area of the first storage system, and transmits a write completion report to the computer,

wherein when one of the plurality of first storage systems instructs the rest of the plurality of first storage systems to defer operations for executing write requests, the operations for executing write requests are deferred in the plurality of first storage systems, the one of the plurality of first storage systems sends information related to a marker to the rest of the plurality of first storage systems, and each first control section stores a marker and the write order information assigned to the marker in a first storage area of a first storage system including the first control section, and the operation for executing write requests are restarted in the plurality of first storage systems,

wherein a second control section of a second storage system, which includes a second logical volume included in a pair, issues a transmission request to a first storage system, which includes a first logical volume included in the same pair as the second logical volume repeatedly, so that in response to one or more transmission requests issued from the second control section, both the write data with the write order information and the marker with the write order information are transmitted from the first storage system to the second storage system, and the second control section stores the write data with the write order information and the marker with the write order information in a second storage area of the second storage system, and

wherein when one of the plurality of second storage systems instructs the rest of the plurality of second storage systems to store write data to second logical volumes, each second control section controls to store write data from a second storage area of a second storage system including the second control section to a second logical volume of the second storage system, according to the write order information assigned to the write data, until reaching the marker.

12. A remote copy system according toclaim 11,

wherein the information related to the marker includes a number, and each first control section stores the marker including the number in the first storage area.

13. A remote copy system according toclaim 12,

wherein the one of the plurality of first storage systems instructs the rest of the plurality of first storage systems to defer operations for executing write requests and sends the information related to the marker to the rest of the plurality of first storage systems repeatedly, and

wherein the one of the plurality of first storage systems increases a value of the number included in the information related to the marker upon sending the information related to the marker to the rest of the plurality of first storage systems.

14. A remote copy system according toclaim 12,

wherein when the one of the plurality of second storage systems instructs the rest of the plurality of second storage systems to store write data to the second logical volumes, the one of the plurality of second storage systems designates the number included in the marker, and each second control section controls to store the write data from the second storage area to the second logical volume according to the storage area to the second logical volume according to the write order information assigned to the write data, until reaching the marker including the number designated by the one of the plurality of second storage systems.

15. A remote copy system according toclaim 11,

wherein a pace of transmission of write data from a first storage system to a second storage system is controlled by using the transmission request issued from a second control section of the second storage system to the first storage system.

16. A remote copy system according toclaim 15,

wherein the pace of the transmission of write data is controlled based on the amount of write data received by the second storage system.

17. A remote copy system according toclaim 15,

wherein the pace of the transmission of write data is controlled based on a load of the second storage system.

18. A remote copy system according toclaim 11,

wherein each first storage area includes a storage area in a logical volume configured by at least one physical storage device.

19. A remote copy system according toclaim 11,

wherein each second storage area includes a storage area in a logical volume configured by at least one physical storage device.

20. A remote copy system according toclaim 11,

wherein the write order information assigned to the write data or the marker in a first storage system is a sequential number, which indicates a receiving order of the write data from the computer.

Images