Patent Application
20090132753
The present invention relates to database systems, and more particularly to systems, methods and computer code for processing a request to UNDO or REDO data replication in a database system.
Database systems typically allow administrators or other authorized users to reset the database to an older state. Generally, such resets require that data be retrieved from an external backup of the database. For example, if the data volumes storing the database are backed up daily, an administrator may be able to reset the database to the previous day's state by retrieving the backup data from the previous day. Such backup techniques are desirable for many purposes, and ensure that data is recoverable in the event of catastrophic system errors. It is also desirable to be able to avoid the loss of data during the replication or copying process.
Many replication management systems allow administrators or other authorized users to restore the data and databases in the event of a crash or other error. For example, in the event of a crash or other error, the last known good version of the database may be retrieved and brought up to date to the time of the crash or error using data from a data log or point in time (PIT) copy of the database. This process takes considerable time, requires many steps and is error prone. Thus it would be beneficial to reduce the time, complexity and inherent errors in locating, loading and restoring a database.
At times, an administrator or other user may wish to more frequently and easily recover or restart a database in an earlier known state. In such situations, the use of existing techniques would decrease the availability of the database system and require valuable time, hardware and other resources. It would be desirable to provide improved systems, methods and computer code for recovering and restarting databases in a known earlier state and to ensure that an immediate prior version is available upon the initiation of the replication process. Such systems, methods and computer code should not incur unduly burdensome overhead. It would also be desirable to provide a system, method and computer readable medium with program instructions for reverting a database to a prior state by use of a single command. In addition, it would be desirable to have the capability to selectively process UNDO and REDO commands during data replication when a target volume is overwritten with data from a source volume thus preventing the loss of valuable data. The present invention addresses such needs and provides such systems, methods and computer code.
The present invention provides a system, method and computer readable medium with program instructions for replicating a volume of data including initiating a point in time copy command, selecting a target storage box to receive a copy of the volume of data, selecting a point in time storage box to receive a copy of the volume of data and managing the data to compare a size of the copy of the volume of data to a size of the point in time storage box. Upon comparison and confirmation that the size of the point in time storage box is large enough to accept the volume of data, the point in time copy of the volume of data is initiated. The data is then managed to selectively UNDO replication of the volume of data and to remove a data replication relationship to restore the point in time storage box copy of the volume of data to the target storage box. Further, the data can be managed to selectively UNDO replication of the volume of data and to remove a data replication relationship to restore the point in time storage box copy of the volume of data to the target storage box upon the user's single command or automatically as controlled by a management function.
In addition, the present invention provides a system, method and computer readable medium with program instructions to select a second point in time storage box to receive a copy of the volume of data from the target storage box. The system, method and computer readable medium with program instructions of the present invention will manage the data to selectively REDO replication of the volume of data to restore the second point in time storage box copy of the volume of data to the target storage box. The system, method and computer readable medium with program instructions can also select multiple point in time storage boxes to each receive a different copy of the volume of data from the target storage box and each of which can be saved such that the copy of each point in time storage box can be returned to the target storage box to restore the respective volume of data. Further, the point in time storage boxes may be flash memory media. In addition, the system, method and computer readable medium with program instructions of the present invention provides that the oldest version of the multiple point in time storage boxes is discarded if storage space is not available to save the next most current multiple point in time storage box upon initiation of the point in time copy.
As shown in
The details of UNDO sequence are shown in
Referring to
At block 416 the first PIT copy is initiated and will continue at block 418 until the UNDO function is selected at block 420 or the first PIT copy is completed at block 422. If the replication process completes at block 422 then the process will stop at block 414 and the system will be available to begin another replication process upon the user's command.
If the UNDO function is available and selected at block 420, then the data replication process will stop and a second PIT storage box will be selected at block 424. At block 426 the size of the volume of data in the target storage box is compared to the size of the second PIT storage box to determine if the REDO function can be carried out. If the size of the volume of data in the target storage box is equal to or smaller than the second PIT storage box, then the REDO function will be available the upon completion of the UNDO sequence. If the volume of data in the target storage box is larger than the second PIT storage box then the REDO function will not be available upon completion of the UNDO process and a warning will be issued at block 428. At this point the UNDO sequence will continue as described below. If the REDO function is available then the target data volume will be copied to a second PIT storage box so the REDO function can be later chosen. The UNDO process will continue at block 432 where the data replication relationships are removed then, at block 434, the first PIT storage box copy is copied to the target storage box reverting the data to the original form prior to beginning the replication process in block 402.
If the REDO function is available at block 436 then the REDO function can be carried out at block 438 if so chosen by the user. If the user chooses not to enable the REDO function the replication process will stop at block 414. If the REDO function is initiated then the 2nd PIT storage box is copied to the target storage box at 438. The data is then in the same form it was at the time the data replication was stopped in block 424 thereby allowing the replication process to continue from where it left off. If the user chooses to initiate the REDO function then data replication will continue at block 440 by carrying out block 402 again or stopping and waiting to choose the next process at 414.
It will be understood that all steps necessary to make the UNDO command available to the user of the system, method and computer readable medium with program instructions for replicating a volume of data will be implemented by the data replication management application. It will be further understood that all such steps will be executed by a single command to the data management application. The user will provide the management application with a pool of volumes for each target storage box that will be used to maintain the data needed to UNDO the data writes and return the target volumes to their original states before they were used as a data replication target.
In operation, when the start replication command is issued in blocks 102, 202 or 402 (
During operation of the system, if the user desires to UNDO the start replication command, the user will issue the UNDO command to the management application. The management application without, further action by the user, will carry out all of the steps necessary to complete the UNDO command and return the original data back to all of the target volumes in the session.
Specifically, as described above, the management application will carry out the following steps during operation: 1.) stop the data replication from continuing; 2.) remove the data replication relationships to enable data to be restored from the PIT copies to the target volumes; 3.) before restoring the data, the application will create another set of PIT copies from the current target volumes to provide implementation of a REDO command after the UNDO command is completed, if desired; 4.) if the allocated resources are not sufficient, a warning will be raised to the user that the REDO command will not be available for this session; and 5.) after the REDO PIT copies are made, the management application will copy the UNDO PIT copies back to the target volumes to revert all target volumes back to the data they contained before replication started. At this point, the user is able to access the target volumes to save data or carry out additional functions including another replication process. The next step would be to either initiate the REDO command and then continue replication as is, or reconfigure the replication session to use different target volumes thus leaving the target volumes as they were originally.
Each time the start command is issued, the management application will again save PIT copies to allow another UNDO command to restore the target volumes. In this manner, if sufficient resources exist, the application may save multiple UNDO points and the user can sequence back through the various UNDO states and return to any desired one. If sufficient resources do not exist to have multiple UNDO PIT copies, then the oldest PIT copies will be recycled when the next start command is issued.
Another embodiment of a method 500 for replicating a volume of data is illustrated at
At least one point in time volume is established at step 540, and the storage capacity of the point in time volume is determined at step 550. The point in time volume is a storage volume, such as a distinct physical drive, drive partition, or system of drives. The storage capacity of the point in time volume is a measure of the number of bits that can be stored on the point in time volume.
The size of the received data block is determined at step 555. In one embodiment, the size of the received data block is measured in the number of bits within the received data block. The received data block is stored on at least one of the storage volumes at step 560. In one embodiment, the received data block is then mirrored to at least one other storage volume.
A point in time copy is stored on the point in time volume at step 565. The point in time copy is a snapshot of the data stored on at least one of the storage volumes. The determined storage capacity of the point in time copy is then modified based on a size of the stored point in time copy at step 570. In one embodiment, the size of the stored point in time copy is subtracted from the storage capacity of the point in time volume.
At least one modified data block is received after storing at least one data block, at step 575. The modified data block is then compared to the stored data block on the at least one storage volume at step 580, and any differences between the stored data block and the modified data block at determined at step 585. If there are any differences, the size of the different portions of the modified data block are determined at step 590, and the size of the different portions of the modified data block is compared to the modified storage capacity of the point in time volume, at step 600.
Based on the comparison, a storage capacity warning is issued when the compared determined size of at least one difference is greater than or equal to the modified storage capacity of the established point in time volume at step 610. Alternatively, at least one difference and a pointer are stored on the point in time volume when the compared determined size of the difference is less than the modified storage capacity of the established point in time volume at step 615.
An undo command is received at the central location at step 620. Based on receiving the undo command, the central location accesses the point in time volume and locates the most recent pointer at step 630. Based on the located pointer, the central location replaces the data block stored on the storage volumes with the corresponding modified data blocks from the point in time volumes at step 630.
Another embodiment of a method for data replication includes setting up a mirrored data server. The mirrored data server includes at least one mirrored drive, its mirror and a point in time server. The point in time server operates to store a point in time copy of data on the mirrored drive. In one embodiment, the point in time server stores a series of modifications to the mirrored drive with a pointer to each modification. In another embodiment, the point in time server stores a series of complete copies of the mirrored server, although those of skill in the art will recognize the storage limitations occasioned by such embodiments. Having set up the mirrored system with a point in time copy, a central location receives an undo data replication command, and accesses the point in time server using a pointer associated with the undo command and restores the mirrored system based on the data associated with the pointer. In the event that data modifications are to be made to the mirrored server at a time when the point in time server has insufficient storage capacity to provide for the storage of the data modifications, a warning will be directed to the user indicating that the undo capacity will either be disabled due to storage concerns, or that the data storage on the point in time server will be altered according to an inventory control rule, such as first-in-first-out, or last-in-first-out, or other such rules.
Those of skill in the art will recognize that the disclosures provided herein provide for an automated integration of an undo capacity for replications of data in a mirrored storage system. To the extent that the point in time storage volumes have sufficient capacity to store the modifications to data blocks stored on the storage volumes, the pointers stored on the point in time volumes provide the mechanism to restore data replication at any point in the past. Thus, multiple ‘levels’ of undo capability are provided, based on the amount of storage capacity. Indeed, the only actual inputs by a user are the identification of storage volumes and point in time volumes, a data block input, and the undo command.
It is contemplated that available techniques such as space efficient flash copy will be used where applicable to reduce the number of volumes needed to achieve the UNDO functionality included in the present invention. Those skilled in the art will recognize that the techniques and embodiments described herein may be implemented with desirable results in all types of data base systems. It is to be understood that several of the steps disclosed in the flow charts of
While the present invention has been described above in terms of specific embodiments, it is to be understood that the invention is not intended to be confined or limited to the preferred embodiments disclosed herein and that the scope of the present invention fully encompasses other embodiments which may become obvious to those skilled in the art. In addition, while specific component values have been show for ease of illustration and description, it should be understood that a variety of combination of values is possible and contemplated by the present invention. Further, while specific connections have been used and shown for ease of description, it should also be understood that a variety of connection points are possible and may vary depending on the specifics of the application and circuit used. These and all other such modifications and changes are considered to be within the scope of the appended claims and that the scope of the present invention is accordingly to be limited by nothing other than the appended claims.
