Multi-time-point rollback complete machine disaster recovery system and method based on continuous data protection

Abstract

The invention discloses a multi-time-point rollback complete machine disaster recovery system and method based on continuous data protection, and relates to the technical field of computer data disaster recovery. The system comprises a source host, a target machine, a control center, a synchronous engine and a switching engine, and full-amount and incremental data capture, transmission and multi-version management are achieved through the synchronous engine; and the switching engine supports one-key switching to a specified time point, dynamically modifies the starting sequence of the target machine to load a source host system, and supports target machine resetting to realize multiple switching. The method covers target machine environment preparation, data synchronization, fault switching and multiple rollback processes, storage occupation is reduced through file backup and difference comparison, and the problems that traditional disaster recovery switching is low in efficiency, high in environment dependence and limited in single-time switching are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer data disaster recovery, and particularly to an entire machine disaster recovery system and method based on continuous data protection. Background Art

[0002] In traditional entire machine disaster recovery systems, there are some defects and drawbacks, mainly reflected in the following aspects: 1) Low switching efficiency: After a source host fails, many solutions require retransmitting the data of the target machine to a third machine and then starting the third machine to take over the services of the source host, resulting in an overly long RTO and poor actual application effects; 2) High environmental dependence: To solve the problem of overly long RTO, some solutions introduce a virtualization environment, where disaster recovery needs to be carried out, increasing the requirements for the disaster recovery environment and the disaster recovery cost; 3) Single-switching limitation: Another major problem with traditional disaster recovery is that it can only perform one switch and does not support rollback switching at multiple time points, resulting in insufficient reliability of disaster recovery. Summary of the Invention

[0003] To solve the above defects and drawbacks in the entire machine disaster recovery system, the present invention proposes a multi-time-point rollback entire machine disaster recovery system and method based on continuous data protection.

[0004] To achieve the purpose of the present invention, the present invention is realized through the following technical solutions.

[0005] A multi-time-point rollback entire machine disaster recovery system and method based on continuous data protection, the system architecture is as follows: Source host: Runs production services and provides services externally, including a system disk and data disks, etc. It can be a virtual machine or a physical machine, and is deployed with the proxy program used in the present invention; Target machine: Can be a virtual machine or a physical machine, used to store the data of the source host, deployed with the proxy program used in the present invention. In addition to its own operating system and the disk for data storage, it also needs to be configured with an additional disk to be consistent with the source host disk, for receiving the stock data and incremental data of the source host, and can be quickly switched; Control center: Provides external services, such as GUI operations, etc. It is the management end of the present invention. The control center controls various behaviors of the proxy by communicating with the proxy; Synchronization engine: Used for data synchronization, such as synchronizing the full amount of data and incremental data of the source host to the target machine, mainly realizing: 1) Transmission of the full amount of data of the source host; 2) Capture and transmission of incrementally changed data; 3) Management of multi-version stored data, such as rollback data and bookmarks, etc.; Switching engine: used for the whole machine switching. When the source host fails and automatic or manual switching is required, it can switch to a specified time point or the latest data, etc. The main functions are as follows: 1) One-key switching function, supporting automatic or manual switching to a specified time point or the latest data; 2) Start the boot module, dynamically modify the operating system boot sequence. After switching, start the system from the operating system disk corresponding to the source host; 3) Target machine reset. When it is necessary to switch to a different time point again, the target machine reset operation can be quickly performed to enter the state before switching and then perform the switching operation again.

[0006] A multi-time-point rollback whole machine disaster recovery system and method based on continuous data protection, including: Target machine environment preparation. If it is a virtualized environment, the target machine can be automatically created according to the disk information of the source host, create disks corresponding to the source host, add them to the target machine, and automatically deploy the proxy program to prepare for running the disaster recovery plan. If it is a non-virtualized environment, the user extracts and constructs the target machine according to the source host disk information; Run the disaster recovery plan, synchronize the source host disk data to the target machine, synchronize the data corresponding to the operating system to the disks corresponding to the target machine and the source host, and synchronize the data disks to the corresponding data disks; Optionally, to support multiple switches, on the target machine, save the data of the operating system volume and data volume of the source host to the corresponding files, which are called backup files here, and update these backup files in real time according to the incremental data. To save space, sparse files (Windows) or hole files (Linux) are used to save. During the operation of the plan, the backup files are always consistent with the data on the corresponding source host disks on the target machine; Capture the real-time incremental change data of the source host and synchronize it to the target machine and save it according to the time series; When the source host fails and switching is required, the control center notifies the target machine proxy program to perform the switching. The switching engine changes the boot loader according to the switching requirements and sets the target machine to boot from the operating system disk corresponding to the source host; Optionally, after the switching is successful, if the user is not satisfied with the data of the currently booted system, for example, the switched time point is too old or too new, the user can reselect a new switching time point to perform the switching again. Since it has been switched, the data on the disks corresponding to the target machine and the source host may have changed. At this time, before switching, compare the differences between the backup files and the changed disk data, overwrite the different parts of the data with the backup files, and then perform the switching process again; Description of the Drawings

[0007] Figure 1 : Correspondence between the disks of the source host and the target machine Detailed implementation manners

[0008] The following describes the implementation manners of the present invention through specific specific examples in combination with the accompanying drawings. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific examples, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0009] Example 1: Initial switch of the whole machine disaster recovery Step 1: Configure the target machine. The target machine can be a virtual machine or a physical machine, and maintain the same type of operating system as the source host. For example, if the operating system of the source host is Windows, then the target machine is also configured as Windows; if the source host is Linux, then the target machine is also Linux. If the target machine is a virtual machine, it can be automatically created through the control center in the present invention, and disks corresponding to the disks of the source host are newly configured. For example, if the source host has two disks, one for the operating system and one for data, then two disks are newly configured on the target machine, one for storing the disk data of the source host operating system and one for storing the disk data corresponding to the source host data disk. If the disaster recovery plan hopes to retain historical data for a longer time, a separate disk can be configured on the target machine to save backups and historical data, etc., as Figure 1 shown as the correspondence between the disks of the source host and the target machine; Step 2: Synchronize the data of the source host operating system and data disk to the corresponding disks of the target machine; Step 3: When a failure occurs in the source host and an automatic or manual switch is required, notify the target machine to perform a switch operation through the control center; Step 4: After the proxy program of the target machine receives the switch request, if it is switched to the latest data, directly modify the boot loader of the target machine, and set the target machine to boot from the operating system disk corresponding to the source host. If it is switched to a specified time point, according to the selected time point, use the saved CDP historical data to roll back the data on the disk corresponding to the source host on the target machine. After rolling back to after the specified time point, then modify the boot loader of the target machine and set the target machine to boot from the operating system disk corresponding to the source host; Step 5: Restart the target machine to complete the switch.

[0010] Example 2: Re-switch When a switch has been performed and it is necessary to switch to data at a different time point again, the steps are as follows: Step 1: Notify the target machine to perform a reset operation on the target machine; Since the switch has been performed, the target machine system at this time has already booted from the operating system disk corresponding to the source host. That is, the target machine at this time is the data of the source host at a certain time point. Therefore, it is necessary to perform a reset operation on the target machine to make the target machine return to the state before the switch. After receiving the operation of resetting the target machine, the target machine modifies the bootloader of the target machine through the proxy program, configures the target machine to boot from the original operating system disk of the target machine, and then restarts the target machine to enter the state before the switch; Step 2: The user re-selects the time point to be switched and notifies the target machine through the control center; Step 3: After receiving the command to switch again, the target machine proxy program performs the following processing according to the time point to be switched: 1) Obtain the backup file information from the historical data disk of the target machine, and perform a differential comparison between the disk data corresponding to the source host on the target machine and the backup file. According to the differential comparison result, overwrite the different data parts of the disk corresponding to the source host; 2) According to the time point to be switched, roll back the data to the specified time point, and then modify the bootloader of the target machine to set the target machine to boot from the operating system disk corresponding to the source host; Step 4: Restart the target machine to complete the switch.

[0011] In summary, a multi-time-point rollback whole-machine disaster recovery system and method based on continuous data protection of the present invention effectively solve the related problems such as high dependence on the traditional disaster recovery environment, low switching efficiency, and inability to perform multiple switches.

[0012] The above embodiments are only illustrative of the principles and effects of the present invention, and are not used to limit the present invention. Any person skilled in the art can modify and change the above embodiments without departing from the spirit and scope of the present invention. Therefore, the scope of the protection of the rights of the present invention shall be as listed in the claims.

Claims

1. An entire machine disaster recovery system and method based on continuous data protection, characterized in that, Including: Source host: Runs production services and deploys agent programs, including a system disk and a data disk; Target machine: Configures a storage space corresponding to the disks of the source host, used to receive full and incremental data, and supports switching to the source host system startup by starting the boot module; Control center: Provides a management interface to control data synchronization and switching operations through the agent program; Synchronization engine: Implements full data transfer of the source host, capture and transfer of incremental data, and management of multi-version stored data; Switching engine: Supports one-key switching to a specified time point or the latest data, dynamically modifies the startup order of the target machine, and supports resetting the target machine for multiple switches.

2. The system according to claim 1, wherein Except for the operating system disk and the backup data file disk, the configured disks of the target machine are additionally configured to be consistent with the source host.

3. The system according to claim 1, wherein The target machine uses sparse files or hole files to save the backup files of the source host operating system and data disk to save storage space.

4. The system according to claim 1, wherein When making multiple switches, the switching engine overwrites the disk data of the target machine through differential comparison and rolls back to the specified time point using the historical backup files.

Images