CN100405312C - Error Control Method in Remote Data Backup - Google Patents

Abstract

The invention provides an error control method in remote data backup, which belongs to the field of data backup and disaster recovery. In the present invention, through the cumulative verification method, each time the local data generates data that needs to be backed up, the local end uses the summary update algorithm to calculate the summary value reflecting the current local data status, and combines the summary value of the current local data status with the data that needs to be backed up Encapsulate it into a backup data unit and send it to the backup end. After the backup end receives the backup data unit, it also uses the digest update algorithm to calculate the backup end digest value that can reflect the data status of the backup end, and compares it with the local digest value. If the value is consistent, it means that the data on the backup end is consistent with the data on the local end, and the backup end can implement data backup; if the summary values at both ends are inconsistent, it means that the data on the backup end is inconsistent with the data on the local end, and error handling is required. By using the cumulative verification method, the consistency of the data at both ends is detected in real time during the backup process to ensure that when the data at both ends is inconsistent, it can be found in time to ensure the validity and correctness of the backup data. The invention has high efficiency and good real-time performance, can be widely used in data backup, data mirroring, data disaster recovery and other fields, and has broad application prospects.

Description

Error Control Method in Remote Data Backup

technical field

The invention proposes an error control method in remote data backup, belongs to the field of data backup and disaster recovery, and is used for real-time detection and control of the consistency between backup data and local data in a backup system.

Background technique

In the backup system, ensuring the correctness of backup data is the most basic requirement of backup technology, so effective error control must be carried out on the backup data during the backup process to ensure the correctness and effectiveness of data backup. In order to check the correctness of the backup data, the traditional method is to use the method of post-event comparison, but this method cannot find the occurrence of errors in the process of data backup in time.

The invention proposes an efficient data error control method, which can be used for real-time detection of data correctness in the remote backup process, and can always ensure the validity of the backup data. In the present invention, the efficient error control of data is realized through cumulative verification and confirmation of summaries of local data and backup data during the backup process. The invention can improve the efficiency of data error control in the backup process and ensure the correctness of the backup data. It can be widely used in data backup, data mirroring, data disaster recovery and other related fields.

Contents of the invention

The invention is mainly used to ensure the consistency between the remote backup data and the local data in the remote data backup process. The present invention adopts the method of accumulative verification. When each local data generates data that needs to be backed up, the local end uses a digest update algorithm to calculate a digest value reflecting the current local data state, and combines the digest value of the current local data state together with the data that needs to be backed up. The data is encapsulated into a backup data unit and sent to the remote end. After receiving the backup data unit, the remote end also uses the digest update algorithm to calculate the remote digest value that can reflect the state of the remote data, and compares it with the local digest value. If both If the end summary values are consistent, it means that the remote data is consistent with the local data, and the backup end can implement data backup; if the summary values at both ends are inconsistent, it means that the remote end data is inconsistent with the local end data, and error handling is required. By using the cumulative verification method, the consistency of the data at both ends is detected in real time during the backup process to ensure that when the data at both ends is inconsistent, it can be found in time to ensure the validity and correctness of the backup data.

Description of drawings

Fig. 1 is the operating principle of the present invention.

Figure 2 is the steps of local initialization.

Figure 3 is the steps of backup terminal initialization.

Figure 4 shows the steps of data transmission at the local end.

Fig. 5 is the steps of error control at the backup end.

Detailed ways

In order to further illustrate the principles and features of the present invention, the following will be described in conjunction with the accompanying drawings.

Before explaining, define the nouns and symbols that will be used:

1) H: Digest algorithm, such as MD5, SHA-1, etc.

2) D _L : The current data status of the local end.

D _R : The current data status of the backup terminal.

3) D: The data change of the local end.

E _D : The summary value of the local change data.

4) E _L : the summary value corresponding to the current state of the local terminal.

E _Lnew : The summary value corresponding to the data change on the local side.

5) E _R : the summary value corresponding to the current state of the backup end.

E _Rnew : The summary value corresponding to the data change on the backup side.

6) R: backup data unit.

Fig. 1 is the operating principle of the present invention.

The present invention mainly realizes the error control function in the remote data backup (mirror image) process. The local monitoring module FMON in Fig. 1 monitors the data changes of the files to be mirrored in the local data center, and encapsulates the current data changes, And passed to the local mirror module LMM. The local mirroring module LMM processes the data change and packages it with other information such as a summary of the data change into a backup unit and sends it to the remote end. The mirror module at the far end receives the backup unit, first implements error control and completes the backup of local data through remote reconstruction at the far end. In the process of data replay, error control technology is used to ensure the consistency of data. Remote playback and error control technologies are completed by the local mirroring module LMM and the remote mirroring module RMM.

Before implementing this method, it is necessary to ensure the consistency of the data on the backup end and the data on the local end. Therefore, the corresponding initialization work needs to be completed before the backup starts. In the initialization phase, LMM and RMM make the data at both ends in a consistent state through the synchronization and consistency detection of the data at the backup end and the data at the local end. When the data at the backup end is in a consistent state with the data at the local end, the error control method can detect in real time whether there is an error in the backup process, so as to ensure the validity of the backup data.

During the backup process, FMON monitors the data changes of the files that need to be mirrored in the local data center, and obtains the current data changes D (mainly write operations, including URL address, file name, file path, offset, write data, etc.), passed to the LMM. LMM calculates the digest value _ED of the data change situation D, and uses the digest update algorithm to calculate the digest value E _Lnew reflecting the current local data state. The LMM encapsulates D, E _D and E _Lnew into a backup unit R and sends it to the remote end. At the remote end, the RMM receives the backup unit R, parses out D, E _D , E _Lnew , and recalculates the relevant abstract values: E _D ′, E _Rnew . If E _D ′=E _D , it indicates that the current data transmission is complete; otherwise, it indicates that the local data change D has been destroyed during transmission and needs to be retransmitted; if E _Rnew =E _Lnew , the corresponding file maintains data consistency, Otherwise, it indicates that the current state of the data on the local end and the data on the backup end are inconsistent, and the data on both ends needs to be re-synchronized.

When the backup is stopped, before the local end exits, it needs to save the summary E _L corresponding to the current local data in the configuration file, and when the backup end exits, it also needs to save the summary E _R corresponding to the current backup data in the configuration file to ensure When the system is back up and running, the system can run smoothly again.

Figure 2 is the steps of local initialization.

When the local end of the present invention is initialized, the generation of the initial summary value of the local end is completed, and the necessary data synchronization work is completed in cooperation with the backup end, so as to ensure that the data at the local end and the backup end are in a consistent state before the backup starts. The local initialization process includes the following steps:

① Steps to obtain the summary value of the local terminal: read the summary value E _L corresponding to the local terminal from the configuration file corresponding to the local terminal:

I. If reading E _L fails, use digest algorithm H to perform digest operation on D _L to obtain its corresponding digest value E _L , namely: E _L =H(D _L ).

II. If the E _L is read successfully, go to ② to execute.

②The step of sending the summary value of the local terminal: Send the summary value E _L of the local terminal to the backup terminal, and determine whether the data of the local terminal and the backup terminal are consistent:

I. If the data on the local end is consistent with that on the backup end, go to ④.

II. If the data on the local end and the backup end are inconsistent, go to ③ to implement data synchronization.

③ Steps to synchronize the data of the backup end and the local end: Synchronize the data at both ends. After the synchronization is completed, go to ②.

④The step of saving the summary value of the local terminal: saving the summary E _L of the local terminal to the configuration file corresponding to the local terminal.

So far, the initialization of the local side is completed, and the backup of local data can be started.

Figure 3 is the steps of backup terminal initialization.

When the backup end is initialized, it completes the generation of the summary of the backup end, and cooperates with the local end to complete the data synchronization work, so that the data at the local and backup ends are in a consistent state before the data backup starts. Backup side initialization includes the following steps:

① Step of receiving the digest value of the local terminal: the backup terminal receives the digest E _L of the local terminal.

② Steps to detect whether the data on the local side is consistent with the data on the backup side, including:

1) The step of obtaining the summary value of the backup terminal: read the summary value E _R corresponding to the data of the current backup terminal from the corresponding configuration file:

I. If reading E _R fails, use digest algorithm H to perform digest operation on _DR to obtain its corresponding digest value E _R , namely: E _R =H(D _R ).

II. If reading E _R is successful, go to 2).

2) The step of comparing the summary value of the local terminal and the backup terminal: compare the summary value E _R of the backup terminal and the summary value E _L of the local terminal, and send the comparison result to the local terminal. The specific steps are as follows:

I. If E _R is consistent with E _L , go to ④.

II. If E _R is inconsistent with E _L , go to ③.

③ Steps to synchronize the data of the backup end and the local end: cooperate with the local end to synchronize the data between the local end and the backup end, and go to ① after the synchronization is completed.

④ The step of saving the summary value of the backup terminal: saving the summary value E _R of the backup terminal to the configuration file corresponding to the backup terminal.

Figure 4 shows the steps of data transmission at the local end.

During the backup process, the local LMM obtains the change of local data from the local monitoring module FMON, uses the summary update algorithm to calculate the relevant data summary value required for error control, and encapsulates the local data change and the calculated summary value Sent to the backup end for the backup unit.

The steps of local data transmission are as follows:

① Steps to obtain the data change at the local end: when the data of the file to be mirrored at the local end changes, obtain the change D of the local data from the local monitoring module.

②The step of calculating the summary value of the data change at the local end: use the summary algorithm H to obtain the summary value E _D corresponding to the data change D at the local end: E _D =H(D).

③ Steps to calculate the summary value of the local terminal:

(a) The step of obtaining the summary value of the local terminal: read the summary value E _L of the local terminal from the configuration file corresponding to the local terminal.

(b) Steps for calculating the new digest value at the local end:

I encapsulates the summary value E _D corresponding to the summary value E _L of the local terminal and the data change of the local terminal;

II uses the digest algorithm to perform digest operation on the encapsulated data to obtain a new digest value E _Lnew at the local end: E _Lnew =H(E _L +E _D ).

(c) The step of updating the digest value at the local end: update the digest value E _L at the local end: E _L =E _Lnew , and save E _L into the corresponding configuration file.

④The step of sending the backup data unit to the backup end: Encapsulating D, E _D and E _Lnew into a backup data unit R: (D+E _D +E _Lnew ), and sending the backup data unit R to the backup end.

⑤ Steps to accept the error control command from the backup end: If the backup success sign is received, go to ⑥: If the backup end retransmission command is received, go to ④; if the backup end error sign is received, report an error and perform error handling .

⑥Continue to process the next step of local end data transmission: Go to ①, continue to process the next step of local end data transmission.

Fig. 5 is the steps of error control at the backup end.

During the backup process, after the remote mirroring module RMM receives the backup unit sent by the local end, it uses the digest update algorithm to implement the corresponding digest operation, and uses the relevant digest information in the backup unit to detect whether there is an error in the data during the transmission process, and determine whether the data at both ends is wrong. Are the data consistent. If there is an error in the data transmission process, the local end is required to retransmit; if the data at both ends is inconsistent, the data will be synchronized; if no error occurs, the remote end will complete the data retransmission on the backup end Offsite refactoring.

Specifically, the steps of backup terminal error control include the following steps:

① Steps of receiving data from the local end: receiving the backup data unit R sent from the local end. Analyze R to get the data change condition D of the local end, the digest value E D corresponding to _D , and the new digest value E _Lnew of the local end.

② Steps to detect whether there is an error in the data transmission process: recalculate the summary E _D ′ of D: E _D ′=H(D); compare E _D ′ and E _D , and determine whether the data transmission process has an error. If E _D ′≠ _ED , it means that there is an error in the data transmission process, and a retransmission instruction is issued to the local end, requiring the local end to retransmit the backup data unit R.

③ Steps to check whether the data on the backup end is consistent with the data on the local end:

(a) The step of obtaining the summary value of the backup terminal: read the current summary value E _R of the backup terminal from the configuration file corresponding to the backup terminal.

(b) Steps for calculating the new summary value of the backup terminal:

I encapsulates the summary value E _R of the backup end and the summary value E _D ' of local data changes;

II uses the digest algorithm to perform digest operation on the encapsulated data, and obtains the new digest value E _Rnew of the backup end. That is: E _Rnew =H(E _R +E _D ').

(c) The step of comparing the new abstract value of the local terminal and the backup terminal: comparing the values of E _Rnew and E _Lnew , sending the comparison result to the local terminal, the specific steps are as follows:

I. If E _Rnew is consistent with E _Lnew , it means that the backup end data is consistent with the local data, go to ④.

II. If E _Rnew is inconsistent with E _L , it means that the backup end data and the local data are in an inconsistent state, and it is necessary to implement data synchronization between the backup end and the local end, report an error to the local end, and perform error handling.

④ Steps to implement data backup: reconstruct the remote data change D′ remotely according to the local data change situation D, and apply D′ to the backup end.

⑤ The step of updating the summary value of the backup terminal: update the summary value E _R of the remote backup terminal: E _R =E _Rnew , and save E _R into the configuration file of the backup terminal.

⑥The step of notifying the local end of the successful backup: sending the backup success sign to the local end.

⑦Continue to process the next step of backup work: go to ① to implement the next step of backup.

Claims (6)

1. the error control method in the long-distance data backup is characterized in that may further comprise the steps: the initialized step of local side;

The initialized step of backup end; The step of local side data transmission; The step of backup end Error Control;

Its feature also is: when specifically implementing, at first carry out initial work, comprise initialized step of local side and the initialized step of backup end, then in actual backup procedure, carried out the step of local side data transmission earlier by local side, the step of backup end Error Control is carried out in the back by the backup end;

Wherein

(1) the initialized step of local side may further comprise the steps:

1. obtain the step of local side digest value;

2. send the step of local side digest value;

3. the step of backed up in synchronization end and local side data;

4. preserve the step of local side digest value;

(2) the initialized step of backup end may further comprise the steps:

1. receive the step of local side digest value;

2. detect local side data and the whether consistent step of backup end data;

3. the step of backed up in synchronization end and local side data;

4. preserve the step of backup end digest value;

(3) step of local side data transmission may further comprise the steps:

1. obtain the step of local side data situation;

2. calculate the step of the digest value of local side data situation;

3. calculate the step of local side digest value;

4. send the step of Backup Data unit to the backup end;

5. accept backup end Error Control instruction step;

6. continue to handle the step of next step local side data transmission;

(4) step of backup end Error Control may further comprise the steps:

1. receive the step of local side data;

2. detect the step whether data transmission procedure makes mistakes;

3. detect the whether consistent step of backup end data and local side data;

4. the step of implementation data backup;

5. upgrade the step of backup end digest value;

6. notify the local side backup successful step;

7. continue to handle the step of next step back-up job.

2. whether consistent the error control method in the described a kind of long-distance data backup of claim 1 is characterized in that in the initialized step of backup end, detect the local side data step with backing up end data and may further comprise the steps:

(1) obtains the step that digest value is held in backup;

(2) compare the step that digest value is held in local side and backup.

3. the error control method in the described a kind of long-distance data backup of claim 1 is characterized in that in the step of local side data transmission, and the step of calculating the local side digest value may further comprise the steps:

(1) step of acquisition local side digest value;

(2) step of the new digest value of calculating local side;

(3) step of renewal local side digest value.

4. the error control method in the described a kind of long-distance data backup of claim 3 is characterized in that in the step of local side data transmission, calculates in the step of local side digest value, and the step of calculating the new digest value of local side may further comprise the steps:

(1) step of the digest value that encapsulation local side digest value is corresponding with the local side data situation;

(2) the data computing of making a summary after utilizing digest algorithm to encapsulation obtains the step of the new digest value of local side.

5. the error control method in the described a kind of long-distance data backup of claim 1 is characterized in that backing up in the step of end Error Control, and whether consistent step may further comprise the steps to detect backup end data and local side data:

(1) obtains the step that digest value is held in backup;

(2) calculate the step that new digest value is held in backup;

(3) compare local side and the step that backs up the new digest value of end.

6. the error control method in the described a kind of long-distance data backup of claim 5, it is characterized in that backing up in the step of end Error Control, detect in the whether consistent step of backup end data and local side data, calculate to back up and hold the step of new digest value may further comprise the steps:

(1) step of the digest value of encapsulation backup end digest value and local data situation of change;

(2) the data computing of making a summary after utilizing digest algorithm to encapsulation obtains backing up the step of the new digest value of end.

Images