KR100307925B1 - Data Back-up Method of the Switching System - Google Patents

Abstract

The present invention relates to a method for backing up data at an exchange, and more particularly, to a method for backing up data at an exchange for processing a PLD (Processor Loading Data) dump and a backup at the exchange.

The method of the present invention comprises the steps of performing a DBMS disk backup hold in response to a request of the DBMS disk backup hold when an operator requests a backup; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; Releasing the DBMS disk backup hold after copying the original PLD file to the disk under another name; And backing up the PLD file copied to the disk to a PLD file of an original name. Alternatively, another method of the present invention includes the step of performing a DBMS disk backup hold in response to a request of the DBMS disk backup hold when an operator requests a backup; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; De-duplicating the disk and separating the disk into an active disk and a backup disk; And releasing the DBMS disk backup hold to back up the PLD file from the backup disk.

Description

Data back-up method of the switching system

The present invention relates to a method for backing up data at an exchange, and more particularly, to a method for backing up data at an exchange for processing a PLD (Processor Loading Data) dump and a backup at the exchange.

In general, most exchanges have a remote system (MTU) and TCP / IP (Transmission Control Protocol / Internet Protocol) for the application and preservation of the PLD, which is the exchange operation data. System), which is a function of copying operation data to an exchange disk, and a backup function of copying operation data existing on the exchange disk to the outside of the exchange for preservation and other purposes.

During the operation of the exchange, the change of the operation data of the exchange occurs frequently or at the request of the operator. The changed operation data is changed and stored in the PLD on the disk through the DBMS (Database Management System).

At the same time, the dump of the exchange causes the double access of the disk PLD, that is, the PLD access for the dump and the database updater access of the DBMS due to the change of the exchange operation data. Therefore, the disk backup of the DBMS is stopped for safety.

In addition, the backup of the exchange stops the disk backup of the DBMS for dual access and consistency and integrity of the backed up PLD. In this case, the PLD exists in the form of a file, and changes continuously occur during the operation of the exchange, and when the backup is performed, changes in the operational data that have not been applied to the front of the PLD file to be backed up first have passed. Problems occur due to the time difference being backed up, such as applied later in the PLD file backed up.

Then, a data backup method in the conventional exchange will be described as follows with reference to the flowchart of FIG. 1.

First, when an operator requests a backup of the exchange, the user receives a backup request (step T1) and requests the DBMS to prohibit disk backup for dual access and consistency and integrity of the backed up PLD (step T2). Check whether the disk backup of the DBMS is prohibited (step T3).

Therefore, when the disk backup of the DBMS is prohibited in the third step (T3), the exchange is operated by a plurality of processors, and the common database that each processor has in common has the inefficiency of accessing a plurality of disks at the time of change. In order to reduce the number of changes, only one processor PLD is stored, and when performing the backup, a duplication related synchronization (Relation Synchronization) is performed to generate the same change in the remaining processor PLD (step T4).

When the duplication related synchronization is completed normally (step T5), if the duplication related synchronization is normally completed in the fifth step (T5), a backup of the PLD file is performed (step T6). The disk backup is prohibited (step T7).

On the other hand, if the backup of the DBMS is not prohibited in the third step T3 or if the duplication related synchronization is not normally completed in the fifth step T5, the backup operation of the exchange is stopped (step T8). .

As described above, a dump and backup of the operation data is performed by using the file transfer function of the MTU and TCP / IP for the purpose of applying and storing the exchange software package during the operation of the exchange. In addition, there are DBMS-related data required for the exchange operation in the PLD during the operation. The DBMS disk backup prohibition function, which is executed for the integrity and safety of dump and backup, causes problems in the operation of the exchange.

As such, the conventional scheme guarantees the stability of the PLD existing on the disk and the consistency and integrity of the PLD data at the time of dumping, but the failure to be changed automatically due to the suspension of the disk backup of the DBMS for changing the exchange operation data. There was a problem that the change of billing-related data could not be accepted, resulting in an error in the operation of the exchange.

The present invention relates to a switch management management method for PLD backup of the exchange to solve the above problems, and to handle the separation of the dump and backup of the PLD, for integrity and safety during dump and backup The purpose of this paper is to solve the problem of exchange operation caused by DBMS's disk backup prohibition function.

1 is a flowchart illustrating a data back-up method in a conventional exchange.

2 is a flowchart illustrating a data backup method in an exchange according to an embodiment of the present invention.

3 is a flowchart illustrating a data backup method in an exchange according to another embodiment of the present invention.

A data backup method in an exchange according to an embodiment of the present invention for achieving the above object comprises the steps of performing a DBMS disk backup hold in response to a request of the DBMS disk backup hold at the backup request of the operator; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; Releasing the DBMS disk backup hold after copying the original PLD file to the disk under another name; And backing up the PLD file copied to the disk to a PLD file of an original name. The DBMS disk backup hold may include applying a log accumulated in a log buffer to a disk; And storing the changed PLD of the DBMS applied to the log buffer after applying the log to the log buffer.

Alternatively, the data backup method in the switch according to another embodiment of the present invention for achieving the above object comprises the steps of performing a DBMS disk backup hold in response to a request of the DBMS disk backup hold in the backup request of the operator; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; De-duplicating the disk and separating the disk into an active disk and a backup disk; And releasing the DBMS disk backup hold to back up the PLD file from the backup disk. When the backup of the PLD file is terminated, the method further includes performing a mirror copy to operate the two disks in redundancy again. In addition, the simultaneous recording function is released from the time point at which the activation disk and the backup disk are separated, and the activation disk is used for switching operation and the backup disk is used for backup.

The present invention does not use the DBMS disk backup prohibition at the time of backup of the exchange operation data and reflects the current PLD change to the disk. To perform PLD backup through DBMS disk backup hold function applied sequentially to the DBMS disk backup, and to copy DBMS disk backup after copying the original PLD file to another file name to reduce the disk backup downtime of the DBMS, PLD backup can be performed through the function of backing up the PLD file to its original name. Alternatively, the active side can be un-redundant to separate the redundant disk to fully accommodate the change of the exchanger operation data during the PLD backup. Is used for switchboard operation and the Stand-by Side is used for backup. At the end of the backup, PLD backup can be performed by maintaining mirror redundancy through mirror copy.

In other words, the present invention allows for separate handling at the dump and backup of the exchange PLD.

First, when the PLD is dumped, the PLD, which is the operation data of the exchange, is newly dumped to the disk when a switch is to be operated continuously and a problem occurs after stopping and re-operating with a new PLD or applying a new software package. Will be performed.

Therefore, during the dump, the PLD database backup change of the disk is prohibited as before, and the database exchange of the disk PLD is allowed from the end of the dump so that a new exchanger can be operated.

Secondly, when looking at the backup of the PLD, the backup of the PLD should be performed without affecting the switch operating continuously without interruption, so the DBMS disk backup stop of the PLD is basically not performed. At this time, in order to ensure the consistency of the PLD data in question, the DBMS instead of stopping the DBMS disk backup of the PLD, the log buffer, which is the database memory of the corresponding processor, to accommodate the change of the operational data. Perform a disk back-up hold function that keeps your changes on the fly.

After the PLD backup is ended, the disk backup hold function is sequentially applied to the disk PLD, and is automatically performed when the disk backup hold function is terminated.

However, due to the limited processor memory used in the DBMS, the method can only accommodate up to a certain amount of operational data changes, and thus, when a large amount of operational data changes occur, changes over the acceptable limit may be lost as in the conventional problem.

Therefore, there are two ways to avoid the possibility as described above, but the first method can not completely prevent the existing problem, but can be easily implemented and operated, so that there is little change of operation data. Applicable to the system, the second method is somewhat more complicated than the first method, but it can solve the existing problem completely and is suitable for the system where a large amount of change of operation data of the exchange occurs.

As a first method, a data backup method in an exchange according to an embodiment of the present invention will be described with reference to the flowchart of FIG. 2 as follows.

First, when the operator requests the backup of the exchange, the manager receives the backup request of the operator (step S1) and requests the disk backup hold of the DBMS to perform the backup (step S2).

In response, the DBMS receiving the disk backup hold request applies the log accumulated in the log buffer to the disk, and performs a disk backup hold function in which incoming DBMS changes continue to be accumulated only in the log buffer (step S3). ). In this case, when a disk backup hold of the DBMS occurs, a disk copy that takes a relatively short time is used to shorten a backup time that takes a relatively long time.

In addition, it is checked whether the disk backup hold function of the DBMS is normally performed (step S4). If the disk backup hold function is normally performed in the fourth step (S4), the change is generated in the same manner in another processor PLD. The master performs the duplication related motive (step S5).

In addition, it is checked whether the duplication related synchronization is completed normally (step S6). If the duplication related synchronization is completed normally in the sixth step S6, the application processor in charge of the backup has another name for the PLD to be backed up. To the disc (step S7).

Accordingly, when copying of the PLD is terminated in the seventh step S7, the disk backup hold of the DBMS is released for normal operation of the exchange (step S8). Here, the logs accumulated in the log buffer are sequentially stored in the disk while the disk backup hold of the DBMS is performed.

Then, another PLD (that is, a PLD having a different name but the same content) copied to the disk is backed up to a normal PLD file name, that is, a PLD file of the original name (step S9).

On the other hand, if the disk backup hold function is not normally performed in the fourth step S4 or the duplication related synchronization is not normally completed in the sixth step S6, the backup operation of the exchange is stopped (step S10). .

As described above, the above-described method is more reliable than the conventional method because copying of a disk file can save a considerable time in time than backup to an existing MT (Magnetic Tape).

As a second method, a data backup method in an exchange according to another embodiment of the present invention will be described with reference to the flowchart of FIG. 3.

First, when an operator requests a backup of the exchange, the manager receives a backup request of the operator (step S11) and requests a disk backup hold of the DBMS (step S12). Thus, the disk backup hold function is performed in the same manner as the first method. (Step S13).

When the disk backup hold function of the DBMS is normally performed (step S14), if the disk backup hold function is normally performed in step 14 (S14), duplication related synchronization is performed (step S15).

In addition, it is checked whether the duplication related synchronization is completed normally (step S16). If the duplication related synchronization is completed normally in the sixteenth step (S16), the duplication of the duplicated disk is immediately stopped to activate the active disk (Active Disk). And logical separation into the backup disk (step S17). Then, the DBMS disk backup hold is released for normal operation of the exchange (step S18).

At this time, the Concurrent Writing function is released from the time point at which the activation disk and the backup disk are separated. In this case, only the activation of the disk, which is operated by redundancy and performed simultaneously, is performed. That is, the disk of the system is operated only on the activation side, not redundancy, and the change of operation data is applied only on the activation side. Therefore, the backup of the PLD will back up the PLD from the disk on the standby side that has been switched in redundancy (step S19).

Accordingly, when the backup of the PLD is terminated in the nineteenth step (S19), the disk separation operation is released to operate the disk in redundancy again (step S20), for the consistency of the active side and the standby side of the disk. All backup operations are terminated while performing mirror copying of both disks (step S21).

On the other hand, if the disk backup hold function is not normally performed in the fourteenth step S14 or if the duplication related synchronization is not normally completed in the sixteenth step S16, the backup operation of the exchange is stopped (step S22). .

As such, the above-described method can completely solve the existing problem.

Alternatively, if you look briefly at the processing operation for changing the exchange operation data of the DBMS, the memory of the processor is changed, and the PLD change storage of the disk is stored in the log buffer of the memory at the producer. ) Will be applied to the disk in order. At this time, the disk backup hold function of the DBMS suspends the role of the consumer.

As described above, even when the PLD backup of the exchange frequently occurs according to the present invention, the change of the exchange operation data is accommodated without interruption, and thus, it may be helpful for stable exchange operation.

Claims (5)

Performing a DBMS disk backup hold according to a request of the DBMS disk backup hold when an operator requests a backup; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; Releasing the DBMS disk backup hold after copying the original PLD file to the disk under another name; And backing up the PLD file copied to the disk to a PLD file of an original name. The method of claim 1, The process of performing DBMS disk backup hold may include applying logs, which have been accumulated in a log buffer, to disk; And storing the changed PLD of the DBMS applied to the log buffer after applying the log to the log buffer. Performing a DBMS disk backup hold according to a request of the DBMS disk backup hold when an operator requests a backup; When the DBMS disk backup hold is completed, performing duplication related synchronization and checking whether the duplication related synchronization is completed; De-duplicating the disk and separating the disk into an active disk and a backup disk; And releasing the DBMS disk backup hold to back up the PLD file from the backup disk. The method of claim 3, And performing a mirror copy when the backup of the PLD file is completed, and operating the two disks by redundancy again. The method of claim 3, And releasing the simultaneous recording function from the time point at which the activation disk and the backup disk are separated, using the activation disk for switching operation, and using the backup disk for backup.