Classifications

• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F11/00—Error detection; Error correction; Monitoring
  • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
  • G06F11/14—Error detection or correction of the data by redundancy in operation
  • G06F11/1402—Saving, restoring, recovering or retrying
  • G06F11/1446—Point-in-time backing up or restoration of persistent data
  • G06F11/1458—Management of the backup or restore process
  • G06F11/1466—Management of the backup or restore process to make the backup process non-disruptive

Definitions

• This invention relates to a method of operating a computer system, and in particular to a method of providing backup copies of data stored in a storage device to guard against the possibility of the storage device becoming faulty or the data becoming corrupted, lost or, more recently, "infected by a computer virus" .
• the invention is especially concerned with the provision of backups for personal computers (P.C.s) .
• file locking and “record locking” have been introduced into operating systems to avoid, in the case of file locking, more than one application writing to a file or, in the case of record locking, two applications modifying the same area of a given file. These methods are sophisticated enough for most applications, particularly if each application only makes relatively minor changes to files which are shared and therefore does not significantly delay other applications wishing to access a file or record that is locked.
• an application for periodically backing up data in a computer system where the storage means are shared is a special case.
• the smallest unit of data that can be backed up is a single file, since neither the backup application nor the operating system has knowledge of the internal structure of a file because it could be associated with any application program.
• the backup copy of the file needs to be an exact image of the original at a particular point in time. Copying a file is, however, not an instantaneous process and the time taken to copy will depend on the size of the file and the speed of the copying operation. This leads to a problem unless the backup application can be sure that no other application can modify the file during the copying process.
• a backup application allocates an area of random access memory (RAM) as a transfer buffer and proceeds by alternately reading blocks of the source file into the buffer and then writing the buffer contents out to a secondary storage means. Unless the file is very small, the transfer buffer will fill up several times and the copying operation will require several read and write operations to make a complete copy of the file.
• RAM random access memory
• the potential problem caused by allowing the file to be modified during the copying process may be illustrated by considering the example of a file which is being copied as a series of 10 blocks. If the file were to be modified by a transaction that required changes to the file which corresponded to blocks 3 and 8 after blocks 1 to 5 had already been copied, the copying process would subsequently copy the change to block 8 but would not copy the change to block 3 since this block had already been copied before the transaction occurred. The backup copy will therefore contain only a partial transaction which may make the backup copy entirely useless, since the application that "owns" it will probably consider the file to be corrupt.
• US Patent No. 5,086,502 discloses a method of backing up each write operation as it is made by making a duplicate write operation to a secondary storage means.
• this method there are two other ways of backing up files on a computer system having shared storage means:-
• Detect modifications to the file After copying the file, the backup application can use the operating system to check if any modifications were made during the copy operation. If modifications were made, then the copy must be repeated or that file skipped.
• a method of operating a microprocessor-based computer system including a random access memory, a central processing unit and storage means, comprises the steps of providing a backup process running on the computer system and which is operable to issue a start signal and a stop signal and to perform at least one backup read operation to read original data from the storage means, providing first instructions stored in said memory to said central processing unit to cause said central processing unit to perform at least one normal write operation to write data periodically to a primary area of said storage means, and at least one normal read operation to read data from said storage means, after receiving said start signal, performing an interception operation to intercept said write operation, and to write data to a secondary storage area of the storage means so as to preserve the original data which would otherwise be overwritten by said write operation, maintaining a record of said writing to the secondary area and, on performing one of said backup and normal read operations, interrogating said record to read data from one of the primary and the secondary storage areas according to the record, whereby the backup process is provided with preserved original data in the state it was
• the delta area is preferably a different area of the same storage device which may for instance be a hard disk.
• the additional operations for performing the invention are advantageously stored and operated at an operating system interface level of the computer system.
• the software for performing the method of the invention is stored at the operating system level, it can be made transparent to the user so as not to affect application software.
• this may comprise a record of the first data, a record of its intended location on the storage device and a record of its actual location, and with the method further comprises the step of obtaining the second data from the intended location.
• start command is issued prior to the backup application program commencing a backup and preferably the stop command is issued after the backup has been completed.
• a method for providing data to a backup process executing on a microprocessor-based computer system comprises the steps of providing first instructions stored in said memory to said central processing unit to cause said central processing unit to perform at least one normal write operation to write data periodically to a primary area of said storage means, and at least one normal read operation to read data from said storage means, receiving a start signal from the backup process, after receiving said start signal, performing a an interception operation to intercept said write operation, and to write data to a secondary storage area of the storage means so as to preserve original data which would otherwise be overwritten by said write operation, maintaining a record of said writing to the secondary area and, on performing one of said backup and said normal read operations, interrogating said record to read data from one of
• Figure 1 is a block diagram of a personal computer and a cartridge tape drive
• Figures 2A and 2B are diagrams illustrating a system memory map at the operating system level of a conventional personal computer and a computer modified to operate in accordance with the method of the invention respectively;
• Figure 3 is a flow chart of a method for recording data in a backup operation.
• Figure 4 is a flow chart of a second embodiment of the method of Figure 3.
• a personal computer 10 typically comprises a central processing unit 12, a random access memory 14, and a non-volatile storage device in the form of a hard disk drive 16.
• a tape unit 18 is coupled to the computer 10.
• part of a backup program application is formed as a recording process which operates to intercept normal read and write operations from and to the disk drive 16 respectively, in order to control the flow of data between the disk drive 16 and the other parts of the computer 10.
• the recording process is activated by the backup application to re-direct all disk write activity to another area of the disk 16 known as a "delta" area.
• a delta mapping table is maintained of all disk write operations that have been re ⁇ directed to the delta area. This table maintains a mapping between the intended destination on the disk 16, of the data associated with each disk write operation and of its actual storage position in the delta area. If a subsequent write operation occurs to an area that is already the subject of an entry in the table, the old data in the delta area is allowed to be overwritten by the new data. In this way, the data on the main part of the disk does not change whilst a backup copy is being made, since all changes are re-directed to the delta area.
• the recording process When a disk read operation is requested by the backup application, it is passed through unchanged so that the data returned to the backup application comes from the original data area on the main part of the disk.
• the recording process first operates to check the table to see if a disk write to that area has occurred since the process started intercepting write operations and, if so, the read operation is re-directed to the corresponding position in the delta area. If no write operation has occurred i.e. that piece of data has not been altered since the write operation interception started, then the read operation is passed through unchanged.
• the recording process is activated to return the computer to normal operation.
• This restoration phase is achieved by applying the changes stored in the delta area to the appropriate original data areas in the main part of the disk 16 using the mappings stored in the table. While the delta changes are being applied, disk write operations continue to be intercepted as described above. When there are no more write operations that have not been applied to the original data area, the interception of write operations is halted and the system operates conventionally.
• the software can allow write operations to the main part of the disk 16 to pass through unchanged after that part of the disk has been backed up since it will no longer matter if the data changes. This reduces the size of the delta mapping table. Also if the area on the main part of the disk to which that operation is directed was already the subject of an entry in the mapping table, that entry may simply be deleted thereby reducing the time taken in the restoration phase.
• the invention is particularly applicable to IBM compatible personal computers, i.e. the majority of personal computers using processor chip type numbers 8086, 8088, 80286, 80286,
• Interrupt 21 hex (INT21h) is an integral part of the operating system.
• the action performed by the routine depends on the parameters passed to the routine upon entry.
• This routine is shown in Figure 2A as INT21h forming part of the operating system in a system memory map, the INT21h entry point being shown by an arrow.
• additional program code is added at the operating system interface level as shown in Figure 2B. In practice, in a DOS environment, this can be loaded into the computer as a device driver using the CONFIG.SYS file.
• the added software has the effect of an instruction to write data being replaced by an alternative set of instructions.
• a similar technique may be used to intercept an instruction to read data from the disk and to replace this also with an alternative set of instructions.
• the program On receipt of an INT21h call, the program first checks, whether the call is a write call. If it is, the program checks in step 20 to see if it is supposed to be intercepting write calls. If not, data is written (step 22) to the original data area 24.
• step 26 the program checks the delta mapping table (step 26) to see if a write has previously been made to that area of the original file 24. If (step 28) a write has been made previously, then that part of the delta area 30 previously mapped to that part of the original file 24 in the delta mapping table, is overwritten with the new data (step 32). If write calls are being intercepted but a write has not previously been made to that part of the original file 24 then the write is added to the table and a new area of the delta area is allocated to that piece of data (step 34). The data is then written to the new delta area (step 36).
• step 38 If the call is a read call, the program checks in step 38 whether it is supposed to be intercepting read calls. If it is not, then data is read from the original file 24 (step 40).
• step 42 If the program is intercepting read calls, then it first checks in step 42 to see if the call came from the backup application program. If the call did come from the backup application, the program passes to step 40 and data is read from the original file 24.
• the program checks the delta mapping table to see if a write to the area of the original file 24 corresponding to that of the read call, has been re-directed to the delta area (step 44). If it has not been re-directed (step 46) the program passes back to step 40 and the data is read from the original file 24. If the read has been re-directed, the data is read from the correct part of the delta area 30 (step 48) .
• a second embodiment of the invention when read and write operations are being intercepted during the recording process, instead of storing the changed data in the delta area, when the first write to a particular area of a file is made, the original data is stored in the delta area and the original file is allowed to be changed. An entry is made in the mapping table which points to the original contents of that area of the original file which is now stored in the delta area. Any subsequent write to the same area of the file is then ignored by the recording process. Read requests from the backup application are then re-directed to the delta area if the mapping table indicates that that area of the original file has been overwritten.
• step 60 if read intercept is active (step 38) and the read is not from the backup application, the data is read from the original data area. This is the reverse of step 42 in the previous embodiment.
• step 62 is inserted after step 28 which reads data from the original data area before it is overwritten for the first time.
• Step 36 of the first embodiment is modified to form new step 64 which incorporates an additional write to the delta area of the data read in step 62.
• Step 32 is modified to form new step 66 which no longer redirects the data to overwrite the delta area but instead writes the data to over the original data area.
• This second method is slower than the first since each file write request becomes a read and two writes as opposed to the single re-directed write of the first method.
• the second method does not require the changes stored in the delta area to be applied to the original file once the backup is complete.
• the delta mapping table and the data stored in the delta area may simply be discarded. Additionally because of this difference, provision does not need to be made to recover from a fatal condition (such as a power failure) occurring when the delta changes have not been fully applied to the original data.
• the interception of read operations can take place within the backup application and need not involve the use of an operating system routine.
• the method may also include the steps of recognizing the user and only backing up that user's files in the way described above. In this way it is not necessary to intercept all data transactions but only those relating to that user.

Landscapes

• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• Quality & Reliability (AREA)
• Physics & Mathematics (AREA)
• General Engineering & Computer Science (AREA)
• General Physics & Mathematics (AREA)
• Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
• Techniques For Improving Reliability Of Storages (AREA)

Priority Applications (10)

Applications Claiming Priority (1)

Publications (2)

Family

ID=10749432

Family Applications (1)

Country Status (3)

Cited By (2)

Citations (2)

Family Cites Families (1)

• 1994
  • 1994-07-20 WO PCT/GB1994/001581 patent/WO1995019599A1/en active IP Right Grant
  • 1994-07-20 KR KR1019940704788A patent/KR100285534B1/ko not_active IP Right Cessation
  • 1994-07-20 JP JP06525841A patent/JP2000514211A/ja active Pending

Patent Citations (2)

Also Published As

Similar Documents

Legal Events