Electrical computers and digital processing systems: memory – Storage accessing and control – Control technique
Reexamination Certificate
2002-02-15
2004-06-08
Sparks, Donald (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Control technique
C711S221000, C707S793000
Reexamination Certificate
active
06748504
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to data storage systems and more specifically to data storage systems that store snapshots (i.e., indications of the status of stored data at particular points in time).
2. Description of Related Art
Many data storage applications store data on electromechanical systems that are prone to physical failure. Magnetic disk drives are an example of such storage systems. Magnetic disk drives utilize a rotating magnetic platter that has a read/write head suspended above but very close to the platter. Data is stored by creating a magnetic recording on the magnetic platter. Contamination on the surface of the magnetic platter often causes damage to the magnetic surface and the recording, thereby rendering the data inaccessible. Other data storage systems are subject to physical or electrical damage and may lose their data.
Many data storage systems organize stored data according to a file metaphor. In these storage systems, related data are stored in a file, and the data storage system stores multiple files. The data storage system then stores references to the multiple files in order to access the data in those files. A single file may be stored in contiguous locations in the data storage device, or the data may be stored in disparate locations. Storage of data in disparate locations in a data storage device often results when a large data file is to be stored on a device that already stores many files and the large data file must be broken up to fit in the free area of the storage device. Data is also often stored in disparate locations when additional data is added to an existing file. The assembly of stored data into files and the structure of those files on a data storage device is referred to as a file system.
Data storage systems often store images or snapshots of the data that is currently stored in the file system. The data contents of a snapshot are the data that is stored within the active file system or a previous snapshot at the time the snapshot was captured. One use of snapshots is to store the state of the file system on another storage system, such as another disk drive or magnetic tape storage system. Another use of file system snapshots is to be able to recreate data that was deleted, i.e., to access previous versions of files that have been deleted or updated.
The data stored within files in a file system have associated metadata to describe the data and allow access to the data. Some existing methods for taking snapshots of a file system defer actually copying the data in the original file system to the snapshot until the data in the original system is modified. Such systems are referred to as “Copy-on-write” systems since the data is not copied to the snapshot data until a write is performed on the original data.
Existing copy-on-write snapshot systems differ in how and when metadata is copied. Existing snapshot systems copy into the snapshot data file some or all of the metadata that describes the data file storage locations at the time the snapshot is made. These existing systems create snapshot data sets that include file references to the original data file in the original file system. This results in multiple references to the same data block in the original file system, the reference in the metadata of the original file system as well as the references in each of the snapshot data sets.
An exemplary file system data structure
400
is illustrated in
FIG. 4
that contains two inodes, one in the active file system and one in a snapshot dataset, that each point to the same data block
406
. The existence of multiple references to a single data block within the original file system impacts the requirements of the original file system. File systems that utilize snapshots that each store a reference to an original data block must maintain an indication of each reference to that data block in order to determine if the data block is in-use or free. Without multiple references, a single bit is able to indicate if a data block is in-use or free. With the multiple references, multiple bits are required to track the multiple references and ensure that no references exist to the data block prior to declaring the data block “free.” This need to track the multiple references complicates the operation of the file system, limits the total number of snapshots, and also complicates, or renders impossible, the implementation of such snapshot system with file systems that do not support tracking multiple references to a data block.
Therefore a need exists to overcome the problems with the prior art as discussed above, and particularly for a way to more efficiently utilize system kernel memory within data processing equipment to support time sensitive processing tasks such as external data communications processing.
SUMMARY OF THE INVENTION
Briefly, in accordance with the present invention, disclosed is a system, method and computer readable medium for deferring copy-on-write of a snapshot. In an embodiment of the present invention, the method includes the generation of snapshot of a source file. Upon modification of a first data block referenced by the source file, the first data block is referenced by the snapshot and a second data block is allocated for the source file. Then, a first variable associated with the source file is set to a value indicating the lack of a complete source file data block and a second variable associated with the source file is set to a value indicating the valid portion of the second data block. Any portion of the second data block that is overwritten is considered valid. The second data block is then modified and the second variable is changed to reflect the modification.
In another embodiment of the present invention, upon reception of a request to read the source file, it determined from the first variable whether there is a complete copy of the source file. If so, the second data block is retrieved. Otherwise, it is determined whether the read request refers to a portion of the source file corresponding to a portion of the second data block that is valid. If so, the corresponding portion of the second data block is retrieved. Otherwise, the portion of the first data block that is requested and found to be invalid in the second data block is copied to the second data block. Subsequently, that portion of the second data block is retrieved.
The foregoing and other features and advantages of the present invention will be apparent from the following more particular description of the preferred embodiments of the invention, as illustrated in the accompanying drawings.
REFERENCES:
patent: 4875159 (1989-10-01), Cary et al.
patent: 6484186 (2002-11-01), Rungta
patent: 2003/0140070 (2003-07-01), Kaczmarski et al.
patent: 2003/0140204 (2003-07-01), Ashton et al.
patent: 2003/0182253 (2003-09-01), Chen et al.
patent: 1003103 (2000-05-01), None
patent: WO00/07104 (2000-02-01), None
patent: WO02/19110 (2002-03-01), None
Sawdon Wayne A.
Schmuck Frank B.
Cutter Lawrence D.
Farrokh Hashem
Fleit Kain Gibbons Gutman Bongini & Bianco P.L.
Gibbons Jon A.
International Business Machines - Corporation
LandOfFree
Deferred copy-on-write of a snapshot does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Deferred copy-on-write of a snapshot, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Deferred copy-on-write of a snapshot will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3308321