System and methods for improved file management in a multi-user environment

• US 5,692,178 A
• Filed: 07/16/1996
• Issued: 11/25/1997
• Est. Priority Date: 08/20/1992
• Status: Expired due to Term

• Current Assignee: Borland Software Corporation (Open Text Corporation)
• Original Assignee: Borland International, Inc. (Open Text Corporation)
• Inventors: Shaughnessy, Steven T.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Alam, Hosain T.
• Application Number:
  US08/682,055
• Time in Patent Office: 497 Days
• Field of Search: 395/608, 395/477, 395/414
• US Class Current: 1/1
• CPC Class Codes: G06F 16/1774 Locking methods, e.g. locking methods for file systems allowing shared and concurrent access to files
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users
• CPC Class Codes: Y10S 707/99956 File allocation

Abstract

A computer system having concurrently shared objects or resources is described. An exemplary embodiment includes a multi-user database management system having information tables and related objects stored in shared directories on a file server. A plurality of lock types, including directory lock, full lock, write lock, prevent full lock, and prevent write lock, are provided for controlling concurrent access. Methods are described for managing locks by creating a special lock file for each shared directory that is accessed. The lock file stores at least one logical lock file having locking or concurrency information specific to a family of related members. The logical lock file itself includes a plurality of entries for specifying concurrency information of associated family members. A shared object or resource is accessed according to the information retrieved from the corresponding logical lock file.

Claims

1. In a multi-client data processing system, an improved method for storing files on a storage device, the method comprising:

storing a plurality of files as a plurality of logical files stored in a single physical file, said physical file residing as a single disk file on a storage device, each logical file being stored as a contiguous sequence of uniform storage blocks in the physical file;

storing with the physical file a directory describing each logical file stored within the physical file;

receiving a first request from a client for information from a particular logical file;

in response to said first request, determining from said directory where said logical file is located in said physical file, and storing at the client information comprising a pointer describing that location, so that a subsequent request for the logical lock file can be satisfied without the directory;

receiving a second request from another client for storing in the logical file new information which causes overflow of the logical file;

in response to said second request, storing the new information by;

(i) appending to one end of the physical file a sufficient number of new uniform storage blocks for storing the new information,

(ii) storing information for the logical lock file together with the new information being added in said new uniform storage blocks,

(iii) marking storage blocks which previously comprised the logical file as invalid, and

(iv) storing with the invalidated storage blocks a pointer which points to said storage blocks being appended to said physical file; and

as each client requests access to information stored in the logical file, updating the pointer stored at the client to reference a location in the physical file where the new storage blocks have been appended.

2. The method of claim 1, wherein each uniform storage block is allocated as a multiple of a disk sector size.

3. The method of claim 2, wherein said appending step includes appending the uniform storage blocks as a contiguous sequence of storage blocks, so that said logical file can always be read from disk in a single disk read operation.

4. The method of claim 1, wherein each logical file itself includes a directory describing contents stored at that file.

5. The method of claim 1, further comprising:

storing a high water mark in the physical file for indicating where new storage blocks can be appended.

6. The method of claim 1, wherein said directory stores information for locating said logical files, said directory including a plurality of entries each of which stores a starting address for one logical file.

7. The method of claim 1, wherein each particular logical file includes a header storing a flag for indicating whether the particular logical file has moved within said physical file.

8. The method of claim 7, wherein said logical file header for the particular logical file stores a pointer to a new logical file if said flag indicates that the logical file has moved.

9. The method of claim 1, wherein each at least one logical file stores concurrency information specific to a particular resource shared among multiple clients.

10. The method of claim 1, wherein said physical file is a separate disk file stored on a server computer, and wherein one physical file is created for each directory of the server computer which has resources being shared among multiple clients.

11. The method of claim 1, wherein each logical file comprises a number of uniform storage blocks allocated within said physical file as a multiple of a common disk unit.

12. The method of claim 1, wherein said physical file comprises a physical lock file and wherein said plurality of logical files comprise a plurality of logical lock files, each logical lock file storing information for accessing particular resources of a database, said physical lock file having a plurality of uniform storage areas each of which stores no more than one logical lock file.

13. A multi-client data processing system comprising:

a server computer storing a plurality of files as a plurality of logical files stored in a single physical file, said physical file residing as a single disk file on a storage device of said server computer, each logical file being stored as a contiguous sequence of uniform storage blocks in the physical file, said physical file including a directory describing each logical file stored within the physical file;

a client computer capable of generating a first request from a client for information from a particular logical file;

means, responsive to said first request, for determining from said directory where said logical file is located in said physical file, said means storing at the client computer information comprising a pointer describing that location, so that a subsequent request for the logical lock file can be satisfied without the directory;

means for receiving a second request from another client computer requesting storage of new information in the logical file, said new information causing overflow of the logical file;

means, responsive to said second request, for storing the new information by;

(i) appending to one end of the physical file a sufficient number of new uniform storage blocks for storing the new information,

(ii) storing information for the logical lock file together with the new information being added in said new uniform storage blocks,

(iii) marking storage block which previously comprised the logical file as invalid, and

(iv) storing with the invalidated storage blocks a pointer which points to said storage blocks being appended to said physical file; and

means for updating the pointer stored at the client computer to reference a location in the physical file where the new storage blocks have been appended, as each client requests access to information stored in the logical file.

14. The system of claim 13, wherein each uniform storage block is allocated as a multiple of a disk sector size.

15. The system of claim 14, wherein means for storing the new information includes:

means for appending the uniform storage blocks as a contiguous sequence of storage blocks, so that said logical file can always be read from disk in a single disk read operation.

16. The system of claim 13, wherein each logical file itself includes a directory describing contents stored at that file.

17. The system of claim 13, further comprising:

means for storing a high water mark in the physical file for indicating where new storage blocks can be appended.

18. The system of claim 13, wherein said directory stores information for locating said logical files, said directory including a plurality of entries each of which stores a starting address for one logical file.

19. The system of claim 13, wherein each particular logical file includes a header storing a flag for indicating whether the particular logical file has moved within said physical file.

20. The system of claim 19, wherein said logical file header for the particular logical file stores a pointer to a new logical file if said flag indicates that the logical file has moved.

21. The system of claim 13, wherein each at least one logical file stores concurrency information specific to a particular resource shared among multiple clients.

22. The system of claim 13, wherein said physical lock file is a separate disk file stored on the server computer, and wherein one physical lock file is created for each directory of the server computer which has resources being shared among multiple clients.

23. The system of claim 13, wherein each logical file comprises a number of uniform storage blocks allocated within said physical file as a multiple of a common disk unit.

24. The system of claim 13, wherein said physical file comprises a physical lock file and wherein said plurality of logical files comprise a plurality of logical lock files, each logical lock file storing information for accessing particular resources of a database, said physical lock file having a plurality of uniform storage areas each of which stores no more than one logical lock file.