System and method for efficient cache management in a distributed file system

• US 6,119,151 A
• Filed: 03/07/1994
• Issued: 09/12/2000
• Est. Priority Date: 03/07/1994
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Jaji, Sebnem, Ward, Richard Byron, Cantrell, Thomas George, Shaheen, Amal Ahmed
• Primary Examiner(s): Maung, Zarni
• Application Number:
  US08/206,706
• Time in Patent Office: 2,381 Days
• Field of Search: 395/500, 395/200, 395/800, 395/600, 709/216, 709/213, 711/100
• US Class Current: 709/216
• CPC Class Codes: G06F 16/172 Caching, prefetching or hoarding of files

Abstract

A cache manager efficiently supporting both connected and disconnected client operation. A single integrated cache is maintained for file system data, status and pathname information. File system requests are satisfied by the cache if possible and passed to a remote file system for servicing if necessary. The cache manager is independent of operating system file syntax and distributed file system protocol. Disconnected operation is supported through an aggressive caching policy and by logging of file system modifications. Modification logs are created when the client is disconnected by logging objects that optimize the log as it is created. Cache file system objects can have multiple parents each with different naming syntax because the object name is not cached.

Claims

1. A method of managing a file system cache in a client computer system operating under a first operating system, the method comprising the steps of:

intercepting operating system requests for a file system object in a distributed file system;

transforming said requests to remove operating system dependent syntax;

testing a cache in a storage means of said client for the presence of file system object data based on said transformed request;

satisfying said transformed request for said file system object data, if cache data exists;

testing to determine whether a connection exists to said distributed file system;

generating a request for said file system object from said distributed file system, without regard to distributed file system protocol, if no cache data exists and there is a connection;

transmitting said request to said distributed file system; and

rejecting said file system object request, if no cache data exists and there is no connection.

2. The method of claim 1 wherein the step of transmitting said request comprises the steps of:

determining a distributed file system protocol for said requested file system object;

transforming said operating system independent request into a request complying with said determined distributed file system protocol; and

sending said transformed operating system independent request to said distributed file system.

3. The method of claim 1 further comprising the steps of:

performing the following steps whenever said testing step determines that no connection exists between said client computer system and said distributed file system;

testing said file system object to determine a request type;

activating a request logging object for said request type;

invoking a modify log method to log said file system object request; and

logging said file system object request by performing log optimization processing for said request type.

4. A computer file system comprising:

cache means for storing file system object data in a storage means;

cache management means for satisfying operating system requests for file system objects using said cache means, said cache management means being independent of operating system syntax and independent of distributed file system protocol, said cache management means implemented within a client computer system, said cache management means further comprising;

means for intercepting an operating system request, from said first operating system, for a file system object stored in a remote server computer system operating under a second operating system in said distributed file system;

means for transforming said request to remove operating system dependent syntax;

means for testing a cache in a storage means of said client computer system for the presence of file system object data based on said transformed request;

means for satisfying said file system object request from said cache, if said file system object data based on said transformed request is present in said cache;

network means detachably connecting said cache management means to a file system object server; and

remote file access means for accessing file system object requests through said network means.

5. The system of claim 4, further comprising:

object modification logging means for maintaining a log of file system modifications for each operating system request processed by said cache management means whenever said cache means is disconnected from said file system object server.

6. A computer system cache apparatus for maintaining file system information, comprising:

storage means for storing a plurality of file system object data entries;

link means for linking each of said plurality of file system object data entries to a file volume; and

second link means for linking each of said plurality of file system object data entries to either a location in local data storage, a directory cache in local memory, or a symbolic link in local memory, wherein said plurality of file system object data entries are independent of any file system object name and are referencable by multiple file system object parents each having different naming syntax.

7. The computer system of claim 6, further comprising:

logging means for logging for each file volume all changes made to objects referenced by said plurality of file system object data entries when said cache is disconnected from said file volume.

8. The computer system of claim 6, further comprising:

interface means for accepting operating system requests for one or more file system object data entries and transforming the requests into operating system independent requests.

9. The computer system of claim 8, wherein said directory cache comprises:

means for storing resolved directory entries for requested file system data objects; and

means for storing resolved directory entries for objects hierarchically immediately below each path component of said requested file system data objects.

10. The computer system of claim 6, further comprising:

logging means for logging for each file volume all changes made to objects referenced by said plurality of file system object data entries when said cache is disconnected from said file volume.

11. The computer system of claim 6, further comprising:

interface means for accepting operating system requests for one or more file system object entries data and transforming the requests into operating system independent requests.

12. The computer system of claim 11, wherein said directory cache comprises:

means for storing resolved directory entries for requested file system data objects; and

means for storing resolved directory entries for objects hierarchically immediately below each path component of said requested file system data objects.