Parallel file system and method with a metadata node

• US 5,974,424 A
• Filed: 07/11/1997
• Issued: 10/26/1999
• Est. Priority Date: 07/11/1997
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Curran, Robert, Shmueli, Boaz, Schmuck, Frank B., Wyllie, James Christopher
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Shah, Sanjiv
• Application Number:
  US08/893,723
• Time in Patent Office: 837 Days
• Field of Search: 707/200, 707/201, 707/8
• US Class Current: 1/1
• CPC Class Codes: G06F 16/1858 Parallel file systems, i.e. file systems supporting multiple processors
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users

Abstract

A computer system having a shared disk file system running on multiple computers each having their own instance of an operating system and being coupled for parallel data disks. One of the computers serves as a metadata node and manages metadata for parallel read and write actions for a particular file. Metadata is transferred from and to this file from the other computers via this medadata node computer.

Claims

1. The method for a computer system having a shared disk environment for a plurality of files stored in said shared disk environment and in which multiple computer nodes can each directly access, update or enlarge the same file in parallel, the method of operating the shared disk environment, including the steps of:

transferring metadata for said same file to and from said shared disk environment via one of said multiple computer nodes;

transferring metadata for the same file to and from said shared disk environment to other than said one of said multiple computer nodes via said one of said multiple computer nodes;

wherein for the management of metadata information for a file in said shared disk environment system a single computer node is selected as said one of said multiple computer nodes for that file;

wherein the method includes reflecting the metadata on disk by keeping cached a copy of the metadata on the metadata computer node;

wherein each computer node has an operating system instance which includes a daemon for deferred sync which periodically flushes dirty data directly to disk and flushes dirty metadata to the single computer node, and wherein the single computer node flushes dirty metadata to disk when it gets a signal from said sync daemon.

2. The method for a computer system according to claim 1, wherein said metadata includes the file size, the file access and modification times, and the addresses of the file'"'"'s data blocks.

3. The method for a computer system according to claim 1, wherein said one of said multiple computer nodes for the file is responsible for handling all input output activity of the metadata from and to one or more disks on which the metadata resides.

4. The method for a computer system according to claim 1, wherein other computer nodes of said multiple computer system communicate via said single computer node in order to fetch or update metadata information.

5. The method for a computer system according to claim 1, wherein said single computer node is elected as the first computer node that accesses an associated file system identified by said metadata, permitting all disks that make up the file system to be accessed by multiple processors.

6. The method for a computer system according to claim 1, wherein all other computer nodes also keep a cached copy of the metadata which they have read via the single computer node, which they augment as needed.

7. The method for a computer system according to claim 1, wherein each element of metadata has its own pattern of usage and special characteristic'"'"'s which are determined by said single computer node for updating periodically.