Methods to perform disk writes in a distributed shared disk system needing consistency across failures

• US 7,930,278 B2
• Filed: 02/20/2007
• Issued: 04/19/2011
• Est. Priority Date: 02/13/1998
• Status: Expired due to Term

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Srihari, Vinay, Brower, David, Bridge, William H., Bamford, Roger J., Chandrasekaran, Sashikanth, Chan, Wilson Wai Shun, Macnaughton, Neil
• Primary Examiner(s): Le; Debbie
• Assistant Examiner(s): Tran; Bao G
• Application Number:
  US11/708,857
• Publication Number:
  US 20070174541A1
• Time in Patent Office: 1,519 Days
• Field of Search: 707/204, 707/1, 707/202, 707/687, 707/610, 711/145, 711/114, 711/135, 711/113
• US Class Current: 707/695
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 12/0804 with main memory updating G06F12/0806 takes precedence
• CPC Class Codes: G06F 12/0815 Cache consistency protocols
• CPC Class Codes: G06F 12/0817 using directory methods
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users
• CPC Class Codes: Y10S 707/99953 Recoverability
• CPC Class Codes: Y10S 707/99954 Version management

Abstract

Techniques are provided for managing caches in a system with multiple caches that may contain different copies of the same data item. Specifically, techniques are provided for coordinating the write-to-disk operations performed on such data items to ensure that older versions of the data item are not written over newer versions, and to reduce the amount of processing required to recover after a failure. Various approaches are provided in which a master is used to coordinate with the multiple caches to cause a data item to be written to persistent storage. Techniques are also provided for managing checkpoints associated with the caches, where the checkpoints are used to determine the position at which to begin processing recovery logs in the event of a failure.

Claims

1. A method for managing caches in a system with multiple caches that may contain different copies of a data item, comprising the steps of:

modifying the data item in a first cache to create a modified version of the data item;

wherein said data item is one of a plurality of data items stored in persistent storage;

wherein each data item of the plurality of data items is mastered by a master;

wherein the master for each data item of the plurality of data items maintains information about said each data item to coordinate writing of said each data item to persistent storage;

writing the modified version of the data item to persistent storage;

in response to writing the modified version of the data item to persistent storage, performing the steps of;

a node associated with the first cache determining whether any other cache in said multiple caches had created a dirty version of said data item; and

if any other cache in said multiple caches had created a dirty version of said data item, then the node associated with the first cache informing the master of said data item that said modified version of the data item has been written to persistent storage; and

if no other cache in said multiple caches had created a dirty version of said data item, then the step of writing the modified version of the data item is performed without informing said master that said modified version of the data item has been written to persistent storage;

wherein the method is performed by one or more computing devices.

2. The method of claim 1 wherein the step of determining whether any other cache in said multiple caches had created a dirty version of said data item includes inspecting a global dirty flag associated with said data item.

3. The method of claim 2 wherein:

prior to modifying the data item in said first cache, the data item had been modified in a second cache;

the data item is not persistently stored between being modified in said second cache and being modified in said first cache; and

a node associated with the second cache causes said global dirty flag to be set to indicate that said data item is globally dirty.

4. The method of claim 3 wherein the node associated with the second cache causes said global dirty flag to be set in response to transferring said data item from said second cache to another cache of said multiple caches.

5. A computer readable storage storing instructions for managing caches in a system with multiple caches that may contain different copies of a data item, the instructions comprising instructions for performing the steps of:

modifying the data item in a first cache to create a modified version of the data item;

wherein said data item is one of a plurality of data items stored in persistent storage;

wherein each data item of the plurality of data items is mastered by a master;

wherein the master for each data item of the plurality of data items maintains information about said each data item to coordinate writing of said each data item to persistent storage;

writing the modified version of the data item to persistent storage;

in response to writing the modified version of the data item to persistent storage, performing the steps of;

a node associated with the first cache determining whether any other cache in said multiple caches had created a dirty version of said data item; and

if any other cache in said multiple caches had created a dirty version of said data item, then the node associated with the first cache informing the master of said data item that said modified version of the data item has been written to persistent storage; and

if no other cache in said multiple caches had created a dirty version of said data item, then the step of writing the modified version of the data item is performed without informing said master that said modified version of the data item has been written to persistent storage.

6. The computer-readable storage of claim 5 wherein the step of determining whether any other cache in said multiple caches had created a dirty version of said data item includes inspecting a global dirty flag associated with said data item.

7. The computer-readable storage of claim 6 wherein:

prior to modifying the data item in said first cache, the data item had been modified in a second cache;

the data item is not persistently stored between being modified in said second cache and being modified in said first cache; and

a node associated with the second cache causes said global dirty flag to be set to indicate that said data item is globally dirty.

8. The computer-readable storage of claim 7 wherein the node associated with the second cache causes said global dirty flag to be set in response to transferring said data item from said second cache to another cache of said multiple caches.