Data revision synchronization in a dispersed storage network

• US 8,819,179 B2
• Filed: 11/24/2010
• Issued: 08/26/2014
• Est. Priority Date: 10/09/2007
• Status: Active Grant

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Cleversafe Incorporated (International Business Machines Corporation)
• Inventors: Resch, Jason K., Mark, Zachary J., Baptist, Andrew, Volvovski, Ilya
• Primary Examiner(s): Joo, Joshua
• Assistant Examiner(s): Naji, Younes
• Application Number:
  US12/954,232
• Publication Number:
  US 20110071988A1
• Time in Patent Office: 1,371 Days
• US Class Current: 709/218
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• CPC Class Codes: G06F 2211/1028 Distributed, i.e. distributed RAID systems with parity
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

A method begins by a processing module determining dispersed storage (DS) units that are storing a set of encoded data slices associated with a data segment and sending a revision level check request message to each of the DS units. The method continues with the processing module receiving revision level check response messages from at least some of the DS units within a time period to produce received revision level check response messages and determining concurrency of a revision level of the set of encoded data slices stored by the DS units based on the received revision level check response messages. The method continues with the processing module initiating a revision concurrency process when less than a threshold number of encoded data slices have concurrency of the revision level.

Claims

1. A method comprises:

determining dispersed storage (DS) units that are storing a set of encoded data slices associated with a data segment;

sending a revision level check request message to each of the DS units;

receiving revision level check response messages from at least some of the DS units within a time period to produce received revision level check response messages;

determining concurrency of a revision level of the set of encoded data slices stored by the DS units based on the received revision level check response messages;

when less than a threshold number of encoded data slices have concurrency of the revision level, initiating a revision concurrency process; and

when at least the threshold number, but less than a pillar width number, of encoded data slices have concurrency of the revision level, sending a write finalize message to each of the DS units, wherein the write finalize message includes a revision level and a slice name of a corresponding one of the set of encoded data slices.

2. The method of claim 1, wherein the initiating a revision concurrency process comprises:

when less than the threshold number of encoded data slices have concurrency of a latest revision level, determining whether the threshold number of encoded data slices have concurrency of a previous revision level; and

when at least the threshold number of encoded data slices have concurrency of the previous revision level, sending an undo transaction request message to each of the DS units, wherein the undo transaction request message includes the latest revision level and a slice name of a corresponding one of the set of encoded data slices.

3. The method of claim 2 further comprises:

when at least the threshold number of encoded data slices does not have concurrency of the previous revision level, determining whether the revision level check response message has been received from each of the DS units;

when the revision level check response message was not received from one of the DS units, sending a retry revision level check request message to the one of the DS units; and

when the revision level check response message was received from each of the DS units, initiating a rebuilding process or a re-write process.

4. The method of claim 1, wherein the determining the DS units comprises at least one of:

receiving a multiple revisions message from a revisions agent regarding the data segment, wherein the data segment is identified based on one or more of a revision count, a slice name, and a revision level number;

selecting the DS units from a list of DS units; and

selecting the DS units from a list of data segments for revisions testing.

5. The method of claim 1, wherein a revision level check response message of the revision level check response messages comprises at least one of:

a revision count;

a revision level number; and

a slice length.

6. A computer comprises:

an interface; and

a processing device operable to;

determine dispersed storage (DS) units that are storing a set of encoded data slices associated with a data segment;

send, via the interface, a revision level check request message to each of the DS units;

receive, via the interface, revision level check response messages from at least some of the DS units within a time period to produce received revision level check response messages;

determine concurrency of a revision level of the set of encoded data slices stored by the DS units based on the received revision level check response messages;

initiate a revision concurrency process when less than a threshold number of encoded data slices have concurrency of the revision level; and

send, via the interface, a write finalize message to each of the DS units, wherein the write finalize message includes a revision level and a slice name of a corresponding one of the set of encoded data slices when at least the threshold number, but less than a pillar width number, of encoded data slices have concurrency of the revision level.

7. The computer of claim 6, wherein the processing device further functions to initiate the revision concurrency process by:

determining whether the threshold number of encoded data slices have concurrency of a previous revision level when less than the threshold number of encoded data slices have concurrency of a latest revision level; and

sending, via the interface, an undo transaction request message to each of the DS units, wherein the undo transaction request message includes the latest revision level and a slice name of a corresponding one of the set of encoded data slices when at least the threshold number of encoded data slices have concurrency of the previous revision level.

8. The computer of claim 7, wherein the processing device further functions to:

determine whether the revision level check response message has been received from each of the DS units when at least the threshold number of encoded data slices does not have concurrency of the previous revision level;

send, via the interface, a retry revision level check request message to the one of the DS units when the revision level check response message was not received from one of the DS units; and

initiate a rebuilding process or a re-write process when the revision level check response message was received from each of the DS units.

9. The computer of claim 6, wherein the processing device further functions to determine the DS units by at least one of:

receiving, via the interface, a multiple revisions message from a revisions agent regarding the data segment, wherein the data segment is identified based on one or more of a revision count, a slice name, and a revision level number;

selecting the DS units from a list of DS units; and

selecting the DS units from a list of data segments for revisions testing.

10. The computer of claim 6, wherein the revision level check response message of the revision level check response messages comprises at least one of:

a revision count;

a revision level number; and

a slice length.