Data replication method

• US 8,180,729 B2
• Filed: 04/05/2011
• Issued: 05/15/2012
• Est. Priority Date: 11/02/2007
• Status: Active Grant

• Current Assignee: Pivotal Software, Inc. (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: Schuchardt, Bruce J.
• Primary Examiner(s): NGUYEN, PHONG H
• Application Number:
  US13/079,933
• Publication Number:
  US 20110184911A1
• Time in Patent Office: 406 Days
• Field of Search: 707/610, 707/615, 707/616, 707/619, 707/626, 707/613, 709/224
• US Class Current: 707/613
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 16/275 Synchronous replication
• CPC Class Codes: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06

Abstract

To ensure data consistency for a new data replica created for a computing system, the transmission and receipt of messages altering the data are monitored and replication of the data is permitted when all changes to the data that have been transmitted to the data have been received.

Claims

1. A method of replicating a data region of a distributed data storage system, wherein the distributed data storage system comprises a plurality of computing nodes, each computing node capable of storing data regions and supporting processes that access data regions, the method comprising:

transmitting a message indicating an intention to replicate a data region managed by a first process of a first computing node for a second process, wherein the message is transmitted to a plurality of processes that currently access, through at least one communication channel, the data region and thereby cause each of the processes in the plurality of processes to transmit subsequent update operations made by the process to the data region to the second process;

monitoring a status of each of the communication channels to determine whether a number of in-flight update operations transmitted to the data region are received and applied to the data region, wherein in-flight update operations are transmitted prior to the transmission of the message, but not applied by the data region prior to the transmission of the message; and

replicating the data region upon confirmation that the number of in-flight update operations has been received by and applied to the data region to create a replica of the data region for the second process, wherein the subsequent update operations are applied to the replica of the data region by the second process and the in-flight update operations are not applied to the replica of the data region by the second process.

2. The method of claim 1, further comprising the steps of causing the second process to temporarily store the subsequent update operations transmitted by the processes until replication of the data region has been completed;

and performing the stored subsequent update operations after the data region has been replicated for the second process.

3. The method of claim 1, wherein the second process corresponds to a second computing node.

4. The method of claim 1, wherein at least one of the plurality of processes further accesses a previous replica of the data region residing at a computing node other than the first computing node.

5. The method of claim 4, wherein the previous replica is a cache of the data region stored at the computing node corresponding to the at least one process accessing the previous replica.

6. The method of claim 1, wherein, for each communication channel, the monitoring step utilizes a transmission monitor component at the corresponding process that logs a number of update operations transmitted by the corresponding process and a reception monitor component at the data region that logs a number of update operations received by the data region.

7. The method of claim 1, wherein the data region is a subset of data accessible through the distributed data storage system.

8. A non-transitory computer-readable storage medium comprising instructions that when executed on a plurality of computing nodes of a distributed data storage system, each computing node capable of storing data regions and supporting processes that access data regions, cause the distributed data storage system to replicate a data region by performing the steps of:

transmitting a message indicating an intention to replicate a data region managed by a first process of a first computing node for a second process, wherein the message is transmitted to a plurality of processes that currently access, through at least one communication channel, the data region and thereby cause each of the processes in the plurality of processes to transmit subsequent update operations made by the process to the data region to the second process;

monitoring a status of each of the communication channels to determine whether a number of in-flight update operations transmitted to the data region are received and applied to the data region, wherein in-flight update operations are transmitted prior to the transmission of the message but not applied by the data region prior to the transmission of the message; and

replicating the data region upon confirmation that the number of in-flight update operations has been received by and applied to the data region to create a replica of the data region for the second process, wherein the subsequent update operations are applied to the replica of the data region by the second process and the in-flight update operations are not applied to the replica of the data region by the second process.

9. The non-transitory computer-readable medium of claim 8, further including instructions to perform the steps of causing the second process to temporarily store the subsequent update operations transmitted by the processes until replication of the data region has been completed;

and performing the stored subsequent update operations after the data region has been replicated for the second process.

10. The non-transitory computer-readable medium of claim 8, wherein the second process corresponds to a second computing node.

11. The non-transitory computer-readable medium of claim 8, wherein at least one of the plurality of processes further accesses a previous replica of the data region residing at a computing node other than the first computing node.

12. The non-transitory computer-readable medium of claim 11, wherein the previous replica is a cache of the data region stored at the computing node corresponding to the at least one process accessing the previous replica.

13. The non-transitory computer-readable medium of claim 8, wherein, for each communication channel, the monitoring step utilizes a transmission monitor component at the corresponding process that logs a number of update operations transmitted by the corresponding process and a reception monitor component at the data region that logs a number of update operations received by the data region.

14. The non-transitory computer-readable medium of claim 8, wherein the data region is a subset of data accessible through the distributed data storage system.

15. A distributed data storage system comprising a plurality of computing nodes, wherein each computing node is capable of storing data regions and supporting processes that access data regions and wherein the plurality of computing nodes are configured to perform the steps of:

transmitting a message indicating an intention to replicate a data region managed by a first process of a first computing node for a second process, wherein the message is transmitted to a plurality of processes that currently access, through at least one communication channel, the data region and thereby cause each of the processes in the plurality of processes to transmit subsequent update operations made by the process to the data region to the second process;

monitoring a status of each of the communication channels to determine whether a number of in-flight update operations transmitted to the data region are received and applied to the data region, wherein in-flight update operations are transmitted prior to the transmission of the message but not applied by the data region prior to the transmission of the message; and

replicating the data region upon confirmation that the number of in-flight update operations has been received by and applied to the data region to create a replica of the data region for the second process, wherein the subsequent update operations are applied to the replica of the data region by the second process and the in-flight update operations are not applied to the replica of the data region by the second process.

16. The distributed data storage system of claim 15, wherein the plurality of computing nodes are further configured to perform the steps of causing the second process to temporarily store the subsequent update operations transmitted by the processes until replication of the data region has been completed;

and performing the stored subsequent update operations after the data region has been replicated for the second process.

17. The distributed data storage system of claim 15, wherein the second process corresponds to a second computing node.

18. The distributed data storage system of claim 15, wherein at least one of the plurality of processes further accesses a previous replica of the data region residing at a computing node other than the first computing node.

19. The distributed data storage system of claim 18, wherein the previous replica is a cache of the data region stored at the computing node corresponding to the at least one process accessing the previous replica.

20. The distributed data storage system of claim 15, wherein, for each communication channel, the monitoring step utilizes a transmission monitor component at the corresponding process that logs a number of update operations transmitted by the corresponding process and a reception monitor component at the data region that logs a number of update operations received by the data region.