Pre-existing content replication

• US 7,620,721 B2
• Filed: 02/28/2006
• Issued: 11/17/2009
• Est. Priority Date: 02/28/2006
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Suriyanarayanan, Guhan, Liu, Huisheng, Bjørner, Nikolaj S., Balakrishnan, Shobana M.
• Primary Examiner(s): Chankong; Dohm
• Assistant Examiner(s): Hussain; Tauqir
• Application Number:
  US11/364,235
• Publication Number:
  US 20070220067A1
• Time in Patent Office: 1,358 Days
• Field of Search: 709223-226, 707/E17.005, 707/E17.007, 707/E17.032
• US Class Current: 709/226
• CPC Class Codes: G06F 16/178 Techniques for file synchronisation in file systems
• CPC Class Codes: G06F 16/1844 Management specifically adapted to replicated file systems
• CPC Class Codes: H04L 63/102 Entity profiles
• CPC Class Codes: H04L 63/104 Grouping of entities

Abstract

Aspects of the subject matter described herein relate to replicating initial content in resource replication systems. In aspects, each member participating in a replica group obtains a content set state from a directory service. The content set state indicates whether the member is allowed to synchronize resources with another member participating in the replica group. If the member is designated as a primary member, the member may replicate out resources to its downstream partners. If a member is not designated as a primary member, the member may not replicate out resources to its downstream partners until it has successfully synchronized resources with an upstream partner. Fence values may be set which control which resources win (and are replicated out) during synchronization.

Claims

1. A computer-readable medium comprised of computer storage media having computer-executable instructions for implementing a replication method of helping to insure correct initialization behavior for a pre-existing content set having resource data such as a set of files under a root directory that exists on a computing system of a member participating in a replica group, the replica group providing the resources of the pre-existing content set to the computing systems of members of the replica group, and wherein the executable instructions for implementing the replication method are comprised of:

beginning a replication service on a computing system of a first member of said replica group, and when the replication service begins executing, requesting from the computing system'"'"'s directory service configuration information for said pre-existing content set at the first member'"'"'s computing system;

the replication service then receiving the requested configuration information from the directory service, at least one attribute of the configuration information indicating if the first member is designated as the primary member for said pre-existing content set, and if the first member is designated as a primary member for the pre-existing content set, then;

at the replication service, setting state data as initial-primary for said pre-existing content set;

at the replication service, marking resource metadata records of each resource for said pre-existing content set with a fence value of initial-primary; and

at the replication service, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in an initial-primary state; and

if the first member is not designated as the primary member for said pre-existing content set, then;

at the replication service of the first member, setting state data as initial-sync-needed for said pre-existing content set so that the state data can be later used by the replication service to determine if the replication service should accept replication requests from downstream members of the replica group;

at the replication service, marking resource metadata records of each resource for said pre-existing content set of the first member with a fence value of initial-sync-needed; and

at the replication service of the first member, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in an initial-sync-needed state;

thereafter, when a replication service at a downstream member of the replica group establishes a connection with the first member of the replica group and requests replication of its content set, permitting such replication to occur from the first member if the state of the pre-existing content set at the first member is in the initial-primary state, otherwise returning an error; and

if the downstream member has successfully completed replication of its content set so as to be synchronized with the first member, the downstream member'"'"'s content set state then being set to normal so that other downstream members of the replica group my synchronize to any upstream member for which the content set of that member is set to either initial-primary or normal.

2. The computer-readable medium of claim 1, wherein the computer-executable instructions further comprise instructions by which the content set of each member of the replica group maintains at its replication service a version vector that represents the machine'"'"'s current latest version of each resource within its content set and the latest versions for each resource replicated from other members of the replica group.

3. The computer-readable medium of claim 2, wherein the computer-executable instructions further comprise instructions by which a downstream member of the replica group transmits its version vector when connecting to an upstream member, and the upstream member then subtracting the version vector of the downstream from its version vector so that only those resources of the downstream member are replicated for which the version of the resource is different from that of the upstream member.

4. The computer-readable medium of claim 3, wherein the computer-executable instructions further comprise instructions by which resource versions of a content set are stored as metadata in the replication service of a member.

5. The computer-readable medium of claim 3, wherein the computer-executable instructions further comprise instructions by which a downstream member, prior to requesting those resources of the upstream member for which the version of the resource is different from that of the downstream member, the downstream member first requests tombstones from the upstream member in order to delete any resources no longer active in the content set.

6. The computer-readable medium of claim 2, wherein two members of the replica group which are connecting to complete replication of their content sets with one another have already synchronized with another member of the replica group so that said two members each have content sets that are in the initial-primary or normal state.

7. The computer-readable medium of claim 1, wherein the computer-executable instructions further comprise instructions by which the content set of only one member of the replica group may be set to a normal state while the content set state of all other members of the replica group is in an initial-sync-needed state.

8. The computer-readable medium of claim 1, further comprising requesting that the directory service of the first member update information that indicates when the first member is no longer a designated member of the replica group from which resources are to be propagated.

9. A replication method of helping to insure correct initialization behavior for a pre-existing content set having resource data such as a set of files under a root directory that exists on a computing system of a member participating in a replica group, the replica group providing the resources of the pre-existing content set to the computing systems of the members of the replica group, the method comprising:

beginning a replication service on a computing system of a first member of said replica group, and when the replication service begins executing, requesting from the computing system'"'"'s directory service configuration information for said pre-existing content set at the first member'"'"'s computing system;

the replication service then receiving the requested configuration information from the directory service, at least one attribute of the configuration information indicating if the first member is designated as the primary member for said pre-existing content set, and if the first member is designated as a primary member for the pre-existing content set, then;

at the replication service, setting state data as initial-primary for said pre-existing content set;

at the replication service, marking resource metadata records of each resource for said pre-existing content set with a fence value of initial-primary; and

at the replication service, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in a initial-primary state; and

if the first member is not designated as the primary member for said pre-existing content set, then;

at the replication service of the first member, setting state data as initial-sync-needed for said pre-existing content set so that the state data can be later used by the replication service to determine if the replication service should accept replication requests from downstream members of the replica group;

at the replication service, marking resource metadata records of each resource for said pre-existing content set of the first member with a fence value of initial-sync-needed; and

at the replication service of the first member, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in an initial-sync-needed state;

thereafter, when a replication service at a downstream member of the replica group establishes a connection with the first member of the replica group and requests replication of its content set, permitting such replication to occur from the first member if the state of the pre-existing content set at the first member is in the initial-primary state, otherwise returning an error; and

if the downstream member has successfully completed replication of its content set so as to be synchronized with the first member, the downstream member'"'"'s content set state then being set to normal so that other downstream members of the replica group my synchronize to any upstream member for which the content set of that member is set to either initial-primary or normal.

10. The method of claim 9, wherein the content set of each member of the replica group maintains at its replication service a version vector that represents the machine'"'"'s current latest version of each resource within its content set and the latest versions for each resource replicated from other members of the replica group.

11. The method of claim 9, wherein a downstream member of the replica group transmits its version vector when connecting to an upstream member, and the upstream member then subtracting the version vector of the downstream from its version vector so that only those resources of the downstream member are replicated for which the version of the resource is different from that of an upstream member.

12. The method of claim 11, wherein resource versions of a content set are stored as metadata in the replication service of a member.

13. The method of claim 9, wherein a downstream member, prior to requesting those resources of the upstream member for which the version of the resource is different from that of the downstream member, the downstream member first requests tombstones from the upstream member in order to delete any resources no longer active in the content set.

14. A computer-readable medium comprised of computer storage media having computer-executable instructions for implementing a replication method of helping to insure correct initialization behavior for a pre-existing content set having resource data such as a set of files under a root directory that exists on a computing system of a member participating in a replica group, the replica group providing the resources of the pre-existing content set to the computing systems of members of the replica group, and wherein the executable instructions for implementing the replication method are comprised of:

beginning a replication service on a computing system of a first member of said replica group, and when the replication service begins executing, requesting from the computing system'"'"'s directory service configuration information for said pre-existing content set at the first member'"'"'s computing system;

the replication service then receiving the requested configuration information from the directory service, at least one attribute of the configuration information indicating if the first member is designated as the primary member for said pre-existing content set, and if the first member is designated as a primary member for the pre-existing content set, then;

at the replication service, setting state data as initial-primary for said pre-existing content set;

at the replication service, marking resource metadata records of each resource for said pre-existing content set with a fence value of initial-primary; and

at the replication service, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in an initial-primary state; and

if the first member is not designated as the primary member for said pre-existing content set, then;

at the replication service of the first member, setting state data as initial-sync-needed for said pre-existing content set so that the state data can be later used by the replication service to determine if the replication service should accept replication requests from downstream members of the replica group;

at the replication service, marking resource metadata records of each resource for said pre-existing content set of the first member with a fence value of initial-sync-needed; and

at the replication service of the first member, entering into an event log an entry that indicates that said pre-existing content set has been successfully initialized, and is now in an initial-sync-needed state;

thereafter, when a replication service at a downstream member of the replica group establishes a connection with the first member of the replica group and requests replication of its content set, permitting such replication to occur from the first member if the state of the pre-existing content set at the first member is in the initial-primary state, otherwise returning an error; and

if the downstream member has successfully completed replication of its content set so as to be synchronized with the first member, the downstream member'"'"'s content set state then being set to normal so that other downstream members of the replica group are available to synchronize to any upstream member for which the content set of that member is set to either initial-primary or normal;

the fence value for each replicated resource of the content set being set to normal at the replication service of the downstream member that has successfully completed replication of the pre-existing content set;

thereafter when any downstream member of the replica group establishes connection with any upstream member of the replica group, if the content set of the upstream member is in the initial-primary or normal state, allowing replication of the content set to occur with the requesting downstream member, and using the state of the fence value for each resource to resolve any conflicts according to the following;

a resource with a normal fence state takes precedence over a resource marked with an initial-primary fence state, a resource with a primary-initial state takes precedence over a resource with an initial-sync-needed state.

15. The computer-readable medium of claim 14, wherein the computer-executable instructions further comprise instructions by which the content set of each member of the replica group maintains at its replication service a version vector that represents the machine'"'"'s current latest version of each resource within its content set and the latest versions for each resource replicated from other members of the replica group.

16. The computer-readable medium of claim 15, wherein the computer-executable instructions further comprise instructions by which a downstream member of the replica group transmits its version vector when connecting to an upstream member, and the upstream member then subtracting the version vector of the downstream from its version vector so that only those resources of the downstream member are replicated for which the version of the resource is different from that of the upstream member.

17. The computer-readable medium of claim 16, wherein the computer-executable instructions further comprise instructions by which resource versions of a content set are stored as metadata in the replication service of a member.

18. The computer-readable medium of claim 16, wherein the computer-executable instructions further comprise instructions by which a downstream member, prior to requesting those resources of the upstream member for which the version of the resource is different from that of the downstream member, the downstream member first requests tombstones from the upstream member in order to delete any resources no longer active in the content set.

19. The computer-readable medium of claim 15, wherein two members of the replica group which are connecting to complete replication of their content sets with one another have already synchronized with another member of the replica group so that said two members each have content sets that are in the initial-primary or normal state.

20. The computer-readable medium of claim 14, wherein the computer-executable instructions further comprise instructions by which the content set of only one member of the replica group may be set to a normal state while the content set state of all other members of the replica group is in an initial-sync-needed state.

21. The computer-readable medium of claim 14, further comprising requesting that the directory service of the first member update information that indicates when the first member is no longer a designated member of the replica group from which resources are to be propagated.

22. The computer-readable medium of claim 14, wherein the computer-executable instructions further comprise instructions by which conflicts are resolved by giving the state of a fence value a timestamp, and wherein any resource with a fence value marked with the most recent timestamp will be given precedence.