Synchronization of group state data when rejoining a member to a primary-backup group in a clustered computer system

• US 20030056013A1
• Filed: 09/14/2001
• Published: 03/20/2003
• Est. Priority Date: 09/14/2001
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: Miller, Robert, Morey, Vicki Lynn, Williams, Laurie Ann, Laschkewitsch, Clinton Gene
• Granted Patent: US 7,231,461 B2
• Time in Patent Office: Days
• US Class Current: 709/249
• CPC Class Codes: H04L 67/01 Protocols
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

An apparatus, program product and method to synchronize group state data in a primary-backup group in connection with the rejoining of a member to the primary-backup group in a clustered computer system. Each member in the group includes a copy of replicated group state data for the primary-backup group. In connection with rejoining the member, it is determined whether the rejoining member is the primary member for the primary-backup group. Then, a selection is made between member and group overwrite operations based upon such determination. The member overwrite operation includes overwriting the copy of the replicated group state data for the rejoining member with data from the copy of the replicated group state data for an existing member in the primary-backup group. The group overwrite operation includes overwriting the copy of the replicated group state data for the existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member.

Claims

1. A method for rejoining a member to a primary-backup group in a clustered computer system, wherein the primary-backup group is of the type including a primary and at least one backup member, each of which including a copy of replicated group state data for the primary-backup group, the method comprising:

(a) determining whether the rejoining member is the primary member for the primary-backup group; and

,

(b) selecting between member and group overwrite operations based upon such determination, wherein the member overwrite operation includes overwriting the copy of the replicated group state data for the rejoining member with data from the copy of the replicated group state data for an existing member in the primary-backup group, and wherein the group overwrite operation includes overwriting the copy of the replicated group state data for the existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member.

2. The method of claim 1, wherein determining whether the rejoining member is the primary member for the primary-backup group comprises determining whether the copy of replicated group state data for the primary-backup group held by the rejoining member indicates that the rejoining member is the primary member.

3. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the existing member in the primary-backup group is valid, wherein selecting between the member and group overwrite operations comprises selecting the group overwrite operation if the copy of the group state data for the existing member is not valid.

4. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the existing member in the primary-backup group is valid, wherein selecting between the member and group overwrite operations comprises selecting the member overwrite operation if the copy of the group state data for the existing member is valid.

5. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the existing member in the primary-backup group indicates that the rejoining member is the primary member, wherein selecting between the member and group overwrite operations comprises selecting the group overwrite operation if the copy of the group state data for the existing member indicates that the rejoining member is the primary member.

6. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the existing member in the primary-backup group indicates that the rejoining member is the primary member, wherein selecting between the member and group overwrite operations comprises selecting the member overwrite operation if the copy of the group state data for the existing member indicates that the rejoining member is not the primary member.

7. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the rejoining member indicates that the rejoining member owns a resource, wherein selecting between the member and group overwrite operations comprising selecting the group overwrite operation if the copy of the group state data for the rejoining member indicates that the rejoining member owns a resource.

8. The method of claim 1, further comprising determining whether the copy of the replicated group state data for the rejoining member indicates that the rejoining member owns a resource, wherein selecting between the member and group overwrite operations comprising selecting the member overwrite operation if the copy of the group state data for the rejoining member indicates that the rejoining member does not own a resource.

9. The method of claim 1, further comprising registering the rejoining member with a cluster infrastructure.

10. The method of claim 1, further comprising sending a membership change message to the existing member of the primary-backup group.

11. The method of claim 1, further comprising sending data from the registering member'"'"'s copy of the group state data to the existing member in the primary-backup group.

12. The method of claim 1, further comprising sending data from the existing member'"'"'s copy of the group state data to the rejoining member.

13. An apparatus, comprising:

(a) a memory; and

,

(b) a program resident in the memory, the program configured to rejoin a member to a primary-backup group in a clustered computer system, wherein the primary-backup group is of the type including a primary and at least one backup member, each of which including a copy of replicated group state data for the primary-backup group, wherein the program is configured to determine whether the rejoining member is the primary member for the primary-backup group; and

, to select between member and group overwrite operations based upon such determination, wherein the member overwrite operation includes overwriting the copy of the replicated group state data for the rejoining member with data from the copy of the replicated group state data for an existing member in the primary-backup group, and wherein the group overwrite operation includes overwriting the copy of the replicated group state data for the existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member.

14. The apparatus of claim 13, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group is valid;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member is not valid.

15. The apparatus claim 13, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group indicates that the rejoining member is the primary member;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member indicates that the rejoining member is the primary member.

16. The apparatus of claim 13, wherein the program is further configured to determine whether the copy of the replicated group state data for the rejoining member indicates that the rejoining member owns a resource;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the rejoining member indicates that the rejoining member owns a resource.

17. A program product, comprising:

(a) a program configured to rejoin a member to a primary-backup group in a clustered computer system, wherein the primary-backup group is of the type including a primary and at least one backup member, each of which including a copy of replicated group state data for the primary-backup group, wherein the program is configured to determine whether the rejoining member is the primary member for the primary-backup group; and

, to select between member and group overwrite operations based upon such determination, wherein the member overwrite operation includes overwriting the copy of the replicated group state data for the rejoining member with data from the copy of the replicated group state data for an existing member in the primary-backup group, and wherein the group overwrite operation includes overwriting the copy of the replicated group state data for the existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member; and

,

(b) a signal bearing medium bearing the program.

18. The program product of claim 17, wherein the signal bearing medium includes at least one of a recordable medium and a transmission medium.

19. The program product of claim 17, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group is valid;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member is not valid.

20. The program product of claim 17, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group indicates that the rejoining member is the primary member;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member indicates that the rejoining member is the primary member.

21. The program product of claim 17, wherein the program is further configured to determine whether the copy of the replicated group state data for the rejoining member indicates that the rejoining member owns a resource;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the rejoining member indicates that the rejoining member owns a resource.

22. A clustered computer system, comprising:

(a) a plurality of nodes coupled to one another over a network;

(b) a plurality of member jobs defining a primary-backup group in the clustered computer system, and wherein the primary-backup group is of the type including a primary and at least one backup member, each of which includes a copy of replicated group state data for the primary-backup group, configured to be executed by at least one of the plurality of nodes; and

,

(c) a program configured to be executed on at least one of the plurality of nodes to rejoin a member to the primary-backup group, wherein the program is configured to determine whether the rejoining member is the primary member for the primary-backup group; and

, to select between member and group overwrite operations based upon such determination, wherein the member overwrite operation includes overwriting the copy of the replicated group state data for the rejoining member with data from the copy of the replicated group state data for an existing member in the primary-backup group, and wherein the group overwrite operation includes overwriting the copy of the replicated group state data for the existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member.

23. The cluster computer system of claim 22, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group is valid;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member is not valid.

24. The clustered computer system of claim 22, wherein the program is further configured to determine whether the copy of the replicated group state data for the existing member in the primary-backup group indicates that the rejoining member is the primary member;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the existing member indicates that the rejoining member is the primary member.

25. The clustered computer system of claim 22, wherein the program is further configured to determine whether the copy of the replicated group state data for the rejoining member indicates that the rejoining member owns a resource;

and, to select between the member and group overwrite operations by selecting the group overwrite operation if the copy of the group state data for the rejoining member indicates that the rejoining member owns a resource.

26. A method for rejoining a member to a primary-backup group in a clustered computer system, wherein the primary-backup group is of the type including a primary and at least one backup member, each of which including a copy of replicated group state data for the primary-backup group, the method comprising:

(A) overwriting the copy of the replicated group state data for an existing member in the primary-backup group with data from the copy of the replicated group state data for the rejoining member in response to at least one of;

(i) determining that the copy of the replicated group state data for the rejoining member indicates that the rejoining member is the primary member for the primary-backup group and that the copy of the replicated group state data for an existing member indicates that the rejoining member is the primary member for the primary-backup group;

(ii) determining that the copy of the replicated group state data for the rejoining member indicates that the rejoining member is the primary member for the primary-backup group, that the copy of the replicated group state data for an existing member indicates that the rejoining member is not the primary member for the primary-backup group, and that the rejoining owner owns a resource; and

(iii) determining that the copy of the replicated group state data for the rejoining member indicates that the rejoining member is not the primary member for the primary-backup group and that an existing member does not have a valid state; and

(B) overwriting the copy of the replicated group state data for the rejoining member in the primary-backup group with data from the copy of the replicated group state data for an existing member in response to at least one of;

(i) determining that the copy of the replicated group state data for the rejoining member indicates that the rejoining member is the primary member for the primary-backup group, that the copy of the replicated group state data for an existing member indicates that the rejoining member is not the primary member for the primary-backup group, and that the rejoining owner does not own a resource; and

(ii) determining that the copy of the replicated group state data for the rejoining member indicates that the rejoining member is not the primary member for the primary-backup group and that an existing member has a valid state.