Upgrading network traffic management devices while maintaining availability

• US 8,438,253 B2
• Filed: 05/25/2012
• Issued: 05/07/2013
• Est. Priority Date: 08/18/2008
• Status: Active Grant

• Current Assignee: F5 Networks, Inc. (F5 Incorporated)
• Original Assignee: F5 Networks, Inc. (F5 Incorporated)
• Inventors: Szabo, Paul I., Skene, Bryan D., Amdahl, Saxon
• Primary Examiner(s): Zong, Ruolei
• Application Number:
  US13/480,943
• Publication Number:
  US 20130047026A1
• Time in Patent Office: 347 Days
• Field of Search: 709220-222, 709/219, 709/225, 709/226, 709227-229, 714/1, 714 41- 45, 717168-178
• US Class Current: 709/220
• CPC Class Codes: G06F 11/203 using migration
• CPC Class Codes: G06F 11/2033 switching over of hardware resources
• CPC Class Codes: G06F 8/656 while running
• CPC Class Codes: H04L 41/0668 by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure

Abstract

A method, system, machine-readable storage medium, and apparatus are directed towards upgrading a cluster by bifurcating the cluster into two virtual clusters, an “old” virtual cluster (old active cluster) and a “new” virtual cluster (new standby cluster), and iteratively upgrading members of the old cluster while moving them into the new cluster. While members are added to the new cluster, existing connections and new connections are seamlessly processed by the old cluster. Optionally, state mirroring occurs between the old cluster and the new cluster once the number of members of the old and new clusters are approximately equal. Once a threshold number of members have been transferred to the new cluster, control and processing may be taken over by the new cluster. Transfer of control from the old cluster to the new cluster may be performed by failing over connectivity from the old cluster to the new cluster.

Claims

1. A method operated in conjunction with one or more processors to update network devices, comprising:

selecting a first member within a plurality of members in a first cluster;

bifurcating the first cluster, wherein the bifurcation comprises;

creating a second cluster, wherein the second cluster includes the selected first member;

activating the first member within the second cluster, wherein the first member is activated with a defined configuration upgrade; and

iteratively upgrading each other member in the first cluster by;

employing cluster state mirroring to move an existing connection managed by a member to be joined in the second cluster by at least one other member in the first cluster;

removing the member to be joined in the second cluster from the first cluster;

upgrading the member to be joined in the second cluster with the defined upgrade; and

joining the upgraded member as a member to the second cluster; and

determining if a defined failover criteria is satisfied, and if the defined failover criteria is satisfied, failing over connectivity from the first cluster to the second cluster such that the second cluster becomes an active cluster for managing connections.

2. The method of claim 1, wherein employing cluster state mirroring further comprises allowing an existing connections within the first cluster to drain by redirecting new connections from the first cluster to the second cluster.

3. The method of claim 1, wherein employing cluster state mirroring further comprises reducing an keep-alive value to a connection to bleed the connection from the first cluster.

4. The method of claim 1, wherein employing cluster state mirroring further includes mirroring configuration state, session state, and connection state information to the second cluster.

5. The method of claim 1, wherein employing cluster state mirroring further comprising transferring of Opens Systems Interconnection layer 4 connection data between the first cluster and the second cluster.

6. The method of claim 1, wherein employing cluster state mirroring comprises using intra-cluster and inter-cluster state mirroring.

7. The method of claim 1, wherein failing over connectivity further comprises assigning a data link layer address of the first cluster to the second cluster, such that the second cluster takes on the data link layer address of the first cluster.

8. An apparatus comprising a non-transitory computer readable medium, having computer-executable instructions stored thereon, that in response to execution by a computing device, cause the computing device to perform operations, comprising:

selecting a first member within a plurality of members in a first cluster;

bifurcating the first cluster, wherein the bifurcation comprises;

creating a second cluster, wherein the second cluster includes the selected first member;

activating the first member within the second cluster, wherein the first member is activated with a defined configuration upgrade; and

iteratively upgrading each other member in the first cluster by;

employing cluster state mirroring to move an existing connection managed by a member to be joined in the second cluster by at least one other member in the first cluster;

removing the member to be joined in the second cluster from the first cluster;

upgrading the member to be joined in the second cluster with the defined upgrade; and

joining the upgraded member as a member to the second cluster; and

determining if a defined failover criteria is satisfied, and if the defined failover criteria is satisfied, failing over connectivity from the first cluster to the second cluster such that the second cluster becomes an active cluster for managing connections.

9. The apparatus of claim 8, wherein employing cluster state mirroring further comprises allowing an existing connections within the first cluster to drain by redirecting new connections from the first cluster to the second cluster.

10. The apparatus of claim 8, wherein employing cluster state mirroring further comprises reducing an keep-alive value to a connection to bleed the connection from the first cluster.

11. The apparatus of claim 8, wherein employing cluster state mirroring further includes mirroring configuration state, session state, and connection state information to the second cluster.

12. The apparatus of claim 8, wherein failing over connectivity comprises performing an address resolution protocol.

13. The apparatus of claim 8, wherein when it is determined that the first cluster and the second cluster include a defined number of members, and inter-state mirroring of the first cluster'"'"'s connection state is at a determined level, then enabling the second cluster to enter a hot standby mode.

14. The apparatus of claim 8, wherein failing over connectivity further comprises assigning a data link layer address of the first cluster to the second cluster, such that the second cluster takes on the data link layer address of the first cluster.

15. The apparatus of claim 8, wherein employing cluster state mirroring further comprises draining existing connections in the first cluster by performing a protocol-specific action to shutdown at least one existing connection.

16. A system, comprising:

a first cluster having a plurality of members, the first cluster being an active cluster to manage network connections; and

one or more processors performing actions, including;

selecting a first member within the plurality of members in the first cluster;

bifurcating the first cluster, wherein the bifurcation comprises;

creating a second cluster, wherein the second cluster includes the selected first member;

activating the first member within the second cluster, wherein the first member is activated with a defined configuration upgrade; and

iteratively upgrading each other member in the first cluster by;

employing cluster state mirroring to move an existing connection managed by a member to be joined in the second cluster by at least one other member in the first cluster;

removing the member to be joined in the second cluster from the first cluster;

upgrading the member to be joined in the second cluster with the defined upgrade; and

joining the upgraded member as a member to the second cluster; and

determining if a defined failover criteria is satisfied, and if the defined failover criteria is satisfied, failing over connectivity from the first cluster to the second cluster such that the second cluster becomes an active cluster for managing connections.

17. The system of claim 16, wherein employing cluster state mirroring further comprises allowing an existing connections within the first cluster to drain by redirecting new connections from the first cluster to the second cluster.

18. The system of claim 16, wherein employing cluster state mirroring further comprises reducing an keep-alive value to a connection to bleed the connection from the first cluster.

19. The system of claim 16, wherein employing cluster state mirroring further includes mirroring configuration state, session state, and connection state information to the second cluster.

20. The system of claim 16, wherein state mirroring further comprising transferring of Opens Systems Interconnection layer 4 connection data between the first cluster and the second cluster.