Generic resynchronization between persistent management store and dynamic configuration

• US 8,332,497 B1
• Filed: 02/20/2007
• Issued: 12/11/2012
• Est. Priority Date: 02/20/2007
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Kownacki, Ronald W., Gladish, Jacob, Comer, Michael
• Primary Examiner(s): Patel, Ashok
• Assistant Examiner(s): EDWARDS, LINGLAN E
• Application Number:
  US11/676,894
• Time in Patent Office: 2,121 Days
• Field of Search: 709/221, 709/201, 709/203, 709/217, 709/218, 709/223, 709/248, 707610-614, 707/617
• US Class Current: 709/223
• CPC Class Codes: G06F 16/1824 implemented using Network-attached Storage [NAS] architecture distributed or networked storage systems G06F3/067; protocols for distributed storage of data in a network H04L67/1097
• CPC Class Codes: G06F 16/1844 Management specifically adapted to replicated file systems
• CPC Class Codes: G06F 16/2308 Concurrency control transaction processing G06F9/466

Abstract

A cluster of nodes share configuration data that characterizes the cluster as an entity (virtualization) to a client. Management configuration data, VSC, is entered into a replicating database, RDB. That data is distributed and stored into replicating databases in each node, and then reconciled with the corresponding cluster configuration data in dynamic memory. The dynamic memory is accessed by clients. In a preferred embodiment, the nodes are storage servers.

Claims

1. A resynchronization process for a cluster of nodes interconnected over a cluster switching fabric, the resynchronization process configured to reconcile dynamic memory data in each node of the cluster with management configuration data, the resynchronization process comprising:

loading the management configuration data into a replicated database (RDB) on each node of the cluster that shares a common network address so that clients can interact with any node in the cluster, wherein each node of the cluster includes at least one of a network element and a disk element, and wherein the management configuration data is owned by the cluster of nodes as a single entity; and

in response to loading the management configuration data into the RDB on each node of the cluster, resynchronizing the dynamic memory data with the management configuration data in each RDB on each node of the cluster, wherein the management configuration data is shared among each node of the cluster, and wherein the resynchronizing comprises, for each node of the cluster;

comparing the dynamic memory data in the node of the cluster with the management configuration data loaded into the RDB on the node of the cluster, and

in response to the comparison, adding, removing or modifying the dynamic memory data in the node of the cluster so that the dynamic memory data replicates the management configuration data loaded into the RDB on the node of the cluster thus allowing a requesting client to

access any node of the cluster to obtain an up-to-date version of the management configuration data;

detecting a failure in the resynchronizing; and

disabling at least one of a client application and a protocol from using the management configuration data associated with the detected failure.

2. The process of claim 1 further comprising:

changing the management configuration data in a single node;

distributing the changed management configuration data to each other node of the cluster; and

resynchronizing the dynamic memory data of all nodes of the cluster.

3. The process of claim 1 further comprising:

structuring the management configuration data in a persistent memory as classes and instantiations therein; and

structuring the management configuration data in a dynamic memory in each node as corresponding classes and corresponding instantiations.

4. The process of claim 1, further comprising:

managing the cluster of nodes from any node of the cluster.

5. The process of claim 1 further comprising:

organizing the management configuration data and the dynamic memory data as objects.

6. The process of claim 1 further comprising storing new cluster configuration data into a RDB of a first node of the cluster, wherein contents of the RDB of the first node are distributed and loaded into the RDB in each other node of the cluster.

7. A system for resynchronizing dynamic memory data with management configuration data across a cluster of nodes interconnected over a cluster switching fabric, the system comprising:

a replicated database (RDB) on each node of the cluster containing the management configuration data that is owned by the cluster of nodes as a single entity, wherein each node of the cluster includes at least one of a network element and a disk element and wherein each node of the cluster shares a common network address so that clients can interact with any node in the cluster;

a dynamic memory in each node of the cluster, wherein the dynamic memory stores the dynamic memory data associated with at least one of client applications and protocols;

a comparator that compares the dynamic memory data in each node with the management configuration data in the RDB on each node;

a reconciliation processor that, in response to the comparison, adds, removes or modifies the dynamic memory data in each node of the cluster to match contents of the management configuration data in the RDB on each node of the cluster, thereby allowing a requesting client to access any node of the cluster to obtain an up-to-date version of the management configuration data;

a processor to detect a failure to replicate the dynamic memory data and the management configuration data; and

a first program that disables the at least one of client applications and protocols from using the management configuration data associated with the detected failure.

8. The system of claim 7 wherein the management configuration data and the dynamic memory data are objects.

9. The system of claim 7 further comprising:

an interface where an administrator enters a change into the management configuration data in the RDB of a specific node; and

a second program that distributes the change to each other node of the cluster, wherein the reconciliation processor operates in each of the nodes of the cluster to match the dynamic memory data with the management configuration data in the RDB of each of the nodes of the cluster.

10. The system of claim 7 wherein the management configuration data is held as classes and instantiations, and in the dynamic memory data has corresponding classes and corresponding instantiations therein.

11. The system of claim 7, wherein the management configuration data may be entered from any node in the cluster.

12. The system of claim 7 wherein the management configuration data and the dynamic memory data are held as objects.

13. The system of claim 7 further comprising a port in a node in the cluster, wherein the port allows new configuration data to be stored into the RDB of the node.

14. A non-transitory computer readable medium containing executable program instructions executed by a processor, comprising:

program instructions that load management configuration data into a replicated database (RDB) on each node in a cluster interconnected over a cluster switching fabric, wherein each node in the cluster includes at least one of a network element and a disk element, wherein the management configuration data is owned by the cluster as a single entity, and wherein each node in the cluster shares a common network address so that clients can interact with any node in the cluster;

program instructions that resynchronize content of a dynamic memory in each of the nodes in the cluster with the management configuration data in response to loading the management configuration data into the RDB on each node in the cluster, wherein the program instructions that resynchronize comprise;

program instructions that compare the dynamic memory content in each node in the cluster with the management configuration data loaded into the RDB on each node in the cluster;

in response to the comparison, program instructions that add, remove or modify the dynamic memory content in each node in the cluster so that the dynamic memory replicates the management configuration data loaded into the RDB on each node in the cluster, thus allowing a requesting client to access any node of the cluster to obtain an up-to-date version of the management configuration data;

program instructions that detect a failure in the resynchronizing; and

program instructions that disable at least one of a client application and a protocol from using the management configuration data associated with the detected failure.

15. A method, comprising:

maintaining a plurality of nodes interconnected as a cluster over a cluster switching fabric, wherein each node in the cluster includes at least one of a network element and a disk element, and shares a common network address so that a client can interact with any node in the cluster;

loading management configuration data into a replicated database (RDB) on each node of the cluster, wherein the management configuration data is owned by the plurality of nodes interconnected as the cluster as a single entity;

in response to loading the management configuration data into the RDB on each node of the cluster, comparing contents of a dynamic memory in each node with the management configuration data loaded into the RDB on that same node;

in response to the comparison indicating that no data is located in the contents of the dynamic memory, adding the management configuration data loaded into the RDB to the dynamic memory;

in response to the comparison resulting in no difference between the contents of the dynamic memory and the management configuration data loaded into the RDB of the node, leaving the contents of the dynamic memory as is;

in response to the comparison resulting in a difference between the contents of the dynamic memory and the management configuration data loaded into the RDB of the node, modifying the contents of the dynamic memory to be the same as the management configuration data loaded into the RDB of the node, thus allowing a requesting client to access any node of the plurality of nodes interconnected as the cluster to obtain an up-to-date version of the management configuration data;

detecting a failure in the desynchronizing; and

disabling at least one of a client application and a protocol from using the management configuration data associated with the detected failure.