System and method for preserving state for a cluster of data servers in the presence of load-balancing, failover, and fail-back events

• US 20060212453A1
• Filed: 03/18/2005
• Published: 09/21/2006
• Est. Priority Date: 03/18/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Eshel, Marc M., Naik, Manoj Premanand, Gomez, Juan Carlos
• Granted Patent: US 7,962,915 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/2025 using centralised failover control functionality
• CPC Class Codes: G06F 11/2035 without idle spare hardware
• CPC Class Codes: G06F 11/2046 where the redundant components share persistent storage G06F11/2043 takes precedence
• CPC Class Codes: G06F 16/1774 Locking methods, e.g. locking methods for file systems allowing shared and concurrent access to files
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1034 Reaction to server failures by a load balancer
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• 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

A state management system preserves a state for a cluster of file servers in a cluster file system in the presence of load balancing, failover, and fail-back events. The system provides a file and record locking solution for a clustered network attached storage system running on top of a cluster file system. The system employs a lock ownership scheme in which ownership identifiers are guaranteed to be unique across clustered servers and across various protocols the clustered servers may be exporting. The system supports multi-protocol clustered NAS gateways, NAS gateway server failover and fail-back, and load-balancing architectures. The system further eliminates a need for a lock migration protocol, resulting in improved efficiency and simplicity.

Claims

1. A method for preserving a state for a cluster of data servers with a shared storage, the method comprising:

interacting with at least one client via a communication channel, to generate state information;

recognizing an event that causes a data server to fail;

establishing at least one of the data servers as a replacement server for the failed data server;

initiating a recovery of the failed data server;

providing the replacement server with an identity of the failed data server and of clients in communication with the failed data server, prior to failure;

providing the replacement server with the state information associated with the clients at the time of the data server failure;

redirecting the clients to the replacement server;

the replacement server preserving the provided state information associated with the redirected clients; and

the replacement server serving the redirected clients;

2. The method of claim 1, further comprising upon recovery of the failed server, performing failback of the clients to the recovered server.

3. The method of claim 1, further comprising load balancing a plurality of clients across available data servers while maintaining a consistent state.

4. The method of claim 3, wherein the state comprises any one of a file lock or a record lock.

5. The method of claim 2, further comprising enforcing a plurality of requests for locks across any one or more of an initiated data server failover, server failback, and load balancing, and redirecting the clients to the recovered server.

6. The method of claim 4, further comprising providing a selective grace period during the recovery of the failed server, wherein additional requests for locks that do not conflict with failed nodes locks, for files are services.

7. The method of claim 1, wherein the clients transparently interact with a plurality of servers other than the failed server.

8. The method of claim 1, wherein recognizing the event that caused the data server to fail, comprises recognizing additional successive failures of additional data servers that are managed by the shared storage data.

9. The method of claim 1, further comprising notifying the clients associated with the failed server of the recognized event.

10. The method of claim 1, wherein the shared storage data is a distributed file system.

11. The method of claim 2, wherein the client and the data servers concurrently use a plurality of protocols for communication;

and

wherein a shared state is kept consistent and enforced.

12. The method of claim 2, wherein a client registration state with the servers is maintained through failover and fail-back events.

13. The method of claim 2, wherein multiple concurrent failures of the servers are tolerated if at least one server remains operational.

14. The method of claim 3, where the clients and the server engage in asynchronous operations on the state and keep the state consistent in the face of server failures.

15. A computer program product including a plurality of executable instruction codes on a computer readable medium, for preserving a state for a cluster of data servers with a shared storage, the computer program product comprising:

a first set of instruction codes for interacting with at least one client via a communication channel, to generate state information;

a second set of instruction codes for recognizing an event that causes a data server to fail;

a third set of instruction codes for establishing at least one of the data servers as a replacement server for the failed data server;

a fourth set of instruction codes for initiating a recovery of the failed data server;

a fifth set of instruction codes for providing the replacement server with an identity of the failed data server and of clients in communication with the failed data server, prior to failure;

a sixth set of instruction codes for providing the replacement server with the state information associated with the clients at the time of the data server failure;

a seventh set of instruction codes for redirecting the clients to the replacement server;

wherein the replacement server preserves the provided state information associated with the redirected clients; and

wherein the replacement server serving the redirected clients;

16. The computer program product of claim 15, further comprising upon recovery of the failed server, an eight set of instruction code performs failback of the clients to the recovered server.

17. The computer program product of claim 15, further comprising a ninth set of instruction codes for load balancing a plurality of clients across available data servers while maintaining a consistent state.

18. A system for preserving a state for a cluster of data servers with a shared storage, the system comprising:

a data server for interacting with at least one client via a communication channel, to generate state information;

a clustering module for recognizing an event that causes a data server to fail;

the clustering module establishing at least one of the data servers as a replacement server for the failed data server;

a cluster leader initiating a recovery of the failed data server;

the cluster leader providing the replacement server with an identity of the failed data server and of clients in communication with the failed data server, prior to failure;

the cluster leader further providing the replacement server with the state information associated with the clients at the time of the data server failure;

the cluster leader redirecting the clients to the replacement server;

the replacement server preserving the provided state information associated with the redirected clients; and

the replacement server serving the redirected clients;

19. The system of claim 18, further comprising upon recovery of the failed server, the cluster leader performing failback of the clients to the recovered server.

20. The system of claim 18, wherein the state comprises any one of a file lock or a record lock.