Method and apparatus for handling failures of resource managers in a clustered environment

Abstract

The present invention provides a system and method for committing a transaction. Briefly described, one embodiment is a method comprises assigning a resource manager a globally unique identifier (GUID), the resource manager assigned to complete the transaction; detecting a failure such that the resource manager cannot complete the transaction; and assigning a backup resource manager the GUID when the failure is detected such that the transaction is completed by the backup resource manager.

What is claimed is:  
     
         1 . A method for committing a transaction, the method comprising the steps of: 
 assigning a resource manager a globally unique identifier (GUID), the resource manager assigned to complete the transaction;    detecting a failure such that the resource manager cannot complete the transaction; and    assigning a backup resource manager the GUID when the failure is detected such that the transaction is completed by the backup resource manager.    
     
     
         2 . The method of  claim 1 , further comprising the steps of: 
 associating a beginner transaction manager facility (TMF) with the resource manager;    broadcasting a prepare signal to a plurality of other TMFs that each receive the prepare signal;    communicating a ready signal from each of the other TMFs back to the beginner TMF; and    broadcasting a commit signal,    wherein the resource manager is replaced by the backup resource manager when the detected failure occurs after the ready signals are communicated and before the commit signal is broadcasted.    
     
     
         3 . The method of  claim 2 , wherein the step of broadcasting the commit signal further comprises the step of broadcasting the commit signal by the backup resource manager.  
     
     
         4 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of the resource manager.  
     
     
         5 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of a component associated with the resource manager.  
     
     
         6 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of another node associated with the resource manager.  
     
     
         7 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of communication from the resource manager.  
     
     
         8 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of a processor unit associated with the resource manager.  
     
     
         9 . The method of  claim 1 , wherein the step of detecting the failure further comprises the step of detecting loss of a process associated with the resource manager.  
     
     
         10 . A transaction processing system which commits a transaction comprising: 
 a resource manager having an assigned globally unique identifier (GUID) and assigned to complete the transaction; and    a backup resource manager that is assigned the GUID when the resource manager is lost such that the transaction is completed by the backup resource manager.    
     
     
         11 . The transaction processing system of  claim 10 , wherein the transaction is performed using a two-phase transaction protocol.  
     
     
         12 . The transaction processing system of  claim 10 , further comprising; 
 a beginner transaction manager facility (TMF), associated with the resource manager, that causes a prepare signal to be broadcasted; and    a plurality of other TMFs that each receive the prepare signal and each communicate a ready signal back to the beginner TMF, wherein the resource manager is replaced by the backup resource manager when the resource manager is lost after the communicated ready signals and before a commit signal is received.    
     
     
         13 . A system for committing a two-phase protocol transaction, comprising: 
 means for assigning a resource manager a globally unique identifier (GUID), the resource manager assigned to complete the two-phase protocol transaction;    means for detecting a failure such that the resource manager cannot complete the two-phase protocol transaction; and    means for assigning a backup resource manager the GUID when the failure is detected such that the two-phase protocol transaction is completed by the backup resource manager.    
     
     
         14 . A transaction processing system which commits a two-phase protocol transaction comprising: 
 means for specifying a beginner resource manager having an assigned globally unique identifier (GUID) and assigned to complete the two-phase protocol transaction; and    means for specifying a backup resource manager that is assigned the GUID when the beginner resource manager is lost such that the two-phase protocol transaction is completed by the backup resource manager.