Fault-tolerant transaction processing in a distributed system using explicit resource information for fault determination

A method of tracking and controlling data within a system, said system including at least two system CPUs, comprising the steps of:

• receiving a transaction in one of said two system CPUs, said transaction being capable of altering said data in said system;
• associating with said transaction a control data structure;
• sending at least one request associated with said transaction to another of said two system CPUs;
• indicating in the control data structure that an other one of said two system CPUs is participating in the transaction;
• detecting failure of said other system CPU;
• and
• examining the control data structure for the indication, and if found deciding whether to abort the transaction.

A process of tracking and controlling data in a distributed a computing system using explicit resource information for fault determination, the computing system including a plurality of communicatively intercoupled system processor units each executing an associated transaction manager facility, the process comprising the steps of:

• starting a transaction at a beginner one of the processor units;
• registering said transaction via a start transaction call to a beginner transaction manager facility executed by said beginner processor unit;
• creating an information data structure associated with said transaction;
• sending at least one request associated with said transaction to at least one resource executed by an associated working one of the processor units, said request indicating that said beginner processor unit is associated with said transaction;
• notifying the transaction manager facility at each said working processor unit that said working processor unit hosts a resource that is participating in said transaction;
• notifying said beginner transaction manager facility that said working processor unit hosts a resource that is participating in said transaction;
• writing transaction control information to said data structure, said control information indicating that each said working processor unit is a participant in said transaction;
• detecting a failed one of the working processor units;
• examining the information data structure to determine if said failed working processor unit is a participant in said transaction;
• and
• if said failed working processor unit is a participant in said transaction, determining whether to abort said transaction.