Abstract: Method embodiments for triggering error injection into a function under test using a serialization resource are provided. A test process invokes the function under test immediately after relinquishing exclusive control of the serialization resource. An error-injection process injects the error into the running function after gaining exclusive control of the serialization resource from the test process. The error-injection process may add a delay to inject the error. If the processes are repeated, the error-injection process may vary the delay, perhaps randomly, over a specified time window to thoroughly exercise the function's error recovery routine.

Abstract: Generation of diagnostic information of a computer-implemented system is made early so that the data is closer to the causation of errors or for performance analysis. At least one selected activity of the system is monitored from initiation of the activity, and the monitoring is for successful completion. Early collection of diagnostic information is provided by comparing the time of the activity without successful completion to an initial trigger, where the initial trigger is less than the time period for a time-out for the activity. If the time of the activity without successful completion exceeds the initial trigger, diagnostic information is collected and an initial dump of the diagnostic information is taken. In one example, a notification that the dump of diagnostic information has been taken is directed to the host or diagnostic terminal.

Abstract: Method, system, and computer program product embodiments of a time-based trace facility for facilitating software debugging without interfering with the run-time behavior, performance or resource usage of the traced software are provided. The trace facility resides in a different address space than the target address space and uses different time-slices of CPU resources to execute the instructions in the address spaces. The trace facility uses a cross-memory mode to read the state data from the target address space in accordance with a time schedule. The trace facility writes the state data to a trace area, which may be located in either the target or trace address spaces or external storage. With this approach, the trace facility can read a large amount of state data frequently to construct the type of historical record needed to analysis run-time behavior, performance and resource usage.