Executive Summary
The Fault Replication System prepares and executes error replay on the Maintenance (Server) environment.
Services Provided
- Fault replication: Given a ticket ID from the TRAC server, the fault replication system will replay that error report. It will check if the ticket is replayable and if so start the step by step replay. This will display the execution of the application stopping and each user input point. It will also provide a textual display of context information that was recorded during the app execution. In a more advanced stage, the system may connect to eclipse and synchronize the application code window with the error replay and allow for other stopping points in the step by step execution besides the user input points.
Services Needed
- From ???:
- Context Information (context events and related faults) relevant to the particular fault/ticket being replayed.
Open Questions and Doubts
- Context augmented debugging
How are context data collected by the Context System presented to the user?
How is this realized? Where does the video/ view of UI at a certain point in time come from (is it generated, captured, ...)?
- User context and distributed logging
How do these concepts go together? Do they?
- It is unclear to us at INESC where the relevant context information (context events and related faults) will be fetched from. Is it from the Data Store' Is it a log-like file that was transferred at error reporting time?
Shared Data Structures
- Error Report: Description of a fault including: app, relevant configuration and relevant log files.
- Context Event: An event with a textual description and a timestamp/process id that can be correlated to a ticket replay.
- Related Fault: The id of a fault, in this case that is related to the one being replayed.
Component-specific Errors
see [Component Specific Error Types]
Limitations
see [Component Limitations]