MicroRacer

This repository contains the source code for MicroRacer.

Project Structure

.
├── auto-trace-query            # Scripts for automatic export of system traces
├── data                        # Directory for raw data files
├── output                      # Directory for output results
├── train-ticket-workload       # TrainTicket workload
├── utils                       # Utility methods
├── ./reqflow_analyze.py        # Generation of conflicting request pairs
├── ./reqflow_construct.py      # Request flow construction
├── ./interleaving_testing.py   # Interleaving testing
├── ./main.py                   # Main program
├── ./object.py                 # Class definitions for request flow construction
├── ./trace_preprocess.py       # Preprocessing of trace data
├── ./requirements.txt
└── ./TrainTicket_Faults.md     # Description of reproduced faults

Reproduction Steps

This repository provides the trace files collected from TrainTicket (in five versions of fault reproduction) generated by the APM tool SkyWalking. You can follow the steps to construct request flows, generate potential conflict request pairs, and perform early validation and automatic interleaving testing.

1. Set up the environment: pip install -r requirements.txt

2. Construct request flows and generate potential conflicting request pairs:

   • Set the trace_dir and output_dir paths in the ./main.py file (located at the bottom of the file) to the paths of the raw Trace files and the output results, respectively.
   • Run python main.py to construct request flows and generate potential conflicting request pairs.
   • You can find the JSON files starting with candidatePairs_ in the output directory, which will be used for subsequent automatic interleaving testing (case 4 does not have this file because the pruning result is such that the problematic request pairs are reported in early validation).
   • Optional: Print debug logs if the environment variable DEBUG=1.

3. Deploy TrainTicket (see Train-Ticket-Benchmark/README.md)

4. Perform automatic interleaving testing:

   • Edit db_helper.py to set up your database connection details:

     DB_HOST = 'localhost'
     DB_USER = 'root'
     DB_PASS = 'root'
     DB_NAME = 'your_database_name'

   • Configure the paths and other variables in interleaving_testing.py:

     json_file_path = "path_to_candidatePairs_file"
     backup_dir = "path_to_backup_directory"
     restore_dir = "path_to_restore_directory"

     The testing results, including logs and any detected inconsistencies, will be saved in the specified backup_dir and response_log_path.

   • Backup Database: Run db_helper.py to create a backup of your current database state.

     python db_helper.py

   • Interleaving Testing: Execute interleaving_testing.py to run tests on the specified request pairs.

     python interleaving_testing.py

     The system will read from a JSON file containing the pairs of requests to be tested. It will send requests to the specified services, perform database backups and restorations in between, and log the responses and service logs for comparison.

5. Output

   • The testing results, including logs and any detected inconsistencies, will be saved in the specified backup_dir and response_log_path.