Skip to content

Latest commit

 

History

History
29 lines (21 loc) · 1.52 KB

README.md

File metadata and controls

29 lines (21 loc) · 1.52 KB
Raw

Log Processor for Pravega System Tests

This script is a utility for ease the debugging of Pravega system tests. The idea of this script is to generate a single file that contains summaries of error messages from the test driver log, the controller log and the segment store log in parallel, thus helping the developer to understand the sequence of errors across multiple subsystems. By default, the time frame granularity is 1 second. Note that multiple errors may be displayed in a single time frame. Moreover, the script also takes care of extracting all the log files which are partitioned and compressed.

How To

The script is simple to use:

  • Download all the *.tar.gz files from the Jenkins build you want to debug and store them in my_log_dir.
  • Modify the constant LOGS_DIR in the script to point to my_log_dir.
  • Execute python processor.py.

*Note for Windows users: Due to how log files are packaged, unziped files may have path lengths larger than the max path length in Windows.

Customization

There are several parameter that can be customized in this script:

  • MAX_PREVIOUS_LINES: Lines of logs to be included in the error summary that precede the actual error (good for getting some context about what is going on).
  • MAX_TRACE_LINES: Length of the error trace to be included in the error summary.
  • ERROR_KEYWORDS_TO_CATCH: The script looks for one of the keywords in this list to label a line as the start of an error to be included in the final output.
  • OUTPUT_FILE: Name of the output file to be generated.