When doing some quick tests on NVMe, I ran into the situation where bench-fio created empty csv files, apparently because --loginterval was too close to the duration of an individual benchmark trial.
When that happened, fio-plot threw an exception with the error message "statistics.StatisticsError: mean requires at least one data point" (in parse_raw_cvs_data of dataimport.py), which was not very helpful.
Perhaps it would be good to print a more helpful error message on this exception, or implement some sanity checking to avoid this situation?
When doing some quick tests on NVMe, I ran into the situation where bench-fio created empty csv files, apparently because --loginterval was too close to the duration of an individual benchmark trial.
When that happened, fio-plot threw an exception with the error message "statistics.StatisticsError: mean requires at least one data point" (in parse_raw_cvs_data of dataimport.py), which was not very helpful.
Perhaps it would be good to print a more helpful error message on this exception, or implement some sanity checking to avoid this situation?