- Benchmark.bonnie omits results for some targets
It is hard to predict when this will occur. Depending on which result has this issue, the parser.py may choke on the data.
There is now a spec called 'more-data', which provides more data for the machine to parse, but I suspect the issue is more complicated than just adding more files will fix. I believe it may be an issue with the Linux file cache being too big for the test, and reads coming back nearly instantaneously. (But if that were the case, what value is coming back less than .5???)
Anyway, it needs more investigation.
At a minimum the parser or criteria evaluator should not fall over when data is missing.