Use FIO on different devices and collect the FIO results to OUTPUT
DIR=/data1 OUTPUT=./log1 ./run_fio.sh
DIR=/data2 OUTPUT=./log2 ./run_fio.sh
python parse_latency.py -i ./log1/com -i ./log1/com
It will generate following files in the directory './output'
read100-lat-mean.csv
read100-lat-p99.csv
read70-lat-mean.csv
read70-lat-p99.csv
write100-lat-mean.csv
write100-lat-p99.csv
write30-lat-mean.csv
write30-lat-p99.csv
read100
means 100% read, read70
means 70% read + 30% write, and so on.
python parse_bw.py -i ./log1/com -i ./log1/com
It will generate following files in the directory './output'
read100-bandwidth.csv
read70-bandwidth.csv
write100-bandwidth.csv
write30-bandwidth.csv
python parse_var_bs.py -i ./log1/var-bs -i ./log2/var-bs
It will generate following files in the directory './output'
read100-bs-lat-mean.csv
read100-bs-lat-p99.csv
You can use ./gen_output.sh
to generate all the CSV files directly, like:
./gen_output.sh ./log1 ./log2
After you generate all CSV files, you can use many tools to convert them to Markdown table, or paste them to the Excel.
- Use pyplot to generate the chart