RPS decreasing after increasing readahead size
After the readahead size has been increased, total rps that picodata can handle (number of requests per test duration) decreases. Here some data:
|readahead| rps |
|—————————|—————————————————————————————————————————————————|
|16320 |3305|2976|2903|2909|2911|2929|2861|3222|2866|3122|
|—————————|————|————|————|————|————|————|————|————|————|————|
|1024000 |2522|2384|2391|2485|2332|2153|2397|2388|2370|2372|
Prerequisites:
- Picostress
- Sceanrio.yml test2.yml
- Command: yarn build:run -e path=test2.yml -e clients=localhost:3301 -e outdir=./summary -e percerr=100 -e mode=const_rps -e duration=60s -e rps=6000 -e vu=10 -e maxvu=300