If I have scripts that run docker launch commands in parallel, the docker mechanism seems to process these commands sequentially. Since starting a minimum image of a container with a โdocker stopperโ takes about 100 ms to start, does this mean that issuing commands parallel to starting 1000 containers will take the docker engine 100 ms x 1000 = 100 s or almost 2 minutes? Is there any reason why the docker engine works in parallel, rather than in parallel? How do people get around this?
source share