Is there currently a way to limit
cpuset available to a Docker Cloud service not just
I have services that automatically employ all CPU’s available.
I’m running 20 core machines, and, correct me if I’m wrong, but I believe it would be better to have 4 instances that each get a
cpuset of 5 cores rather than 4 instances that each run on all 20 cores with
cpushares rationing off processor %.
- Maybe performance would balance out either way…haven’t tested my theories.
- I could always limit this in the application, but prefer configuration over code.