3 ms·

Many people in the Node.js community have had first-hand experience with with this issue. Node.js has had to turn off OS-based scheduling for its cluster module

Many people in the Node.js community have had first-hand experience with with this issue. Node.js has had to turn off OS-based scheduling for its cluster module because you always ended up with a couple of CPUs taking all the load (and accepting new connections) while most of them remained idle. I thought it was either a bug with the TCP polling mechanism or with the Linux OS scheduler itself. It's good that this issue is finally getting some attention.

melchebo 11y ago
That might be what they call the 'overload on wakeup' bug. Maybe try the patch? I've read that the patch additionally needs a small fix, the goto label position went missing. Probably just before rcu_read_unlock() in that function: http://lxr.free-electrons.com/source/kernel/sched/fair.c#L5139 http://lxr.free-electrons.com/source/kernel/sched/fair.c#L51...