On Wed, Aug 14, 2013 at 03:31:44PM +1000, Fog_Watch wrote: > On Mon, 5 Aug 2013 14:52:07 +1000 > Fog_Watch <db5@exemail.com.au> wrote: >> What changes are needed to prevent a high guest CPU load, when >> combined with an IFB ingress interface, from dropping packets? > Since the last post I've successfully used cpuset to distribute my cores > so that, I think, my host has one entirely to itself: > cset: > Name CPUs-X MEMs-X Tasks Subs Path > ------------ ---------- - ------- - ----- ---- ---------- > root 0-7 y 0 y 175 4 / > as 1-4 n 0 n 61 0 /as > kdc 7 n 0 n 2 0 /kdc > voip 5 n 0 n 70 0 /voip > hf 6 n 0 n 32 0 /hf > The problem remains. > When ingress traverses an IFB, with a high host CPU load, > packets are not dropped, but when a guest CPU load is high > packets are dropped. Okay, what kernel/patch/util-vserver version? Are network related interrupts routed to the host only CPU (0 in your case)? Does it happen on a specific guest? If so, which guest load triggers it? > How do I prevent my guest-related packet drops? Ingress filtering works by dropping packets, so how do you know that the packets are not dropped because the limit was reached? best, Herbert > Regards > Fog_Watch > -- > "A. Because it breaks the logical order of conversation. > Q. Why is top posting bad?"