[vserver] How do I stop a high guest CPU load and ifb0 from dropping internet packets?

From: Fog_Watch <db5_at_exemail.com.au>
Date: Mon 05 Aug 2013 - 05:52:07 BST
Message-ID: <20130805145207.35db591c@mr-greedy-as.fogwatch.com.au>

Hello

A high cpu load from my guests causes internet packets to be dropped.
Please may I have some help fixing this problem.

When guest CPU load is high and internet ingress comes from an
intermediate functional block (IFB), care of Shorewall, packets are
dropped.

For example, when I
$ ping -i .5 <isp name>

And when the 2x4 CPU cores are occupied:
<hostname> ~ # date && netstat -i | head -n 3
Mon Aug 5 14:22:34 EST 2013
Kernel Interface table
Iface MTU RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
TX-OVR Flg
ifb0 1500 319098 0 0 0 319098 0
2248 0 BORU
<hostname> ~ # date && netstat -i | head -n 3
Mon Aug 5 14:22:39 EST 2013
Kernel Interface table
Iface MTU RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
TX-OVR Flg
ifb0 1500 319118 0 0 0 319118 0
2268 0 BORU

That is under high CPU load internet traffic basically stops.

When guest CPU load is high and ingress comes from a normal PPP
interface packets are not dropped.

Finally, when CPU load is high from the host and ingress comes from
ifb0 packets are not dropped.

What changes are needed to prevent a high guest CPU load, when
combined with an IFB ingress interface, from dropping packets?

Regards

Fog_Watch

uname -a
Linux <hostname> 3.9.5-vs2.3.6.5 #8 SMP Mon Jul 15 17:16:49 EST 2013
x86_64 Intel(R) Xeon(R) CPU E5345 @ 2.33GHz GenuineIntel GNU/Linux

-- 
"A. Because it breaks the logical order of conversation.
 Q. Why is top posting bad?"
Received on Mon Aug 5 05:52:19 2013
[Next/Previous Months] [Main vserver Project Homepage] [Howto Subscribe/Unsubscribe] [Paul Sladen's vserver stuff]
Generated on Mon 05 Aug 2013 - 05:52:19 BST by hypermail 2.1.8