Hi ,
I have the following setup :
ovs built with dpdk
root@ubuntu:/home/ubuntu# ovs-vsctl --version
ovs-vsctl (Open vSwitch) 2.9.0
DB Schema 7.15.1
running on ubuntu 17 server edition,
I am trying to configure the meters as follows,
root@ubuntu:~# ovs-ofctl add-meter br-voss
"meter=2,kbps,stats,band=type=drop,rate=43000" -OOpenflow15
root@ubuntu:~# ovs-ofctl add-flow br-voss
"priority=10000,dl_vlan=400,actions=meter:2,output:2" -OOpenflow15
My expectation is that, traffic in excess of 43Mbps defined by the flows be
dropped.
However what i am observing is that
Before adding meters and flow rules
port "vhost-2": rx pkts=110, bytes=8080, drop=0, errs=0, frame=?, over=?,
crc=?
tx pkts=232930268, bytes=28883353232, drop=9628, errs=?, coll=?
After few seconds.
port "vhost-2": rx pkts=110, bytes=8080, drop=0, errs=0, frame=?, over=?,
crc=?
tx pkts=232886922, bytes=28877978328, drop=9628, errs=?, coll=?
As you can see the meter has allowed (28877978328-28883353232)*8 bits which
is
42999232 or roughly 43Mb.
After this i do not see the tx counters for vhost-2 ports getting
increased, It looks like the meter is allowing 43Mb of traffic overall and
dropping rest of the traffic thereafter.
My understanding is that meter should be able to , control the ingress rate
of traffic. Is this correct?
Am i doing something wrong here? the effect is similar with the burst_size
specified.
_______________________________________________
discuss mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
