<div dir="ltr"><br>cc to ovs-discuss<br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">---------- Forwarded message ---------<br>发件人： <strong class="gmail_sendername" dir="auto">shuangyang qian</strong> <span dir="auto"><<a href="mailto:qsyqian@gmail.com">qsyqian@gmail.com</a>></span><br>Date: 2019年11月5日周二 下午6:12<br>Subject: Re: [ovs-discuss] the network performence is not normal when use openvswitch.ko make from ovs tree<br>To: Tonghao Zhang <<a href="mailto:xiangxia.m.yue@gmail.com">xiangxia.m.yue@gmail.com</a>><br></div><br><br><div dir="ltr">thank you for your reply, i just change my kernel version as same as you and do the steps you provide, and get the same result which i metioned at first. The process is like below.<div>on node1:</div><div># ovs-vsctl show<br>4f4b936e-ddb9-4fc6-b0aa-6eb6034d4671<br> Bridge br-int<br> Port br-int<br> Interface br-int<br> type: internal<br> Port "gnv0"<br> Interface "gnv0"<br> type: geneve<br> options: {csum="true", key="100", remote_ip="10.18.124.2"}<br> Port "veth-vm1"<br> Interface "veth-vm1"<br> ovs_version: "2.12.0"<br></div><div># ip netns exec vm1 ip a<br>1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN group default qlen 1000<br> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00<br>2: ovs-gretap0@NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN group default qlen 1000<br> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff<br>3: erspan0@NONE: <BROADCAST,MULTICAST> mtu 1450 qdisc noop state DOWN group default qlen 1000<br> link/ether 32:d9:4f:86:c3:58 brd ff:ff:ff:ff:ff:ff<br>4: ovs-ip6gre0@NONE: <NOARP> mtu 1448 qdisc noop state DOWN group default qlen 1000<br> link/gre6 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00 brd 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00<br>5: ovs-ip6tnl0@NONE: <NOARP> mtu 1452 qdisc noop state DOWN group default qlen 1000<br> link/tunnel6 :: brd ::<br>19: vm1-eth0@if18: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000<br> link/ether 32:4b:51:e2:2b:f4 brd ff:ff:ff:ff:ff:ff link-netnsid 0<br> inet <a href="http://192.168.100.10/24" target="_blank">192.168.100.10/24</a> scope global vm1-eth0<br> valid_lft forever preferred_lft forever<br> inet6 fe80::304b:51ff:fee2:2bf4/64 scope link <br> valid_lft forever preferred_lft forever<br></div><div><br></div><div>on node2:</div><div># ovs-vsctl show<br>53df6c21-c210-4c2c-a7ab-b1edb0df4a31<br> Bridge br-int<br> Port "veth-vm2"<br> Interface "veth-vm2"<br> Port "gnv0"<br> Interface "gnv0"<br> type: geneve<br> options: {csum="true", key="100", remote_ip="10.18.124.1"}<br> Port br-int<br> Interface br-int<br> type: internal<br> ovs_version: "2.12.0"<br></div><div># ip netns exec vm2 ip a<br>1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN group default qlen 1000<br> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00<br>2: ovs-gretap0@NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN group default qlen 1000<br> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff<br>3: erspan0@NONE: <BROADCAST,MULTICAST> mtu 1450 qdisc noop state DOWN group default qlen 1000<br> link/ether 8e:90:3e:95:1b:dd brd ff:ff:ff:ff:ff:ff<br>4: ovs-ip6gre0@NONE: <NOARP> mtu 1448 qdisc noop state DOWN group default qlen 1000<br> link/gre6 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00 brd 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00<br>5: ovs-ip6tnl0@NONE: <NOARP> mtu 1452 qdisc noop state DOWN group default qlen 1000<br> link/tunnel6 :: brd ::<br>11: vm2-eth0@if10: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000<br> link/ether ee:e4:3e:16:6f:66 brd ff:ff:ff:ff:ff:ff link-netnsid 0<br> inet <a href="http://192.168.100.20/24" target="_blank">192.168.100.20/24</a> scope global vm2-eth0<br> valid_lft forever preferred_lft forever<br> inet6 fe80::ece4:3eff:fe16:6f66/64 scope link <br> valid_lft forever preferred_lft forever<br></div><div><br></div><div>and in network namespace vm1 on node1 i start iperf3 as server:</div><div># ip netns exec vm1 iperf3 -s<br></div><div><br></div><div>and in network namespace vm2 on noed2 i start iperf3 as client:</div><div># ip netns exec vm2 iperf3 -c 192.168.100.10 -i 2 -t 10<br>Connecting to host 192.168.100.10, port 5201<br>[ 4] local 192.168.100.20 port 35258 connected to 192.168.100.10 port 5201<br>[ ID] Interval Transfer Bandwidth Retr Cwnd<br>[ 4] 0.00-2.00 sec 494 MBytes 2.07 Gbits/sec 151 952 KBytes <br>[ 4] 2.00-4.00 sec 582 MBytes 2.44 Gbits/sec 3 1007 KBytes <br>[ 4] 4.00-6.00 sec 639 MBytes 2.68 Gbits/sec 0 1.36 MBytes <br>[ 4] 6.00-8.00 sec 618 MBytes 2.59 Gbits/sec 0 1.64 MBytes <br>[ 4] 8.00-10.00 sec 614 MBytes 2.57 Gbits/sec 0 1.88 MBytes <br>- - - - - - - - - - - - - - - - - - - - - - - - -<br>[ ID] Interval Transfer Bandwidth Retr<br>[ 4] 0.00-10.00 sec 2.88 GBytes 2.47 Gbits/sec 154 sender<br>[ 4] 0.00-10.00 sec 2.88 GBytes 2.47 Gbits/sec receiver<br><br>iperf Done.<br></div><div><br></div><div>the openvswitch.ko in both two nodes is:</div><div># modinfo openvswitch<br>filename: /lib/modules/3.10.0-957.el7.x86_64/extra/openvswitch/openvswitch.ko<br>alias: net-pf-16-proto-16-family-ovs_ct_limit<br>alias: net-pf-16-proto-16-family-ovs_meter<br>alias: net-pf-16-proto-16-family-ovs_packet<br>alias: net-pf-16-proto-16-family-ovs_flow<br>alias: net-pf-16-proto-16-family-ovs_vport<br>alias: net-pf-16-proto-16-family-ovs_datapath<br>version: 2.12.0<br>license: GPL<br>description: Open vSwitch switching datapath<br>retpoline: Y<br>rhelversion: 7.6<br>srcversion: 764C8BD051B3182DE71CF29<br>depends: nf_conntrack,tunnel6,nf_nat,nf_defrag_ipv6,libcrc32c,nf_nat_ipv6,nf_nat_ipv4<br>vermagic: 3.10.0-957.el7.x86_64 SMP mod_unload modversions <br>parm: udp_port:Destination UDP port (ushort)<br></div><div><br></div><div>aslo, i uninstall openvswitch-kmod rpm package and use openvswitch.ko in linux kernel, like:</div><div># modinfo openvswitch<br>filename: /lib/modules/3.10.0-957.el7.x86_64/kernel/net/openvswitch/openvswitch.ko.xz<br>alias: net-pf-16-proto-16-family-ovs_packet<br>alias: net-pf-16-proto-16-family-ovs_flow<br>alias: net-pf-16-proto-16-family-ovs_vport<br>alias: net-pf-16-proto-16-family-ovs_datapath<br>license: GPL<br>description: Open vSwitch switching datapath<br>retpoline: Y<br>rhelversion: 7.6<br>srcversion: 6FE05FC439FA9CE7E264684<br>depends: nf_conntrack,nf_nat,libcrc32c,nf_nat_ipv6,nf_nat_ipv4,nf_defrag_ipv6<br>intree: Y<br>vermagic: 3.10.0-957.el7.x86_64 SMP mod_unload modversions <br>signer: CentOS Linux kernel signing key<br>sig_key: B7:0D:CF:0D:F2:D9:B7:F2:91:59:24:82:49:FD:6F:E8:7B:78:14:27<br>sig_hashalgo: sha256<br></div><div><br></div><div>and get the bw is:</div><div># ip netns exec vm2 iperf3 -c 192.168.100.10 -i 2 -t 10<br>Connecting to host 192.168.100.10, port 5201<br>[ 4] local 192.168.100.20 port 35270 connected to 192.168.100.10 port 5201<br>[ ID] Interval Transfer Bandwidth Retr Cwnd<br>[ 4] 0.00-2.00 sec 1.61 GBytes 6.92 Gbits/sec 2793 877 KBytes <br>[ 4] 2.00-4.00 sec 1.56 GBytes 6.70 Gbits/sec 7773 907 KBytes <br>[ 4] 4.00-6.00 sec 1.78 GBytes 7.62 Gbits/sec 4387 952 KBytes <br>[ 4] 6.00-8.00 sec 1.66 GBytes 7.11 Gbits/sec 9365 815 KBytes <br>[ 4] 8.00-10.00 sec 1.68 GBytes 7.20 Gbits/sec 2421 554 KBytes <br>- - - - - - - - - - - - - - - - - - - - - - - - -<br>[ ID] Interval Transfer Bandwidth Retr<br>[ 4] 0.00-10.00 sec 8.28 GBytes 7.11 Gbits/sec 26739 sender<br>[ 4] 0.00-10.00 sec 8.28 GBytes 7.11 Gbits/sec receiver<br><br>iperf Done.<br></div><div><br></div><div>so, the performence also is not normal when use openvswitch.ko installed from openvswitch-kmod rpm package.</div><div><br></div><div>can you show me your build process for openvswith-*.rpm package or give me some link? or your process of install ovs?</div><div><br></div><div>i don't know where is wrong.</div><div><br></div><div>thx.</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">Tonghao Zhang <<a href="mailto:xiangxia.m.yue@gmail.com" target="_blank">xiangxia.m.yue@gmail.com</a>> 于2019年11月5日周二 下午3:59写道：<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Mon, Nov 4, 2019 at 5:14 PM shuangyang qian <<a href="mailto:qsyqian@gmail.com" target="_blank">qsyqian@gmail.com</a>> wrote:<br>
><br>
> Hi:<br>
> I make rpm packages for ovs and ovn with this document:<a href="http://docs.openvswitch.org/en/latest/intro/install/fedora/" rel="noreferrer" target="_blank">http://docs.openvswitch.org/en/latest/intro/install/fedora/</a> . For use the kernel module in ovs tree, i configure with the command: ./configure --with-linux=/lib/modules/$(uname -r)/build .<br>
> Then install the rpm packages.<br>
> when i finished, i check the openvswitch.ko is like:<br>
> # lsmod | grep openvswitch<br>
> openvswitch 291276 0<br>
> tunnel6 3115 1 openvswitch<br>
> nf_defrag_ipv6 25957 2 nf_conntrack_ipv6,openvswitch<br>
> nf_nat_ipv6 6459 2 openvswitch,ip6table_nat<br>
> nf_nat_ipv4 6187 2 openvswitch,iptable_nat<br>
> nf_nat 18080 5 xt_nat,openvswitch,nf_nat_ipv6,nf_nat_masquerade_ipv4,nf_nat_ipv4<br>
> nf_conntrack 102766 10 ip_vs,nf_conntrack_ipv6,openvswitch,nf_conntrack_ipv4,nf_conntrack_netlink,nf_nat_ipv6,nf_nat_masquerade_ipv4,xt_conntrack,nf_nat_ipv4,nf_nat<br>
> libcrc32c 1388 3 ip_vs,openvswitch,xfs<br>
> ipv6 400397 92 ip_vs,nf_conntrack_ipv6,openvswitch,nf_defrag_ipv6,nf_nat_ipv6,bridge<br>
> # modinfo openvswitch<br>
> filename: /lib/modules/4.9.18-19080201/extra/openvswitch/openvswitch.ko<br>
> alias: net-pf-16-proto-16-family-ovs_ct_limit<br>
> alias: net-pf-16-proto-16-family-ovs_meter<br>
> alias: net-pf-16-proto-16-family-ovs_packet<br>
> alias: net-pf-16-proto-16-family-ovs_flow<br>
> alias: net-pf-16-proto-16-family-ovs_vport<br>
> alias: net-pf-16-proto-16-family-ovs_datapath<br>
> version: 2.11.2<br>
> license: GPL<br>
> description: Open vSwitch switching datapath<br>
> srcversion: 9DDA327F9DD46B9813628A4<br>
> depends: nf_conntrack,tunnel6,ipv6,nf_nat,nf_defrag_ipv6,libcrc32c,nf_nat_ipv6,nf_nat_ipv4<br>
> vermagic: 4.9.18-19080201 SMP mod_unload modversions<br>
> parm: udp_port:Destination UDP port (ushort)<br>
> # rpm -qf /lib/modules/4.9.18-19080201/extra/openvswitch/openvswitch.ko<br>
> openvswitch-kmod-2.11.2-1.el7.x86_64<br>
><br>
> Then i start to build my network structure. I have two node，and network namespace vm1 on node1, network namespace vm2 on node2. vm1's veth pair veth-vm1 is on node1's br-int. vm2's veth pair veth-vm2 is on node2's br-int. In logical layer, there is one logical switch test-subnet and two logical switch port node1 and node2 on it. like this:<br>
> # ovn-nbctl show<br>
> switch 70585c0e-3cd9-459e-9448-3c13f3c0bfa3 (test-subnet)<br>
> port node2<br>
> addresses: ["00:00:00:00:00:02 192.168.100.20"]<br>
> port node1<br>
> addresses: ["00:00:00:00:00:01 192.168.100.10"]<br>
> on node1:<br>
> # ovs-vsctl show<br>
> 5180f74a-1379-49af-b265-4403bd0d82d8<br>
> Bridge br-int<br>
> fail_mode: secure<br>
> Port "ovn-431b9e-0"<br>
> Interface "ovn-431b9e-0"<br>
> type: geneve<br>
> options: {csum="true", key=flow, remote_ip="10.18.124.2"}<br>
> Port br-int<br>
> Interface br-int<br>
> type: internal<br>
> Port "veth-vm1"<br>
> Interface "veth-vm1"<br>
> ovs_version: "2.11.2"<br>
> # ip netns exec vm1 ip a<br>
> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1<br>
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00<br>
> inet <a href="http://127.0.0.1/8" rel="noreferrer" target="_blank">127.0.0.1/8</a> scope host lo<br>
> valid_lft forever preferred_lft forever<br>
> inet6 ::1/128 scope host<br>
> valid_lft forever preferred_lft forever<br>
> 14: ovs-gretap0@NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN group default qlen 1000<br>
> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff<br>
> 15: erspan0@NONE: <BROADCAST,MULTICAST> mtu 1450 qdisc noop state DOWN group default qlen 1000<br>
> link/ether 22:02:1b:08:ec:53 brd ff:ff:ff:ff:ff:ff<br>
> 16: ovs-ip6gre0@NONE: <NOARP> mtu 1448 qdisc noop state DOWN group default qlen 1<br>
> link/gre6 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00 brd 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00<br>
> 17: ovs-ip6tnl0@NONE: <NOARP> mtu 1452 qdisc noop state DOWN group default qlen 1<br>
> link/tunnel6 :: brd ::<br>
> 18: vm1-eth0@if17: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1400 qdisc noqueue state UP group default qlen 1000<br>
> link/ether 00:00:00:00:00:01 brd ff:ff:ff:ff:ff:ff link-netnsid 0<br>
> inet <a href="http://192.168.100.10/24" rel="noreferrer" target="_blank">192.168.100.10/24</a> scope global vm1-eth0<br>
> valid_lft forever preferred_lft forever<br>
> inet6 fe80::200:ff:fe00:1/64 scope link<br>
> valid_lft forever preferred_lft forever<br>
><br>
><br>
> on node2:# ovs-vsctl show<br>
> 011332d0-78bc-47f7-be3c-fab0beb08e28<br>
> Bridge br-int<br>
> fail_mode: secure<br>
> Port br-int<br>
> Interface br-int<br>
> type: internal<br>
> Port "ovn-c655f8-0"<br>
> Interface "ovn-c655f8-0"<br>
> type: geneve<br>
> options: {csum="true", key=flow, remote_ip="10.18.124.1"}<br>
> Port "veth-vm2"<br>
> Interface "veth-vm2"<br>
> ovs_version: "2.11.2"<br>
> #ip netns exec vm2 ip a<br>
> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1<br>
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00<br>
> inet <a href="http://127.0.0.1/8" rel="noreferrer" target="_blank">127.0.0.1/8</a> scope host lo<br>
> valid_lft forever preferred_lft forever<br>
> inet6 ::1/128 scope host<br>
> valid_lft forever preferred_lft forever<br>
> 10: ovs-gretap0@NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN group default qlen 1000<br>
> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff<br>
> 11: erspan0@NONE: <BROADCAST,MULTICAST> mtu 1450 qdisc noop state DOWN group default qlen 1000<br>
> link/ether 4a:1d:ca:65:e3:ca brd ff:ff:ff:ff:ff:ff<br>
> 12: ovs-ip6gre0@NONE: <NOARP> mtu 1448 qdisc noop state DOWN group default qlen 1<br>
> link/gre6 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00 brd 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00<br>
> 13: ovs-ip6tnl0@NONE: <NOARP> mtu 1452 qdisc noop state DOWN group default qlen 1<br>
> link/tunnel6 :: brd ::<br>
> 17: vm2-eth0@if16: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1400 qdisc noqueue state UP group default qlen 1000<br>
> link/ether 00:00:00:00:00:02 brd ff:ff:ff:ff:ff:ff link-netnsid 0<br>
> inet <a href="http://192.168.100.20/24" rel="noreferrer" target="_blank">192.168.100.20/24</a> scope global vm2-eth0<br>
> valid_lft forever preferred_lft forever<br>
> inet6 fe80::200:ff:fe00:2/64 scope link<br>
> valid_lft forever preferred_lft forever<br>
><br>
> then i start to use iperf to check the network performence, oh, by the way, i use geneve protocol between the two nodes, ovn-sbctl show is :<br>
> # ovn-sbctl show<br>
> Chassis "c655f877-b7ed-4bb5-a047-23521426d541"<br>
> hostname: "<a href="http://node1.com" rel="noreferrer" target="_blank">node1.com</a>"<br>
> Encap geneve<br>
> ip: "10.18.124.1"<br>
> options: {csum="true"}<br>
> Port_Binding "node1"<br>
> Chassis "431b9efb-b464-42a1-a6dd-7fc6e0176137"<br>
> hostname: "<a href="http://node2.com" rel="noreferrer" target="_blank">node2.com</a>"<br>
> Encap geneve<br>
> ip: "10.18.124.2"<br>
> options: {csum="true"}<br>
> Port_Binding "node2"<br>
><br>
> on node1, in network namespace vm1 i start the iperf3 as server:<br>
> #ip netns exec vm1 iperf3 -s<br>
> on node2, in network namespace vm2 i start the iper3 as client:<br>
> # ip netns exec vm2 iperf3 -c 192.168.100.10<br>
> Connecting to host 192.168.100.10, port 5201<br>
> [ 4] local 192.168.100.20 port 40708 connected to 192.168.100.10 port 5201<br>
> [ ID] Interval Transfer Bandwidth Retr Cwnd<br>
> [ 4] 0.00-1.00 sec 431 MBytes 3.61 Gbits/sec 34 253 KBytes<br>
> [ 4] 1.00-2.00 sec 426 MBytes 3.58 Gbits/sec 0 253 KBytes<br>
> [ 4] 2.00-3.00 sec 426 MBytes 3.57 Gbits/sec 0 253 KBytes<br>
> [ 4] 3.00-4.00 sec 401 MBytes 3.37 Gbits/sec 0 255 KBytes<br>
> [ 4] 4.00-5.00 sec 429 MBytes 3.60 Gbits/sec 0 255 KBytes<br>
> [ 4] 5.00-6.00 sec 413 MBytes 3.46 Gbits/sec 0 253 KBytes<br>
> [ 4] 6.00-7.00 sec 409 MBytes 3.43 Gbits/sec 0 250 KBytes<br>
> [ 4] 7.00-8.00 sec 427 MBytes 3.58 Gbits/sec 0 253 KBytes<br>
> [ 4] 8.00-9.00 sec 417 MBytes 3.49 Gbits/sec 0 250 KBytes<br>
> [ 4] 9.00-10.00 sec 385 MBytes 3.23 Gbits/sec 0 5.27 KBytes<br>
> - - - - - - - - - - - - - - - - - - - - - - - - -<br>
> [ ID] Interval Transfer Bandwidth Retr<br>
> [ 4] 0.00-10.00 sec 4.07 GBytes 3.49 Gbits/sec 34 sender<br>
> [ 4] 0.00-10.00 sec 4.07 GBytes 3.49 Gbits/sec receiver<br>
><br>
> as you see, the bw is only 3.xxGbits/sec, but my physics eth1's bw 10000M:<br>
Hi, I run the ovs on node1 and node2, using geneve tunnel but I didn't<br>
reproduce your issue.<br>
<br>
create the geneve tunnel:<br>
# ovs-vsctl add-br br-int<br>
# ovs-vsctl add-port br-int gnv0 -- set Interface gnv0 type=geneve<br>
options:csum=true options:key=100 options:remote_ip=1.1.1.200<br>
<br>
# ovs-vsctl show<br>
9393485c-c64c-490e-884e-418ff5d90251<br>
Bridge br-int<br>
Port gnv0<br>
Interface gnv0<br>
type: geneve<br>
options: {csum="true", key="100", remote_ip="1.1.1.200"}<br>
Port __tap01<br>
Interface __tap01<br>
Port br-int<br>
Interface br-int<br>
type: internal<br>
<br>
# ip netns exec ns100 ifconfig<br>
__tap00: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1450<br>
inet 2.2.2.100 netmask 255.255.255.0 broadcast 0.0.0.0<br>
inet6 fe80::254:ff:fe00:1 prefixlen 64 scopeid 0x20<link><br>
ether 00:54:00:00:00:01 txqueuelen 1000 (Ethernet)<br>
RX packets 605000 bytes 39951500 (38.1 MiB)<br>
RX errors 0 dropped 0 overruns 0 frame 0<br>
TX packets 819864 bytes 31247862764 (29.1 GiB)<br>
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0<br>
<br>
# ip netns exec ns100 iperf -c 2.2.2.200 -i 2 -t 10<br>
------------------------------------------------------------<br>
Client connecting to 2.2.2.200, TCP port 5001<br>
TCP window size: 482 KByte (default)<br>
------------------------------------------------------------<br>
[ 3] local 2.2.2.100 port 41428 connected with 2.2.2.200 port 5001<br>
[ ID] Interval Transfer Bandwidth<br>
[ 3] 0.0- 2.0 sec 1.85 GBytes 7.93 Gbits/sec<br>
[ 3] 2.0- 4.0 sec 1.94 GBytes 8.33 Gbits/sec<br>
<br>
# modinfo openvswitch<br>
filename: /lib/modules/3.10.0-957.1.3.el7.x86_64/extra/openvswitch.ko<br>
<br>
so, can you use the commands as show above to reproduce your<br>
issue.(the kernel version is different.)<br>
> # ethtool eth1<br>
> Settings for eth1:<br>
> Supported ports: [ FIBRE ]<br>
> Supported link modes: 10000baseT/Full<br>
> Supported pause frame use: Symmetric<br>
> Supports auto-negotiation: No<br>
> Supported FEC modes: Not reported<br>
> Advertised link modes: 10000baseT/Full<br>
> Advertised pause frame use: Symmetric<br>
> Advertised auto-negotiation: No<br>
> Advertised FEC modes: Not reported<br>
> Speed: 10000Mb/s<br>
> Duplex: Full<br>
> Port: Other<br>
> PHYAD: 0<br>
> Transceiver: external<br>
> Auto-negotiation: off<br>
> Supports Wake-on: d<br>
> Wake-on: d<br>
> Current message level: 0x00000007 (7)<br>
> drv probe link<br>
> Link detected: yes<br>
><br>
> when i uninstall the openvswitch-kmod package and use the openvswitch.ko in the upstream linux kernel, like this:<br>
> # lsmod | grep openvswitch<br>
> openvswitch 95805 0<br>
> nf_defrag_ipv6 25957 2 nf_conntrack_ipv6,openvswitch<br>
> nf_nat_ipv6 6459 2 openvswitch,ip6table_nat<br>
> nf_nat_ipv4 6187 2 openvswitch,iptable_nat<br>
> nf_nat 18080 5 xt_nat,openvswitch,nf_nat_ipv6,nf_nat_masquerade_ipv4,nf_nat_ipv4<br>
> nf_conntrack 102766 10 ip_vs,nf_conntrack_ipv6,openvswitch,nf_conntrack_ipv4,nf_conntrack_netlink,nf_nat_ipv6,nf_nat_masquerade_ipv4,xt_conntrack,nf_nat_ipv4,nf_nat<br>
> libcrc32c 1388 3 ip_vs,openvswitch,xfs<br>
> # modinfo openvswitch<br>
> filename: /lib/modules/4.9.18-19080201/kernel/net/openvswitch/openvswitch.ko<br>
> alias: net-pf-16-proto-16-family-ovs_packet<br>
> alias: net-pf-16-proto-16-family-ovs_flow<br>
> alias: net-pf-16-proto-16-family-ovs_vport<br>
> alias: net-pf-16-proto-16-family-ovs_datapath<br>
> license: GPL<br>
> description: Open vSwitch switching datapath<br>
> srcversion: 915B872C96FB1D38D107742<br>
> depends: nf_conntrack,nf_nat,libcrc32c,nf_nat_ipv6,nf_nat_ipv4,nf_defrag_ipv6<br>
> intree: Y<br>
> vermagic: 4.9.18-19080201 SMP mod_unload modversions<br>
><br>
> and do the same test in above, and i get the follow result:<br>
> # ip netns exec vm2 iperf3 -c 192.168.100.10<br>
> Connecting to host 192.168.100.10, port 5201<br>
> [ 4] local 192.168.100.20 port 40652 connected to 192.168.100.10 port 5201<br>
> [ ID] Interval Transfer Bandwidth Retr Cwnd<br>
> [ 4] 0.00-1.00 sec 1000 MBytes 8.39 Gbits/sec 4 290 KBytes<br>
> [ 4] 1.00-2.00 sec 994 MBytes 8.34 Gbits/sec 0 292 KBytes<br>
> [ 4] 2.00-3.00 sec 1002 MBytes 8.41 Gbits/sec 0 287 KBytes<br>
> 4] 3.00-4.00 sec 994 MBytes 8.34 Gbits/sec 0 292 KBytes<br>
> ▽ 4] 4.00-5.00 sec 992 MBytes 8.32 Gbits/sec 0 298 KBytes<br>
> [ 4] 5.00-6.00 sec 994 MBytes 8.34 Gbits/sec 0 305 KBytes<br>
> [ 4] 6.00-7.00 sec 989 MBytes 8.29 Gbits/sec 0 313 KBytes<br>
> [ 4] 7.00-8.00 sec 992 MBytes 8.32 Gbits/sec 0 290 KBytes<br>
> [ 4] 8.00-9.00 sec 996 MBytes 8.36 Gbits/sec 0 303 KBytes<br>
> [ 4] 9.00-10.00 sec 955 MBytes 8.01 Gbits/sec 0 5.27 KBytes<br>
> - - - - - - - - - - - - - - - - - - - - - - - - -<br>
> [ ID] Interval Transfer Bandwidth Retr<br>
> [ 4] 0.00-10.00 sec 9.67 GBytes 8.31 Gbits/sec 4 sender<br>
> [ 4] 0.00-10.00 sec 9.67 GBytes 8.31 Gbits/sec receiver<br>
><br>
> so, i cann't understand why the performence is so poor when i use the kernel module build from the ovs tree.<br>
><br>
> anyone can give me some advice where is wrong?<br>
><br>
> thx!<br>
><br>
> _______________________________________________<br>
> discuss mailing list<br>
> <a href="mailto:discuss@openvswitch.org" target="_blank">discuss@openvswitch.org</a><br>
> <a href="https://mail.openvswitch.org/mailman/listinfo/ovs-discuss" rel="noreferrer" target="_blank">https://mail.openvswitch.org/mailman/listinfo/ovs-discuss</a><br>
</blockquote></div>
</div></div>