[ovs-discuss] Issue when using local_ip with VXLAN tunnels in OVS
Siva Teja ARETI
siva_teja.areti at nuagenetworks.net
Tue Nov 6 23:08:20 UTC 2018
On Tue, Nov 6, 2018 at 5:42 PM Gregory Rose <gvrose8192 at gmail.com> wrote:
>
> On 11/6/2018 8:51 AM, Siva Teja ARETI wrote:
>
> Hi Greg,
>
> Thanks for looking into this.
>
> I have two VMs in my setup each with two interfaces. Trying to setup the
> VXLAN tunnels across these interfaces which are in different subnets. A
> docker container is attached to ovs bridge using ovs-docker utility on each
> VM and doing a ping from one container to another.
>
>
> Hi Siva,
>
> In reading through the documentation and looking at your configuration I
> noticed that when using the
> local_ip option the remote_ip is not set to flow. If the local_ip option
> is specified then remote_ip must
> equal flow.
>
> From the documentation (man ovs-vswitchd.conf.db):
>
> options : local_ip: optional string
> Optional. The tunnel destination IP that received packets
> must
> match. Default is to match all addresses. If specified,
> may be
> one of:
>
> · An IPv4/IPv6 address (not a DNS name), e.g.
> 192.168.12.3.
>
> · The word flow. The tunnel accepts packets sent to
> any of
> the local IP addresses of the system running
> OVS. To
> process only packets sent to a specific IP address,
> the
> flow entries may match on the tun_dst or
> tun_ipv6_dst
> field. When sending packets to a local_ip=flow
> tunnel,
> the flow actions may explicitly set the
> tun_src or
> tun_ipv6_src field to the desired IP address, e.g.
> with a
> set_field action. However, while routing the
> tunneled
> packet out, the local system may override the
> specified
> address with the local IP address configured for the
> out‐
> going system interface.
>
> This option is valid only for tunnels also
> configured
> with the remote_ip=flow option.
>
As I understand this documentation, option local_ip=flow can only be
specified when remote_ip=flow is also specified. Otherwise, ovs-vsctl
throws this error I guess?
[root at vm1 ~]# ovs-vsctl add-port testbr0 vxlan1 -- set interface vxlan1
type=vxlan options:local_ip=flow options:remote_ip=30.30.0.193
options:dst_port=4789
ovs-vsctl: Error detected while setting up 'vxlan1': vxlan1: vxlan type
requires 'remote_ip=flow' with 'local_ip=flow'. See ovs-vswitchd log for
details.
ovs-vsctl: The default log directory is "/var/log/openvswitch".
Please try using the remote_ip=flow option and then configuring the proper
> flow and action.
>
Anyways, I also tried by specifying remote_ip=flow option when creating the
tunnel. But I still the same issue
[root at vm1 ~]# ovs-vsctl add-port testbr0 vxlan0 -- set interface vxlan0
type=vxlan options:local_ip=20.20.0.183 options:remote_ip=flow
options:dst_port=4789
[root at vm1 ~]# ovs-appctl dpif/show
system at ovs-system: hit:69275 missed:1050
testbr0:
2cfb62a9b0f04_l 2/3: (system)
testbr0 65534/1: (internal)
vxlan0 3/2: (vxlan: local_ip=20.20.0.183, remote_ip=flow)
[root at vm1 ~]# ovs-ofctl add-flow testbr0 'table=0 priority=100 in_port=2
actions=set_tunnel:3,set_field:30.30.0.193->tun_dst,output:3'
> Thanks,
>
> - Greg
>
>
> *VM1 details:*
>
> [root at vm1 ~]# ip a
> .......
> 3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
> UP qlen 1000
> link/ether 52:54:00:b8:05:be brd ff:ff:ff:ff:ff:ff
> inet 30.30.0.59/24 brd 30.30.0.255 scope global dynamic eth1
> valid_lft 3002sec preferred_lft 3002sec
> inet6 fe80::5054:ff:feb8:5be/64 scope link
> valid_lft forever preferred_lft forever
> 4: eth2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
> UP qlen 1000
> link/ether 52:54:00:f0:64:37 brd ff:ff:ff:ff:ff:ff
> inet 20.20.0.183/24 brd 20.20.0.255 scope global dynamic eth2
> valid_lft 3248sec preferred_lft 3248sec
> inet6 fe80::5054:ff:fef0:6437/64 scope link
> valid_lft forever preferred_lft forever
> .......
> [root at vm1 ~]# ovs-vsctl show
> ff70c814-d1b0-4018-aee8-8b635187afee
> Bridge "testbr0"
> Port "gre0"
> Interface "gre0"
> type: gre
> options: {local_ip="20.20.0.183", remote_ip="30.30.0.193"}
> Port "testbr0"
> Interface "testbr0"
> type: internal
> Port "2cfb62a9b0f04_l"
> Interface "2cfb62a9b0f04_l"
> ovs_version: "2.9.2"
> [root at vm1 ~]# ip rule list
> 0: from all lookup local
> 32765: from 20.20.0.183 lookup siva
> 32766: from all lookup main
> 32767: from all lookup default
> [root at vm1 ~]# ip route show table siva
> default dev eth2 scope link src 20.20.0.183
> [root at vm1 ~]# ######################### A docker container is attached to
> ovs bridge using ovs-docker utility
> [root at vm1 ~]# docker ps
> CONTAINER ID IMAGE COMMAND CREATED
> STATUS PORTS NAMES
> be4ab434db99 busybox "sh" 5 days ago
> Up 5 days admiring_euclid
> [root at vm1 ~]# nsenter -n -t `docker inspect be4 --format={{.State.Pid}}`
> -- ip a
> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
> inet 127.0.0.1/8 scope host lo
> valid_lft forever preferred_lft forever
> inet6 ::1/128 scope host
> valid_lft forever preferred_lft forever
> 2: gre0 at NONE: <NOARP> mtu 1476 qdisc noop state DOWN qlen 1
> link/gre 0.0.0.0 brd 0.0.0.0
> 3: gretap0 at NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN
> qlen 1000
> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff
> 9: eth0 at if10: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue
> state UP qlen 1000
> link/ether 22:98:41:0f:e8:50 brd ff:ff:ff:ff:ff:ff link-netnsid 0
> inet 70.70.0.10/24 scope global eth0
> valid_lft forever preferred_lft forever
> inet6 fe80::2098:41ff:fe0f:e850/64 scope link
> valid_lft forever preferred_lft forever
>
>
> *VM2 details:*
>
> [root at vm2 ~]# ip a
> ........
> 3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
> UP qlen 1000
> link/ether 52:54:00:79:ef:92 brd ff:ff:ff:ff:ff:ff
> inet 30.30.0.193/24 brd 30.30.0.255 scope global dynamic eth1
> valid_lft 2406sec preferred_lft 2406sec
> inet6 fe80::5054:ff:fe79:ef92/64 scope link
> valid_lft forever preferred_lft forever
> 4: eth2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
> UP qlen 1000
> link/ether 52:54:00:05:93:7c brd ff:ff:ff:ff:ff:ff
> inet 20.20.0.64/24 brd 20.20.0.255 scope global dynamic eth2
> valid_lft 2775sec preferred_lft 2775sec
> inet6 fe80::5054:ff:fe05:937c/64 scope link
> valid_lft forever preferred_lft forever
> .......
> [root at vm2 ~]# ovs-vsctl show
> b85514db-3f29-4f7a-9001-37d70adfca34
> Bridge "testbr0"
> Port "gre0"
> Interface "gre0"
> type: gre
> options: {local_ip="30.30.0.193", remote_ip="20.20.0.183"}
> Port "a0769422cfc04_l"
> Interface "a0769422cfc04_l"
> Port "testbr0"
> Interface "testbr0"
> type: internal
> ovs_version: "2.9.2"
> [root at vm2 ~]# ip rule list
> 0: from all lookup local
> 32766: from all lookup main
> 32767: from all lookup default
> [root at vm2 ~]# ######################### A docker container is attached to
> ovs bridge using ovs-docker utility
> [root at vm2 ~]# docker ps
> CONTAINER ID IMAGE COMMAND CREATED
> STATUS PORTS NAMES
> 86214f0d99e8 busybox:latest "sh" 5 days ago
> Up 5 days peaceful_snyder
> [root at vm2 ~]# nsenter -n -t `docker inspect 862 --format={{.State.Pid}}`
> -- ip a
> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
> inet 127.0.0.1/8 scope host lo
> valid_lft forever preferred_lft forever
> inet6 ::1/128 scope host
> valid_lft forever preferred_lft forever
> 2: gre0 at NONE: <NOARP> mtu 1476 qdisc noop state DOWN qlen 1
> link/gre 0.0.0.0 brd 0.0.0.0
> 3: gretap0 at NONE: <BROADCAST,MULTICAST> mtu 1462 qdisc noop state DOWN
> qlen 1000
> link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff
> 9: eth0 at if10: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue
> state UP qlen 1000
> link/ether ae:ac:14:7a:40:5f brd ff:ff:ff:ff:ff:ff link-netnsid 0
> inet 70.70.0.20/24 scope global eth0
> valid_lft forever preferred_lft forever
> inet6 fe80::acac:14ff:fe7a:405f/64 scope link
> valid_lft forever preferred_lft forever
>
> With this configuration, if I do a ping from docker container on VM1 to
> docker container on VM2 it works.
>
> [root at vm1 ~]# nsenter -n -t `docker inspect be4 --format={{.State.Pid}}`
> -- ping 70.70.0.20
> PING 70.70.0.20 (70.70.0.20) 56(84) bytes of data.
> 64 bytes from 70.70.0.20: icmp_seq=1 ttl=64 time=0.831 ms
> 64 bytes from 70.70.0.20: icmp_seq=2 ttl=64 time=0.933 ms
> 64 bytes from 70.70.0.20: icmp_seq=3 ttl=64 time=0.564 ms
> ^C
> --- 70.70.0.20 ping statistics ---
> 3 packets transmitted, 3 received, 0% packet loss, time 2001ms
> rtt min/avg/max/mdev = 0.564/0.776/0.933/0.155 ms
>
> And the traffic is as expected on VM2.
>
> [root at vm2 ~]# tcpdump -n -i any host 20.20.0.183
>
>
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on any, link-type LINUX_SLL (Linux cooked), capture size 262144
> bytes
> 16:37:32.262553 IP 20.20.0.183 > 30.30.0.193: GREv0, length 102: IP
> 70.70.0.10 > 70.70.0.20: ICMP echo request, id 28158, seq 1, length 64
> 16:37:32.262835 IP 30.30.0.193 > 20.20.0.183: GREv0, length 102: IP
> 70.70.0.20 > 70.70.0.10: ICMP echo reply, id 28158, seq 1, length 64
> 16:37:33.263211 IP 20.20.0.183 > 30.30.0.193: GREv0, length 102: IP
> 70.70.0.10 > 70.70.0.20: ICMP echo request, id 28158, seq 2, length 64
> 16:37:33.263374 IP 30.30.0.193 > 20.20.0.183: GREv0, length 102: IP
> 70.70.0.20 > 70.70.0.10: ICMP echo reply, id 28158, seq 2, length 64
> 16:37:34.264159 IP 20.20.0.183 > 30.30.0.193: GREv0, length 102: IP
> 70.70.0.10 > 70.70.0.20: ICMP echo request, id 28158, seq 3, length 64
> 16:37:34.264252 IP 30.30.0.193 > 20.20.0.183: GREv0, length 102: IP
> 70.70.0.20 > 70.70.0.10: ICMP echo reply, id 28158, seq 3, length 64
>
> But when I change the tunnel type to vxlan, ping fails.
>
> [root at vm1 ~]# ovs-vsctl del-port testbr0 gre0
> [root at vm1 ~]# ovs-vsctl add-port testbr0 vxlan0 -- set interface vxlan0
> type=vxlan options:local_ip=20.20.0.183 options:remote_ip=30.30.0.193
> options:dst_port=4789
> [root at vm1 ~]# ovs-vsctl show
> ff70c814-d1b0-4018-aee8-8b635187afee
> Bridge "testbr0"
> Port "testbr0"
> Interface "testbr0"
> type: internal
> Port "vxlan0"
> Interface "vxlan0"
> type: vxlan
> options: {dst_port="4789", local_ip="20.20.0.183",
> remote_ip="30.30.0.193"}
> Port "2cfb62a9b0f04_l"
> Interface "2cfb62a9b0f04_l"
> ovs_version: "2.9.2"
>
> [root at vm2 ~]# ovs-vsctl del-port testbr0 gre0
> [root at vm2 ~]# ovs-vsctl add-port testbr0 vxlan0 -- set interface vxlan0
> type=vxlan options:local_ip=30.30.0.193 options:remote_ip=20.20.0.183
> options:dst_port=4789
> [root at vm2 ~]# ovs-vsctl show
> b85514db-3f29-4f7a-9001-37d70adfca34
> Bridge "testbr0"
> Port "a0769422cfc04_l"
> Interface "a0769422cfc04_l"
> Port "vxlan0"
> Interface "vxlan0"
> type: vxlan
> options: {dst_port="4789", local_ip="30.30.0.193",
> remote_ip="20.20.0.183"}
> Port "testbr0"
> Interface "testbr0"
> type: internal
> ovs_version: "2.9.2"
>
> Ping fails with this setup
>
> [root at vm1 ~]# nsenter -n -t `docker inspect be4 --format={{.State.Pid}}`
> -- ping 70.70.0.20
> PING 70.70.0.20 (70.70.0.20) 56(84) bytes of data.
> ^C
> --- 70.70.0.20 ping statistics ---
> 6 packets transmitted, 0 received, 100% packet loss, time 4999ms
>
> Expected traffic is not seen on VM2
>
> [root at vm2 ~]# tcpdump -n -i any host 20.20.0.183
>
>
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on any, link-type LINUX_SLL (Linux cooked), capture size 262144
> bytes
> ^C
> 0 packets captured
> 0 packets received by filter
> 0 packets dropped by kernel
>
> Kindly let me know if you need more information.
>
> Siva Teja.
>
> On Tue, Nov 6, 2018 at 10:49 AM Gregory Rose <gvrose8192 at gmail.com> wrote:
>
>>
>> On 11/5/2018 6:10 PM, Siva Teja ARETI wrote:
>>
>> Hi,
>>
>> I am trying to use local_ip option for a VXLAN tunnel using ovs but it
>> does not seem to work. The same works when I use GRE tunnel. I also found a
>> previous discussion from another user who tried the exact same approach.
>> Here is the link to the discussion
>>
>> *https://www.mail-archive.com/ovs-discuss@openvswitch.org/msg03643.html
>> <https://www.mail-archive.com/ovs-discuss@openvswitch.org/msg03643.html>*
>>
>> I am unable to find any working resolution at the end of this discussion.
>> Could you please help?
>>
>>
>> I looked into that but was never able to set up a configuration like the
>> one in that discussion and could
>> not repro the bug.
>>
>> Please provide some details on your usage, configuration and steps to
>> repro and I can look into it.
>>
>> Thanks,
>>
>> - Greg
>>
>>
>> I am using ovs 2.9.2
>>
>> [root at localhost ~]# ovs-vsctl --version
>> ovs-vsctl (Open vSwitch) 2.9.2
>> DB Schema 7.15.1
>>
>> Thanks,
>> Siva Teja.
>>
>>
>> _______________________________________________
>> discuss mailing listdiscuss at openvswitch.orghttps://mail.openvswitch.org/mailman/listinfo/ovs-discuss
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openvswitch.org/pipermail/ovs-discuss/attachments/20181106/ab757567/attachment-0001.html>
More information about the discuss
mailing list