I am proof of concept for a client, the requirements are as follows:
Stretch the ERP (Extended VLAN) network in 2 places, this is necessary due to outdated machinery that has to communicate with a particular IP address; We cannot change the direction of the ERP HA since the industrial machinery has a hard code.
What I have is the following configuration:
2 Ubuntu server 16.04 3 NIC for each OVS server version 2.5.5
They are respectively named OVS1 and OVS2. The server network has 1 NIC for administration (those configuration identities are reflected in the two servers, the IP addressing is different) 1 NIC is assigned to an OVS-bridge AB100, a physical port is assigned to this the bridge NIC 1 is assigned to an OUTBOUND of the OVS bridge, a physical port is assigned to this bridge
Bridges are addresses and settings as follows on each server:
On OVS1 Bridge AB100 it has an IP address of 10.1.1.201 On OVS1 Bridge AB100 it has a VXLAN100 with remote IP 172.29.0.2 On OVS1 Bridge AB100 it has GRE0 with remote IP 172.29.0.4 On OVS1 Bridge OUTBOUND it has an IP address of 172.24.0.1
On OVS2 Bridge AB100 it has an IP address of 10.1.1.202 On OVS1 Bridge AB100 it has a VXLAN100 with remote IP 172.29.0.1 On OVS1 Bridge AB100 it has GRE0 with remote IP 172.29.0.3 On OVS1 Bridge OUTBOUND it has an IP address of 172.24.0.2
OVS1 interface eth0 is for management OVS2 interface eth0 is for management OVS1 interface eth1 is bridged AB100 OVS2 interface eth1 is bridged AB100 OVS1 interface eth2 is bridged OUT OVS2 interface eth2 is bridged OUT
OVS1 interface eth0, management on physical management switch1 OVS2 interface eth0, management on physical management switch1 OVS1 interface eth1, jumper AB100 on physical management switch2 OVS2 interface eth1, jumper AB100 on physical management switch2 OVS1 interface eth2, OUTBOUND jumper on physical management switch3 OVS2 eth2 interface, OUTBOUND jumper on physical management switch3
The eth0 interface is not part of any OVS switch, it is currently working and it is the medium where I can use SSH and manage the devices.
What is connected to them:
OVS1 eth1 interface, AB100 jumper on physical management switch2, Mint desktop with IP 10.1.1.2 OVS2 eth1 interface, AB100 jumper on physical management switch2, Fedora desktop with IP 10.1.1.3 Secondary IP address 10.1.1.254, which will vary depending What the asset is, but for this example, leave it on 10.1.1.2 One of the devices will have a secondary IP address of 10.1.1.254, which is the IP address to which all the factory devices connect.
The connectivity of the two desktops to the local AB100 bridge is successful and they can ping each other. Ping one of IP 172.29.0.1 or 172.29.0.2 respectively reverse unsuccessful Ping one if IP 172.24.0.1 or 172.24.0.2 respectively from its own OVS is unsuccessful
Looking at the ARP-a you can see the IP address but not the MAC address of the respective OVS from the OUTGOING Bridge Looking at the ovs-appctl fdb / show OUTBOUND you can see the MAC address
The result should be that the two desktops should be able to ping each other through the VXLAN or GRE tunnel, but I am at the stage where there is no basic connectivity.
Any help and suggestion would be appreciated.