On Thu, Nov 19, 2020 at 4:48 PM Brendan Doyle <brendan.do...@oracle.com> wrote:
>
> So It works if I manually configure the br-int MTUs:
>
> #ovs-vsctl set interface vnet1 mtu_request=8900
> #ovs-vsctl set interface vnet1 mtu=8900
> #ovs-vsctl set interface vnet2 mtu_request=8900
> #ovs-vsctl set interface vnet2 mtu=8900#ovs-vsctl get interface vnet2 mtu
> 8900
> # ovs-vsctl get interface vnet1 mtu
> 8900
>
>
> But I'm thinking that I should be able to modify the VM KVM XLM, to specify
> an MTU for the openvswitch interface, something like:
>
> <interface type='bridge'>
> <mac address='00:13:97:b6:0f:3c'/>
> <source bridge='br-int'/>
> <virtualport type='openvswitch'>
> <parameters interfaceid='e2ce9e04-25f3-11eb-9180-0010e0c3df92'/>
> </virtualport>
> <model type='virtio'/>
> <address type='pci' domain='0x0000' bus='0x00' slot='0x03'
> function='0x0'/>
> <mtu size='8900'/>
> </interface>
>
>
> Then hopefully the libvert plugin will configure the OVS bridge MTUs to 8900
I think that's what Openstack Nova or any other compute service
would/should do when it boots a VM.
Numan
>
> Brendan
>
> On 19/11/2020 10:53, Brendan Doyle wrote:
>
> Folks,
>
> I'm trying to get OVN working in a 9000 MTU config so :
>
> ovn-nbctl list dhcp-option
> _uuid : 00f684c8-42bb-4fad-a248-f5697926778d
> cidr : "10.225.1.0/24"
> external_ids : {}
> options : {lease_time="3600", mtu="9000", router="10.225.1.1",
> server_id="10.225.1.1", server_mac="00:13:97:60:de:6a"}
>
> The hypervisor phys interface MTU is 9000, so I know the above should specify
> MTU of 8950 to account
> for the geneve header, in any case I manually lower the MTU on the VM node
> but things still don't work.
> And I think the reason is that when I looked at the OVS br-int, I see the OVN
> port MTU is 1500. How
> do I get OVN to create the br-int port with a 9000/8950 MTU?
>
> Is this something we have to manually configure on the OVS bridge?
>
> ovs-vsctl list interface vnet1
> _uuid : 49637687-3870-4eea-911a-e22c6c2ef19a
> admin_state : up
> bfd : {}
> bfd_status : {}
> cfm_fault : []
> cfm_fault_status : []
> cfm_flap_count : []
> cfm_health : []
> cfm_mpid : []
> cfm_remote_mpids : []
> cfm_remote_opstate : []
> duplex : full
> error : []
> external_ids : {attached-mac="00:13:97:32:18:d6",
> iface-id="dc5bca80-2948-11eb-96a2-0010e0c3df92", iface-status=active,
> vm-id="11b14a04-59ed-45e9-9445-665b8847fba7"}
> ifindex : 26
> ingress_policing_burst: 0
> ingress_policing_rate: 0
> lacp_current : []
> link_resets : 1
> link_speed : 10000000
> link_state : up
> lldp : {}
> mac : []
> mac_in_use : "fe:13:97:32:18:d6"
> mtu : 1500
> mtu_request : []
> name : "vnet1"
> ofport : 13
> ofport_request : []
> options : {}
> other_config : {}
> statistics : {collisions=0, rx_bytes=88983, rx_crc_err=0,
> rx_dropped=0, rx_errors=0, rx_frame_err=0, rx_over_err=0, rx_packets=470,
> tx_bytes=33425, tx_dropped=0, tx_errors=0, tx_packets=279}
> status : {driver_name=tun, driver_version="1.6",
> firmware_version=""}
> type : ""
> _______________________________________________
> discuss mailing list
> disc...@openvswitch.org
> https://urldefense.com/v3/__https://mail.openvswitch.org/mailman/listinfo/ovs-discuss__;!!GqivPVa7Brio!OpwDHJWIjB7hTj11XQyHReneJczB2hoqvhMjtBh7eytwLn1qYOJXmefbQ3-oUFaE0AI$
>
>
> _______________________________________________
> discuss mailing list
> disc...@openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
_______________________________________________
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
