Hi,
i am using pve 5.1 with openvswitch, network is set to 9k MTU. Setup looks as this:
So, everything is on 9k MTU. But...sometimes VM start with 1500 MTU (or, change to MTU 1500 via ip link from pve host) and OVS automatically switch to lowest MTU on vmbr1, so this is result:
So communication fails. In OVS documentation is "mtu_request" option to disable automatic change, but it doesn't work on bond1. Any way how to block bond1 from changing MTU?
Or, any way to reconfig to standard linux bridges without needing to configure any vlan? I need OVS feature using any vlan without configuration on PVE hosts, just adding vlan tag in vm's network device config.
i am using pve 5.1 with openvswitch, network is set to 9k MTU. Setup looks as this:
Code:
allow-auto ens2f2
iface ens2f2 inet manual
...
allow-vmbr1 bond1
iface bond1 inet manual
ovs_bridge vmbr1
ovs_type OVSBond
ovs_bonds ens2f2 ...
ovs_options lacp=active bond-mode=balance-tcp
pre-up ( /sbin/ip link set mtu 9000 ens2f2 ....)
up /sbin/ip link set mtu 9000 bond1
auto vmbr1
iface vmbr1 inet manual
ovs_type OVSBridge
ovs_ports bond1
mtu 9000So, everything is on 9k MTU. But...sometimes VM start with 1500 MTU (or, change to MTU 1500 via ip link from pve host) and OVS automatically switch to lowest MTU on vmbr1, so this is result:
Code:
ens2f2 - 9k
bond1 - 1500
vmbr1 - 9k
any internal port on vmbr1 - 1500So communication fails. In OVS documentation is "mtu_request" option to disable automatic change, but it doesn't work on bond1. Any way how to block bond1 from changing MTU?
Or, any way to reconfig to standard linux bridges without needing to configure any vlan? I need OVS feature using any vlan without configuration on PVE hosts, just adding vlan tag in vm's network device config.