| 1 |
--- |
| 2 |
prelude: > |
| 3 |
The ML2 plug-in supports calculating the MTU for instances |
| 4 |
using overlay networks by subtracting the overlay protocol |
| 5 |
overhead from the value of 'path_mtu', ideally the physical |
| 6 |
(underlying) network MTU, and providing the smaller value to |
| 7 |
instances via DHCP. Prior to Mitaka, 'path_mtu' defaults to |
| 8 |
0 which disables this feature. In Mitaka, 'path_mtu' defaults |
| 9 |
to 1500, a typical MTU for physical networks, to improve the |
| 10 |
"out of box" experience for typical deployments. |
| 11 |
features: |
| 12 |
- In Mitaka, the combination of 'path_mtu' defaulting to 1500 and |
| 13 |
'advertise_mtu' defaulting to True provides a value of MTU accounting |
| 14 |
for any overlay protocol overhead on the network to instances using |
| 15 |
DHCP. For example, an instance attaching to a VXLAN network receives |
| 16 |
a 1450 MTU from DHCP accounting for 50 bytes of overhead from the |
| 17 |
VXLAN overlay protocol if using IPv4 endpoints. |
| 18 |
issues: |
| 19 |
- The combination of 'path_mtu' and 'advertise_mtu' only adjusts the |
| 20 |
MTU for instances rather than all virtual network components between |
| 21 |
instances and provider/public networks. In particular, setting |
| 22 |
'path_mtu' to a value greater than 1500 can cause packet loss even |
| 23 |
if the physical network supports it. Also, the calculation does not |
| 24 |
consider additional overhead from IPv6 endpoints. |
| 25 |
upgrade: |
| 26 |
- Operators using the ML2 plug-in with 'path_mtu' defaulting to 0 may need |
| 27 |
to perform a database migration to update the MTU for existing networks |
| 28 |
and possibly disable existing workarounds for MTU problems such as |
| 29 |
increasing the physical network MTU to 1550. |