openstack/neutron
Code Issues Proposed changes

[Docs] Add info about vlan transparent networks' MTU

Browse Source 
MTU configured inside instances which are connected to the networks
with enabled transparent vlan needs to have MTU set to 4 bytes less
than it is provided by DHCP to make space for additional vlan
header.
This is required only if instance will have some vlan tagged interfaces
configured.

Closes-bug: #1906318
Change-Id: Ia6691cbe35a1e874857ec94660ac4b1850586305
This commit is contained in:
Slawek Kaplonski
2021-06-16 14:03:33 +02:00
parent 02451f381b
commit c7e1de09e1
1 changed files with 9 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
doc/source/admin/config-mtu.rst
View File

@@ -129,3 +129,12 @@ while the L3 agent provides an appropriate MTU value to instances using
IPv6. IPv6 uses RA via the L3 agent because the DHCP agent only supports IPv6. IPv6 uses RA via the L3 agent because the DHCP agent only supports
IPv4. Instances using IPv4 and IPv6 should obtain the same MTU value IPv4. Instances using IPv4 and IPv6 should obtain the same MTU value
regardless of method. regardless of method.
Networks with enabled vlan transparency
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In case of networks with enabled vlan transparency, if additional vlan tag is
configured inside guest VM, MTU has to be lowered by 4 bytes to make space for
additional vlan tag in the packet's header.
For example, if network's MTU is set to ``1500``, value configured for the
interfaces in the guest vm should be manually set to ``1496`` or less bytes.