Allow administrators to configure the default MTU used when creating new networks in OSIE, while still supporting higher MTU values for environments that run jumbo frames in the backend.
In our RHOSO environment we run a 9000 MTU backend fabric (configured via global_physnet_mtu).
Currently:
OSIE does not allow MTU to be configured when creating a network
When a network is created without specifying an MTU, it automatically inherits the maximum MTU (9000) from the backend
This results in all newly created networks defaulting to jumbo frames.
Running jumbo frames by default can introduce unnecessary complexity and compatibility issues for tenants and more importantly, the user is not able to reduce the MTU.
Many workloads and network integrations assume 1500 MTU.
While jumbo frames are valuable in certain cases (storage networks, high-performance workloads, etc.), enabling them by default across all tenant networks can lead to issues.
Allow cloud administrators to configure a default MTU value for networks created via OSIE.
Example behaviour:
Backend fabric MTU:
9000
Default network MTU (OSIE setting):
1500
Allowed MTU range:
1500 β 9000
When a user creates a network in OSIE:
If no MTU is specified β network is created with 1500 MTU
Admins (or optionally users) can increase MTU up to the backend maximum
In addition to setting a default MTU, it would be useful to have a toggle to control whether tenants can modify MTU when creating networks.
For example:
Admin settings:
Default network MTU: 1500
Allow tenant MTU override: Enabled / Disabled
This would allow providers to:
Maintain safe defaults
Still allow advanced users to enable jumbo frames when required
Thanks!
Please authenticate to join the conversation.
In Review
π Issues
15 days ago

Kevin.Pannell
Get notified by email when there are changes.
In Review
π Issues
15 days ago

Kevin.Pannell
Get notified by email when there are changes.