Is this a new feature, an enhancement, or a change to existing functionality?
Enhancement
How would you describe the priority of this feature request
High
Please provide a clear description of problem this feature solves
In multi-pod machine-a-tron deployments, each pod independently generates mac addresses starting from the same default base (02:00:00:00:00:00). This causes duplicate MAC addresses across pods, leading to:
- Conflicting machine registrations in NICo
- DHCP lease conflicts
- Unpredictable behavior in site-explorer
Real-world example: Deploying 2 pods with 100 hosts each would generate:
- Pod 0: MACs
02:00:00:00:00:01 through 02:00:00:00:00:64
- Pod 1: MACs
02:00:00:00:00:01 through 02:00:00:00:00:64 (duplicates!)
Feature Description
When deploying multiple machine-a-tron pods, the Helm chart should generate unique mac address pools configurations per pod, ensuring each pod allocates from a distinct mac address range to prevent collisions.
Key capabilities:
- Per-pod MAC pool base - Each pod gets a unique base MAC address derived from pod index
- Configurable pool size -
host_bits controls the size of each pod's MAC pool
- Optional explicit config - Users can override auto-generated pools with explicit values
Two MAC address pools are auto-generated per pod:
mac_address_pool - Standalone MACs for BMC and management interfaces:
hw_mac_address_ranges - Contiguous MAC blocks for NICs
Describe your ideal solution
Add mac pool configuration to machineDefaults:
machineDefaults:
# MAC address pool for simulated machines
macAddressPool:
# Base prefix (first 2 bytes). Pod index is added to byte 3.
# Uses locally administered address (02:xx:xx:xx:xx:xx)
basePrefix: "02:00"
# Number of host bits per pod (~1M macs per pod)
hostBits: 20
Allow per-pod override:
pods:
mat-0:
macAddressPool:
basePrefix: "02:01"
hostBits: 16 # ~65k mac addresses per pod
Additional context
MAC Address Format:
02:YY:PP:XX:XX:XX
02 - Locally Administered Address (LAA) (required for virtual mac addresses)
YY - Fixed prefix
PP - Pod index (0x00, 0x01, 0x02, ...)
XX:XX:XX - Host bits (allocated sequentially)
Code of Conduct
Is this a new feature, an enhancement, or a change to existing functionality?
Enhancement
How would you describe the priority of this feature request
High
Please provide a clear description of problem this feature solves
In multi-pod machine-a-tron deployments, each pod independently generates mac addresses starting from the same default base (
02:00:00:00:00:00). This causes duplicate MAC addresses across pods, leading to:Real-world example: Deploying 2 pods with 100 hosts each would generate:
02:00:00:00:00:01through02:00:00:00:00:6402:00:00:00:00:01through02:00:00:00:00:64(duplicates!)Feature Description
When deploying multiple machine-a-tron pods, the Helm chart should generate unique mac address pools configurations per pod, ensuring each pod allocates from a distinct mac address range to prevent collisions.
Key capabilities:
host_bitscontrols the size of each pod's MAC poolTwo MAC address pools are auto-generated per pod:
mac_address_pool- Standalone MACs for BMC and management interfaces:hw_mac_address_ranges- Contiguous MAC blocks for NICsDescribe your ideal solution
Add mac pool configuration to
machineDefaults:Allow per-pod override:
Additional context
MAC Address Format:
02:YY:PP:XX:XX:XX02 - Locally Administered Address (LAA) (required for virtual mac addresses)
YY - Fixed prefix
PP - Pod index (0x00, 0x01, 0x02, ...)
XX:XX:XX - Host bits (allocated sequentially)
Code of Conduct