What happened:
ln al2023 the max-pods calculation in nodeadm doesn't take into account custom networking configurations (CNI_CUSTOM_NETWORKING_ENABLED) which causes incorrect pod calculations.
What you expected to happen:
The max-pods calculation code for al2023 should consider custom networking configuration and exclude the primary ENI from the calculation when custom networking is enabled, similar to how it worked in AL2 with the max-pods-calculator.sh script.
If I understand it correctly the shell script does this:
Assume: 3 enis, 12 IPs per eni:
In Al2 shell script it would be: 2 * (12-1) + 2 =24
In AL2023 it is currently this: 3 *(12-1) + 2 = 35
Additionally, the current implementation is missing other features from AL2's calculator:
- prefix delegation support
- max ENI configuration support
- network card-specific handling
How to reproduce it (as minimally and precisely as possible):
Environment:
- AWS Region:
- Instance Type(s):
- Cluster Kubernetes version:
- Node Kubernetes version:
- AMI Version:
What happened:
ln al2023 the max-pods calculation in nodeadm doesn't take into account custom networking configurations (CNI_CUSTOM_NETWORKING_ENABLED) which causes incorrect pod calculations.
What you expected to happen:
The max-pods calculation code for al2023 should consider custom networking configuration and exclude the primary ENI from the calculation when custom networking is enabled, similar to how it worked in AL2 with the max-pods-calculator.sh script.
If I understand it correctly the shell script does this:
Assume: 3 enis, 12 IPs per eni:
In Al2 shell script it would be: 2 * (12-1) + 2 =24
In AL2023 it is currently this: 3 *(12-1) + 2 = 35
Additionally, the current implementation is missing other features from AL2's calculator:
How to reproduce it (as minimally and precisely as possible):
Environment: