Routetable table routes are different compared with the ones used during the creation of the object

[adriananeci]

Describe the bug
I've noticed that the routetable rules rules are changed by the ack controller once created.
This affect an ArgoCD application that is creating such objects because it is constantly in an out of sync status.

Steps to reproduce

Create a routetable using

apiVersion: ec2.services.k8s.aws/v1alpha1
kind: RouteTable
metadata:
  annotations:
    ec2.services.k8s.aws/deletion-policy: retain
  name: dev-va6-k8s-public
spec:
  routes:
    - destinationCIDRBlock: 0.0.0.0/0
      gatewayRef:
        from:
          name: dev-va6-k8s
    - destinationCIDRBlock: 10.15.20.0/21
      transitGatewayID: tgw-abc123
  vpcRef:
    from:
      name: dev-va6-k8s

Once created the routes are getting changed into something like:

apiVersion: ec2.services.k8s.aws/v1alpha1
kind: RouteTable
metadata:
  annotations:
    ec2.services.k8s.aws/deletion-policy: retain
  name: dev-va6-k8s-public
spec:
  routes:
    - destinationCIDRBlock: 10.15.20.0/21
      transitGatewayID: tgw-abc123
    - destinationCIDRBlock: 0.0.0.0/0
      gatewayID: igw-abc123
    - destinationPrefixListID: pl-123abc
      gatewayID: vpce-123
    - destinationPrefixListID: pl-abc123
      gatewayID: vpce-xxxx
  vpcRef:
    from:
      name: dev-va6-k8s

Expected outcome
Routetable routes should remain the same after object creation

Environment

• Kubernetes version 1.27
• Using EKS (yes/no), if so version? 1.27
• AWS service targeted (S3, RDS, etc.) EC2

cc @a-hilaly

[a-hilaly]

@adriananeci looks like those are "defaults" created by the ec2 API, can you confirm? If it's the case, then it's an expected behaviour. The controllers can default some fields on Create/Update if they are not specified in the spec.

[a-hilaly]

Mmmmm looks like the gatewayRef also disapeared ...

      gatewayRef:
        from:
          name: dev-va6-k8s

Are you referring to this specific behaviour?

[adriananeci]

Yes and also to the rules with destinationPrefixListID

[adriananeci]

The symptoms look similar with the ones described in #1880

[jas-nik]

Hi @a-hilaly , just checking in, any update on this?

[a-hilaly]

@jas-nik @adriananeci we just shipped a patch in ec2 controller to fix this. Please try 1.2.24

[jas-nik]

Thats awesome. Thank you for the fix! @a-hilaly