#10557 AWS: VPCs / Subnets for Fedora CI EKS staging instances do not work
Closed: Fixed 2 years ago by mvadkert. Opened 2 years ago by mvadkert.

After spending like a 2 month trying to understand why we cannot install EKS cluster to following VPCs and subnets, blaming terraform and our insufficient knowledge, we found out that they are most probably somehow misconfigured. They are both in us-east-2:

  • vpc-0618f77c2f99c9956

    • fedora-ci-eks-stg-subnet-01 / subnet-0b94535bacb5965af
    • fedora-ci-eks-stg-subnet-02 / subnet-0545e3522c521a064
  • vpc-0709d5affa675f857

    • fedora-ci-eks-stg-02-subnet-1 / subnet-06ed190439a0e2037
    • fedora-ci-eks-stg-02-subnet-1 / subnet-05d489836b767cfb6

If we try this VPC / subnets everything works well:

  • vpc-0f6baa3d6bae8d912
    fedora-ci/osci-1 / subnet-0a704a759f7671044
    fedora-ci/osci-2 / subnet-010f90da92f36876e

Could somebody take a look what is configured differently in them? When we try to install clusters with nodes on them, we cannot connect to them and the nodes cannot reach the cluster.

We use terraform [1] for setup. The same code works when we use the working VPC / subnets.

CC @mobrien @kevin

[1] https://gitlab.com/testing-farm/infrastructure/-/merge_requests/35


I see that neither of the 2 VPCs had an internet gateway attached to allow for internet traffic attached. I have added those to both VPCs. Could you try again and let me know.

Metadata Update from @mobrien:
- Issue assigned to mobrien

2 years ago

@mobrien will do but a bit later. I will reopen if needed. I am like 90% sure it is it, because it was flapping on connectivity issues via the internets.

Metadata Update from @mvadkert:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

2 years ago

Log in to comment on this ticket.

Metadata