https://pulumi.com logo
#kubernetes
Title
# kubernetes
f

future-refrigerator-88869

09/02/2021, 8:50 PM
Sharing this here in case it gets more traction πŸ™‚
p

proud-pizza-80589

09/03/2021, 6:45 AM
I know what this is, just read it yesterday in a blog, let me find it
here, you should try the Pod Readiness Gate mentioned here: Pod Readiness Gate
oh, and i do not use this so my statement of β€œi know what this is” might be a slight exaggeration πŸ˜‰
f

future-refrigerator-88869

09/03/2021, 10:42 PM
Been looking at this for a while but I still can't make it work. seems like something doesn't meet the conditions. Also, this would put the POD into a "not ready" state until it gets registered to the ALB. In my case, the ALB is in provisioning state (i'm creating it as well). I still think this is a pulumi error on the awaits. I found another older github issue where the problem was the same for
extensions/v1beta1
d

dry-sugar-63293

10/11/2021, 2:25 PM
I have the same issue too. In fact in my case, the issue is not related to the POD readiness at all. everything works except the status.loadBalancer.ingress[0].ip or hostname both return empty at all times
@future-refrigerator-88869 which ingress-controller are you using? I am using GCE.
f

future-refrigerator-88869

10/16/2021, 7:07 PM
I ended up working around this issue because pulumi team did not yet update the logic behind the await for the ingress. The next version of kubernetes will deprecate the extensions.v1beta.ingress (which is the one that works). I found out that you could technically use an ingress of type
extensions.v1beta.ingress
but it is alreadt marked as deprecated. In my case, I was using
aws-load-balancer-controller
and I switched to
traefik
using custom resource definition for
ingressRoute
which works around the issue of creating and awaiting ingress resources. https://doc.traefik.io/traefik/v2.2/providers/kubernetes-crd/ @dry-sugar-63293
πŸ™Œ 1
4 Views