Closed rocoll closed 11 months ago
Looking to see if we can put a simple check on the format when consuming the FQDN value.
@rocoll is this error being generated from kustomize? Also where are you setting .gelsandbox.aws.unx.sas.com. If it's an ansible variable we could have a check associated. Looking at assigning or closing these older issues out. Thanks again.
Marking as stale/inactive. If there are further questions please open a new GitHub issue.
I screwed up with naming an env var and inadvertently directed viya4-deployment to build Ingress rules referring to a malformed DNS name.
If I run
kubectl apply
directly, I can watch it generate 100+ error messages about this problem. However, Ansible (as directed by viya4-deployment) gave no hint at all about the problem. Its play recap mentions zero errors. From past experience, I understand that a zero error Ansible run doesn't equal a perfectly sound deployment. But scores of error messages about a critical infrastructure component should probably be captured and announced.It took a while to run it down. After completing the failed deployment, the key symptom is that none of the Viya pods start (stuck initializing, crash loop backoff). Digging into them, certframe complains it cannot find "sas-ingress-certificate" (HTTP 404 error). And sure enough, that secret doesn't exist. But the real clue was noticing that none of the ingress rules exist.