openshift-snc: align cluster ingress domain with the nip.io access domain (crc.testing split-brain)

Ouverte
#968 0 commentaires 1 réaction 0 personnes assignées Voir sur GitHub

Les mainteneurs répondent en général sous 4 jours

Personne n'a encore pris cette issue.

Évaluation

Difficulté
5/5
Temps estimé
Plus d'une semaine
Accessibilité débutants
25/100
Type d'issue
Fonctionnalité
Clarté
Plutôt claire
Activité
Active
Stack technique
aws, go, kubernetes

Piste de recherche

Start with pkg/provider/ibmcloud/action/snc/snc.go and trace how the current floating IP, CRC startup, and cluster readiness are handled; inspect the AWS equivalent as well. The issue describes install-time domain setup and post-readiness migration, including console routes and certificates, so first map the existing lifecycle and identify which approach the project can support. Done means cluster-generated domains and routes resolve through nip.io, including after IP rotation; consumers of immutable ingresses.config need install-time configuration or an override.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Description

Problem

mapt ibmcloud openshift-snc create (and the AWS equivalent) runs OpenShift Local with its default domain (apps-crc.testing) but exposes the cluster externally via *.apps.<floating-ip>.nip.io. Names under apps-crc.testing only resolve inside the VM, so every URL the cluster generates — oc get route, operator and app URLs, OAuth redirects — is unreachable from the user's machine.

Verified on a live cluster (openshift-local 4.22)
  • Patching ingresscontroller/default spec.domain works: new hostless routes get the new domain from route admission. status.domain is never re-derived (cosmetic).
  • ingresses.config/cluster spec.domain is immutable after install. Anything reading it — the console operator, the ingress canary controller, e2e deploy scripts using the standard idiom — keeps getting apps-crc.testing, and this cannot be fixed post-install.
  • The console must be served via its custom-route mechanism (console-custom + the default route as redirector); recreating the default route under the new domain collides (HostAlreadyClaimed) and degrades the operator.
  • The router's default certificate must cover the new domain (e.g. via spec.defaultCertificate).
  • Application-operator routes are fine: recreated after the patch, they get the new domain.
Request — two complementary changes

A. Install-time base domain (the complete fix; new clusters only): have the catalog image's self-start service run crc config set base-domain <domain> before the first crc start. Per-instance input already flows via cloud-config (/opt/crc/eip); extend it with the domain — no per-cluster images needed. With nip.io this requires a stable IP (reserved floating IP, or stop-instead-of-destroy between uses). With a real DNS domain (see #273) no stable IP is needed: the baked domain never changes, only the DNS record is updated per boot, and CRC's own certificates stay valid because they are issued for the domain.

B. Post-readiness migration (the only fix for existing clusters): on each provision, with the current floating IP:

  1. patch ingresscontroller.spec.domain to apps.<ip>.nip.io
  2. serve console/downloads via their custom routes
  3. sweep all routes carrying the previous IP or apps-crc.testing hosts
  4. ensure the default ingress certificate covers the new domain

Neither alone is sufficient: A cannot reach snapshot-restored clusters (the domain is baked at first start), B cannot fix ingresses.config readers.

A B
Fixes ingresses.config readers yes no
Works on existing clusters no yes
Needs a stable IP only with nip.io no

Happy to test a PR or preview build; a full runbook from our live-cluster verification is available on request.

Langage dominant
Go
Étoiles
25
Forks
41
Merge moyen
5 j 3 h
PR mergées (30 j)
16

Préparer son environnement

Ce projet ne fournit ni conteneur de développement, ni Dockerfile, ni guide de contribution : l'installation est à votre charge. Commencez par son README, et consultez notre guide de la première contribution pour les étapes générales.

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de redhat-developer/mapt

Toutes les issues de redhat-developer/mapt

Issues similaires

Plus d'issues Go

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.