Include the missing resource id in not-found messages, expire auth to Guest after 401, and document Ingress annotations so nginx does not rewrite 404/405 bodies into branded HTML.
3.1 KiB
Language / Язык: English | Русский
Troubleshooting
/health/ready is 503
- Postgres not up or wrong
DATABASE_URL - Migrations not applied — check migrate initContainer /
make db-migrate - Helm:
kubectl logson the simulator pod (migrate init)
Auth 401
- Wrong user/password/domain (
Default) - Project scope missing for project-scoped APIs
- Token from a different simulator instance (reseed rotates IDs)
- In the Web UI, HTTP 401 clears the local Keystone session and shows Guest in the header; sign in again from Environment
Ingress returns branded HTML 404 / nginx 405 instead of JSON
The simulator answers API errors as JSON (error / itemNotFound / message).
If you see a site HTML “page not found” or plain nginx 405 page, the
Ingress / reverse proxy replaced the upstream body (often via
custom-http-errors on the ingress-nginx controller).
Fix on the Ingress for this host (see
helm/openstack-api-simulator/values-ingress-example.yaml):
annotations:
nginx.ingress.kubernetes.io/proxy-intercept-errors: "false"
nginx.ingress.kubernetes.io/custom-http-errors: "502,503"
Then re-check with Accept: application/json. A missing compute instance should
look like JSON (not HTML), for example:
{"itemNotFound": {"code": 404, "message": "Instance 'missing-id' could not be found"}}
Correct authenticated mutation (OpenStack)
Use a Keystone token header and JSON body (OpenStack APIs are JSON, not form-urlencoded):
# after POST /v3/auth/tokens → X-Subject-Token
TOKEN=...
curl -sS -X POST "https://HOST:8774/v2.1/servers" \
-H "X-Auth-Token: $TOKEN" \
-H "Content-Type: application/json" \
-H "Accept: application/json" \
-d '{"server":{"name":"demo","flavorRef":"...","imageRef":"...","networks":[{"uuid":"..."}]}}'
Keystone/UI via Ingress is usually :443→5000; Nova and other service ports
still need port-forward / LoadBalancer / TCP Ingress unless you only call
through the multi-port gateway Service.
Empty lists after lifecycle probe
Lifecycle DELETE can remove demo-scoped rows. Reload:
make seed-demo
# or Helm:
kubectl exec deploy/… -- python -m app.openstack.seed_cli --profile demo
Wrong service answers on a port
Confirm gateway headers:
curl -sI http://127.0.0.1:8774/ | grep -i openstack
Expect X-OpenStack-Service: nova. If you hit simulator :8080 directly,
set X-OpenStack-Route-Service / X-OpenStack-Service yourself.
Helm port-forward to 5000 fails
Forward the gateway Service, not the simulator Service:
kubectl port-forward svc/<release>-openstack-api-simulator-gateway 5000:5000
Pack activate 404 / empty ops
Ensure contracts/openstack/<series>/ exists in the image and
OPENSTACK_SERIES is a known series name.
Client catalog ports unreachable
Catalog advertises per-service ports. With only Ingress on :443→5000, Nova
:8774 is not automatically published. Port-forward or expose the gateway
Service (see kubernetes.md).