Constant Reloads with SVC backend

[tman5]

Using ingress 3.0.1, we're seeing constant reloads with this message

2024-09-03T16:21:11.086417412-04:00 stderr F 2024/09/03 20:21:11 INFO    service/service.go:157 [transactionID=f6269b17-dde8-436d-ac00-b740858bfcb9] reload required : Service 'clark-developer-featbit/featbit-dev2-els': backend 'clark-developer-featbit_featbit-dev2-els_80' updated: map[HTTPConnectionMode:[ http-server-close]]

We do have a backend object object

---
apiVersion: "ingress.v1.haproxy.org/v1"
kind: Backend
metadata:
  name: featbit-haproxy-backend-config
  namespace: clark-developer-featbit
spec:
  config:
    name: featbit-haproxy-backend-config
    http-server-close: "enabled"
    tunnel_timeout: 3600000

...

Ingress object

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  annotations:
    haproxy.org/cr-backend: clark-developer-featbit/featbit-haproxy-backend-config
    healthcheck: /health/liveness
    meta.helm.sh/release-name: featbit-dev2
    meta.helm.sh/release-namespace: clark-developer-featbit
  creationTimestamp: "2023-09-29T15:46:58Z"
  generation: 2
  labels:
    app.kubernetes.io/instance: featbit-dev2
    app.kubernetes.io/managed-by: Helm
    app.kubernetes.io/name: featbit-dev2
    argocd.argoproj.io/instance: featbit-dev2
    helm.sh/chart: featbit-0.4.1
  name: featbit-dev2-els-ingress
  namespace: clark-developer-featbit
  resourceVersion: "376035891"
  uid: d96c0bfc-bc4e-4629-acd6-2e300e42c41a
spec:
  ingressClassName: haproxy
  rules:
  - host: featbit-eval-2.dev.internal
    http:
      paths:
      - backend:
          service:
            name: featbit-dev2-els
            port:
              number: 80
        path: /
        pathType: ImplementationSpecific

[ivanmatmati]

Thanks @tman5 for reporting, We've seen this bug and will provide a fix.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[beepee9]

@ivanmatmati We noticed #678 was linked to this issue, but is now closed while this one remains open. Should the fix for #678 have resolved this issue as well? We did attempt to upgrade to latest version, but we are still seeing the constant reloads.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[bpereira-wss]

@hdurand0710 I noticed this issue was linked to #678 but that one has been closed. We have tried upgrading to the latest release and we are still experiencing constant reloads. Previously it was about every 5 mins, but now it is about every 10 mins. Would it be possible to revisit this?

[hdurand0710]

@bpereira-wss , @beepee9
It is linked ... but #678 fixes only the Balance.URIPathOnly issue, not all issues and the root cause of constant reloads.
A more general fix for constant reloads to due to wrong diffs detected using CRDs is under development and should be soon available.

[bpereira-wss]

@bpereira-wss , @beepee9 It is linked ... but #678 fixes only the Balance.URIPathOnly issue, not all issues and the root cause of constant reloads. A more general fix for constant reloads to due to wrong diffs detected using CRDs is under development and should be soon available.

Thank you! Is there a different github issue we should keep an eye out for or will it be posted here once that is pushed through?

[hdurand0710]

@bpereira-wss ,
It will be posted here once merged.