Loxilb Config Issue for AMF pod

[chugh50]

Hello,I am trying to setup loxilb loadbalancer for amf pods ,So this is my service setup

apiVersion: v1
kind: Service
metadata:
name: amf-loxilb-service
#namespace: kube-system
annotations:
loxilb.io/liveness: "yes"
loxilb.io/lbmode: "onearm"
loxilb.io/probetype : "sctp"
#loxilb.io/probetype: "none"
loxilb.io/staticIP: "10.10.19.69"
#loxilb.io/sticky-sessions: "false"
spec:
loadBalancerClass: loxilb.io/loxilb
selector:
#what: amf-app
app.kubernetes.io/instance: divyamls
app.kubernetes.io/name: helm-divyam-amf
ports:

• port: 38412
  targetPort: 38412
  protocol: SCTP
  type: LoadBalancer
  externalTrafficPolicy: Local
  sessionAffinity: None

So whenever i set some probetype value i get sctp connection request (all procedure) in tcpdump but it didnot reach to my amf pods logs but if i set probetype to none i am able to recieve the sctp connection request in tcpdump as well as in amf logs (and there is no difference in tcpdump of both the cases)

can someone look into that why setting probtype giving this error

[TrekkieCoder]

Hi @chugh50,

I would suggest you test with a service yaml as follows :

apiVersion: v1
kind: Service
metadata:
  name: sctp-lb1
  annotations:
    loxilb.io/liveness: "yes"
    loxilb.io/usepodnetwork : "yes"
spec:
  externalTrafficPolicy: Local
  loadBalancerClass: loxilb.io/loxilb
  selector:
    what: sctp-ds-test
  ports:
    - port: 55003
      protocol: SCTP
      targetPort: 9999
  type: LoadBalancer
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: sctp-ds-ds
  labels:
    what: sctp-ds-test
spec:
  replicas: 2
  selector:
    matchLabels:
      what: sctp-ds-test
  template:
    metadata:
      labels:
        what: sctp-ds-test
    spec:
      containers:
      - name: sctp-ds-pod
        image: loxilbio/sctp-darn:latest
        command: ["sctp_darn","-H", "0.0.0.0","-P", "9999", "-l"]
        ports:
        - containerPort: 9999

Please note that if you kill the pods with kubectl delete pods, the pods are respawned quite fast before loxilb get detect they are even dead. You can also test replicaset size using :

kubectl edit deployment sctp-ds-ds

And then manually set "replicas" field. Also, feel free to try variations of loxilb.io/probetimeout and loxilb.io/proberetries.