TCP/UDP Loadbalancer doesn't work in 2.19.0 and 2.19.1

[hzproe]

Setup Details

CIS Version : 2.19.0
Build: f5networks/k8s-bigip-ctlr:2.19.0
BIGIP Version: Big IP x.x.x
AS3 Version: 3.x
Agent Mode: AS3/CCCL
Orchestration: K8S/OSCP
Orchestration Version:
Pool Mode: Cluster
Additional Setup details: <Platform/CNI Plugins/ cluster nodes/ etc>

Description

We have a configuration where we want to support TCP and UDP on the same Port (LoadBalancer).
But as soon as the BigIP Controller is restarted, the VIP is removed from one of the LoadBalancer Services in Kubernetes.
This happens from version 2.19.0 onwards.

Current workaround is to fallback to 2.18.1 but this is not sustainable for the environment, as I have to keep our controllers up to date for operational and security reasons.

Steps To Reproduce

1. Create a LoadBalancer with port 53 and protocol TCP

apiVersion: v1
kind: Service
metadata:
annotations:
cis.f5.com/health: '{"interval": 5, "timeout": 10}'
cis.f5.com/ip:
labels:
app.kubernetes.io/name: powerdns
name: powerdns-tcp
namespace: pdns
spec:
allocateLoadBalancerNodePorts: true
clusterIP: 10.100.223.128
clusterIPs:

• 10.100.223.128
  externalTrafficPolicy: Cluster
  internalTrafficPolicy: Cluster
  ipFamilies:
• IPv4
  ipFamilyPolicy: SingleStack
  ports:
• name: dns-tcp
  nodePort: 31527
  port: 53
  protocol: TCP
  targetPort: 53
  selector:
  app.kubernetes.io/name: powerdns
  sessionAffinity: None
  type: LoadBalancer

2. Create a LoadBalancer with port 53 and protocol UDP

apiVersion: v1
kind: Service
metadata:
annotations:
cis.f5.com/health: '{"interval": 5, "timeout": 10}'
cis.f5.com/ip:
labels:
app.kubernetes.io/name: powerdns
name: powerdns-tcp
namespace: pdns
spec:
allocateLoadBalancerNodePorts: true
clusterIP: 10.100.223.128
clusterIPs:

• 10.100.223.128
  externalTrafficPolicy: Cluster
  internalTrafficPolicy: Cluster
  ipFamilies:
• IPv4
  ipFamilyPolicy: SingleStack
  ports:
• name: dns-tcp
  nodePort: 31527
  port: 53
  protocol: UDP
  targetPort: 53
  selector:
  app.kubernetes.io/name: powerdns
  sessionAffinity: None
  type: LoadBalancer

Expected Result

VIPs is up for TCP and UDP protocol

Actual Result

One of the LoadBalancers will go down with the message

Normal ExternalIP 61s k8s-bigip-ctlr F5 CIS unassigned Service LB Ingress status IP for service: powerdns-tcp in namespace:pdns

Diagnostic Information

<Configuration files, error messages, logs>
Note: Sanitize the data. For example, be mindful of IPs, ports, application names and URLs
Note: The following F5 article outlines the information required when opening an issue.
https://support.f5.com/csp/article/K60974137

Observations (if any)

[trinaths]

Created [CONTCNTR-5240] for internal tracking.

[lavanya-f5]

@hzproe dev build: lavanyasirigudi/k8s-bigip-ctlr:fix_lb_sameip_different_protocol. Please share your feedback with the build