Kamailio on AWS Kubernetes with NLB - sr-users

15 Oct 2021

Hello,

I'm trying to deploy Kamailio on Kubernetes (AWS EKS with Network Load Balancer) but
I'm not able to register using TCP.
I'm using official Docker image 5.5.0-buster.
I did some tests with Docker Compose and it worked fine, but on K8S it does not :-(

Here are my logs:

Listening on
             udp: 127.0.0.1:5060
             udp: 172.31.4.29:5060
             udp: 172.31.0.221:5060
             tcp: 127.0.0.1:5060
             tcp: 172.31.4.29:5060
             tcp: 172.31.0.221:5060
Aliases:
             tcp: ip-172-31-0-221.eu-central-1.compute.internal:5060
             tcp:
172-31-4-29.kube-prometheus-stack-prometheus-node-exporter.infra-prometheus.svc.cluster.local:5060
             tcp: localhost4.localdomain4:5060
             tcp: localhost4:5060
             tcp: localhost.localdomain:5060
             tcp: localhost:5060
             udp: ip-172-31-0-221.eu-central-1.compute.internal:5060
             udp:
172-31-4-29.kube-prometheus-stack-kubelet.kube-system.svc.cluster.local:5060
             udp: localhost4.localdomain4:5060
             udp: localhost4:5060
             udp: localhost.localdomain:5060
             udp: localhost:5060

 0(1) INFO: <core> [core/tcp_main.c:4983]: init_tcp(): using epoll_lt as the io
watch method (auto detected)
 0(1) INFO: rr [../outbound/api.h:52]: ob_load_api(): unable to import bind_ob - maybe
module is not loaded
 0(1) INFO: rr [rr_mod.c:188]: mod_init(): outbound module not available
 0(1) INFO: <core> [main.c:3030]: main(): processes (at least): 41 - shm size:
67108864 - pkg size: 8388608
 0(1) INFO: <core> [core/udp_server.c:154]: probe_max_receive_buffer(): SO_RCVBUF is
initially 212992
 0(1) INFO: <core> [core/udp_server.c:206]: probe_max_receive_buffer(): SO_RCVBUF is
finally 425984
 0(1) INFO: <core> [core/udp_server.c:154]: probe_max_receive_buffer(): SO_RCVBUF is
initially 212992
 0(1) INFO: <core> [core/udp_server.c:206]: probe_max_receive_buffer(): SO_RCVBUF is
finally 425984
 0(1) INFO: <core> [core/udp_server.c:154]: probe_max_receive_buffer(): SO_RCVBUF is
initially 212992
 0(1) INFO: <core> [core/udp_server.c:206]: probe_max_receive_buffer(): SO_RCVBUF is
finally 425984
29(35) INFO: jsonrpcs [jsonrpcs_sock.c:443]: jsonrpc_dgram_process(): a new child 0/35
31(37) INFO: ctl [io_listener.c:213]: io_listen_loop(): io_listen_loop:  using epoll_lt io
watch method (config)
35(41) INFO: <core> [core/parser/parse_fline.c:147]: parse_first_line(): method not
followed by SP
35(41) ERROR: <core> [core/parser/parse_fline.c:271]: parse_first_line():
parse_first_line: bad message (offset: 0)
35(41) ERROR: <core> [core/parser/msg_parser.c:747]: parse_msg(): ERROR: parse_msg:
message=<>
35(41) ERROR: <core> [core/receive.c:374]: receive_msg(): core parsing of SIP
message failed (194.230.146.66:9055/2)

And my service:

apiVersion: v1
kind: Service
metadata:
  name: kamailio-tcp
  annotations:
    service.beta.kubernetes.io/aws-load-balancer-nlb-target-type: ip
    service.beta.kubernetes.io/aws-load-balancer-proxy-protocol: '*'
    service.beta.kubernetes.io/aws-load-balancer-scheme: internet-facing
    service.beta.kubernetes.io/aws-load-balancer-target-group-attributes:
preserve_client_ip.enabled=true
    service.beta.kubernetes.io/aws-load-balancer-type: external
  labels:
    app.kubernetes.io/instance: test-sip
    app.kubernetes.io/name: kamailio-tcp
    app.kubernetes.io/version: 5.5.0-buster
spec:
  type: LoadBalancer
  externalTrafficPolicy: Local
  ports:
    - name: sip-tcp
      nodePort: 30567
      port: 5060
      protocol: TCP
      targetPort: 5060
  selector:
    app.kubernetes.io/instance: test-sip
    app.kubernetes.io/name: kamailio

I have a Netwok Load Balancer create with a listener for TCP 5060 but when I try to
connect with Zoiper or Linphone, it does not work :-(

I used tcpdump on my pod and it received well some SIP commands.
Any ideas of what could be wrong here?

Thanks in advance!

Cedric