我在kubernetes集群中使用consul并且想要查询其集群外部的DNS端口，但不确定如何在haproxy中代理请求 .

带有外部IP的Google负载均衡器配置

---
kind: Service
apiVersion: v1
metadata:
  name: consul-lb
spec:
  type: LoadBalancer
  ports:
  - port: 80
    protocol: TCP
    name: tcp
  selector:
    app: haproxy-ingress-lb

haproxy pod可以使用检索dns记录

dns +noall +answer @10.8.1.2 -p 8600 abc.service.cluster.local SRV
abc.service.cluster.local. 0    IN  SRV 1 1 8070 2d210by7.addr.dc1.cluster.local.

在集群外我可以卷曲http：// 并通过haproxy检索后端nginx服务器，以便正常工作

tcpdump host {Any consul PodIP} -vv doesn't reveal much that I can tell

00:53:51.609118 IP haproxy-voltest-gvjfx.51486 > consul-2.consul.default.svc.cluster.local.8600: Flags [S], seq 219267660, win 28400, options [mss 1420,sackOK,TS val 1016956519 ecr 0,nop,wscale 7], length 0
00:53:51.609413 IP consul-2.consul.default.svc.cluster.local.8600 > haproxy-voltest-gvjfx.51486: Flags [S.], seq 2564495418, ack 219267661, win 28160, options [mss 1420,sackOK,TS val 1016956773 ecr 1016956519,nop,wscale 7], length 0
00:53:51.609821 IP haproxy-voltest-gvjfx.51486 > consul-2.consul.default.svc.cluster.local.8600: Flags [R.], seq 1, ack 1, win 222, options [nop,nop,TS val 1016956520 ecr 1016956773], length 0

我的haproxy配置是

defaults
  mode    tcp 
  option  tcplog
  option log-health-checks
  retries 3
  option  redispatch
  timeout connect  5000
  timeout client  10000
  timeout server  10000

  # make sure log-format is on a single line
  log global


frontend http-in
    bind *:80
    mode tcp
    option tcplog    
    log global 
    # Default backend to be used
    default_backend consuldns


backend consuldns
   mode tcp
   option tcplog
   timeout connect 10s
   timeout server 1m
   server server1 10.8.2.9:8600 check

提前致谢 ，

编辑：1

这是我在外部尝试的命令

dig @{external load balancer ip} -p 80 abc.service.cluster.local SRV
;; connection timed out; no servers could be reached
dns kubernetes haproxy consul