[BUG]tcp query timeout, but udp query is ok

[joy717]

Issue

i'm using nodelocaldns, and the upstream is coredns.
it's wired that when all nodes reboot, after some minutes, the tcp dns query for nodelocaldns will be time out, the udp query is okay.

netshoot-66bc59cdd7-wrj9t:~# while true; do date; dig kubernetes.default.svc.cluster.local +short +retries=0 +tcp && sleep 5 || break; done;
...
...
Wed Jan  7 10:19:22 UTC 2026
10.233.0.1
Wed Jan  7 10:19:27 UTC 2026
10.233.0.1
Wed Jan  7 10:19:32 UTC 2026
10.233.0.1
Wed Jan  7 10:19:37 UTC 2026
10.233.0.1
Wed Jan  7 10:19:42 UTC 2026
10.233.0.1
Wed Jan  7 10:19:47 UTC 2026
10.233.0.1
Wed Jan  7 10:19:52 UTC 2026
;; communications error to 169.254.25.10#53: timed out

; <<>> DiG 9.20.10 <<>> kubernetes.default.svc.cluster.local +short +retries=0 +tcp
;; global options: +cmd
;; no servers could be reached

the nodelocaldns cm:
10.233.0.3 is the coredns svc ip
169.254.25.10 is the ip for nic nodelocaldns

apiVersion: v1
data:
  Corefile: |
    cluster.local:53 {
        errors
        cache {
            success 9984 30
            denial 9984 5
        }
        reload
        loop
        bind 169.254.25.10
        forward . 10.233.0.3 {
            force_tcp
        }
        prometheus :9253
        health 169.254.25.10:9254
    }
    in-addr.arpa:53 {
        errors
        cache 30
        reload
        loop
        bind 169.254.25.10
        forward . 10.233.0.3 {
            force_tcp
        }
        prometheus :9253
    }
    ip6.arpa:53 {
        errors
        cache 30
        reload
        loop
        bind 169.254.25.10
        forward . 10.233.0.3 {
            force_tcp
        }
        prometheus :9253
    }
    .:53 {
        errors
        cache 30
        reload
        loop
        bind 169.254.25.10
        forward . 223.5.5.5
        prometheus :9253
    }
kind: ConfigMap

if I query via coredns directly, it's ok
if I delete the nodelocaldns pods, it's ok
if I add plugin pprof, it's ok

i debuged with a netshoot container, and i find that the tcp handshake is establishmented, but no data from nodelocaldns to client.

Can somebody help me?
Thanks in advanced.

---

k8s-dns-node-cache:1.21.1
coredns:v1.10.1
kubernetes: v1.24.6