Question

我在我的K8S集群上运行了声纳浮标，DNS测试失败，但是我无法复制该问题。如何，我如何理解它们为什么失败？

K8S版本-v1.9.11

Infra-Azure，Acs引擎v0.23.1

Sonobuoy版本-v0.11.6（最新支持K8S v1.9.11）。

运行命令-

sonobuoy run --kube-conformance-image "gcr.io/heptio-images/kube-conformance:v1.9"

失败的DNS测试：

[sig-network] DNS should provide DNS for services  [Conformance]
[sig-network] DNS should provide DNS for the cluster  [Conformance]

这是对K8S配置进行基准测试并在投入使用之前对其进行测试的一部分。

我手动运行的测试（我从日志中提取了该测试）：

dig +notcp +noall +answer +search netperf-w2 A
dig +tcp +noall +answer +search netperf-w2 A
dig +notcp +noall +answer +search netperf-w2.network-test A
dig +tcp +noall +answer +search netperf-w2.network-test A
dig +notcp +noall +answer +search netperf-w2.network-test.svc A
dig +tcp +noall +answer +search netperf-w2.network-test.svc A
dig +notcp +noall +answer +search _http._tcp.netperf-w2.network-test.svc SRV
dig +tcp +noall +answer +search _http._tcp.netperf-w2.network-test.svc SRV 
dig +notcp +noall +answer +search _http._tcp.test-service-2.network-test.svc SRV
dig +tcp +noall +answer +search _http._tcp.test-service-2.network-test.svc SRV

#podARec=$$(hostname -i| awk -F. '{print $$1"-"$$2"-"$$3"-"$$4".network-test.pod.cluster.local"}');
dig +notcp +noall +answer +search 10-240-1-76.network-test.pod.cluster.local A
dig +tcp +noall +answer +search 10-240-1-76.network-test.pod.cluster.local A
dig +notcp +noall +answer +search 103.187.0.10.in-addr.arpa. PTR
dig +tcp +noall +answer +search 103.187.0.10.in-addr.arpa. PTR

日志显示很多：

Jan  8 10:17:47.221: INFO: Unable to read wheezy_tcp@PodARecord from pod dns-test-7dbef827-132e-11e9-bd06-e20b33d4fc6d: the server could not find the requested resource (get pods dns-test-7dbef827-132e-11e9-bd06-e20b33d4fc6d)
Jan  8 10:17:47.346: INFO: Unable to read jessie_tcp@PodARecord from pod dns-test-7dbef827-132e-11e9-bd06-e20b33d4fc6d: the server could not find the requested resource (get pods dns-test-7dbef827-132e-11e9-bd06-e20b33d4fc6d)
Jan  8 10:17:47.356: INFO: Lookups using dns-test-7dbef827-132e-11e9-bd06-e20b33d4fc6d failed for: [wheezy_tcp@PodARecord jessie_tcp@PodARecord]

它的结尾是：

• Failure [619.219 seconds]
[sig-network] DNS
/workspace/anago-v1.9.4-beta.0.53+bee2d1505c4fe8/src/k8s.io/kubernetes/_output/dockerized/go/src/k8s.io/kubernetes/test/e2e/network/framework.go:22
should provide DNS for services  [Conformance] [It]
/workspace/anago-v1.9.4-beta.0.53+bee2d1505c4fe8/src/k8s.io/kubernetes/_output/dockerized/go/src/k8s.io/kubernetes/test/e2e/framework/framework.go:648

Expected error:
    <*errors.errorString | 0xc42026ab20>: {
        s: "timed out waiting for the condition",
    }
    timed out waiting for the condition
not to have occurred

/workspace/anago-v1.9.4-beta.0.53+bee2d1505c4fe8/src/k8s.io/kubernetes/_output/dockerized/go/src/k8s.io/kubernetes/test/e2e/network/dns.go:170

最终，我希望了解故障的根本原因，并对如何更好地调试它有所了解。

The test在Kubernetes的仓库中。

Sonobuoy DNS测试失败，无法手动复制失败

0 个答案: