使用“對等名稱沖突”在 Kubernetes 中設置 um WeaveNet 時出錯

Question

我正在設置 Kubernetes-Cluser，但無法正確建立編織網絡。

我有 3 個節點：rowlf（主）、rizzo 和 fozzie。 豆莢看起來不錯：

NAMESPACE     NAME                                READY     STATUS    RESTARTS   AGE
kube-system   pod/etcd-rowlf                      1/1       Running   0          32m
kube-system   pod/kube-apiserver-rowlf            1/1       Running   9          33m
kube-system   pod/kube-controller-manager-rowlf   1/1       Running   0          32m
kube-system   pod/kube-dns-686d6fb9c-kjdxt        3/3       Running   0          33m
kube-system   pod/kube-proxy-6kpr9                1/1       Running   0          9m
kube-system   pod/kube-proxy-f7nk5                1/1       Running   0          33m
kube-system   pod/kube-proxy-nrbbl                1/1       Running   0          21m
kube-system   pod/kube-scheduler-rowlf            1/1       Running   0          32m
kube-system   pod/weave-net-4sj4n                 2/2       Running   1          21m
kube-system   pod/weave-net-kj6q7                 2/2       Running   1          9m
kube-system   pod/weave-net-nsp22                 2/2       Running   0          30m

但是編織狀態顯示失敗：

$ kubectl exec -n kube-system weave-net-nsp22 -c weave -- /home/weave/weave --local status

Version: 2.3.0 (up to date; next check at 2018/06/14 00:30:09)

Service: router
Protocol: weave 1..2
Name: 7a:8f:22:1f:0a:17(rowlf)
Encryption: disabled
PeerDiscovery: enabled
Targets: 1
Connections: 1 (1 failed)
Peers: 1
TrustedSubnets: none

Service: ipam
Status: ready
Range: 10.32.0.0/12
DefaultSubnet: 10.32.0.0/12

首先，我不明白為什么連接被標記為失敗。 在日志中，我發現了這兩行：

INFO: 2018/06/13 17:22:59.170536 ->[172.16.20.12:54077] connection accepted
INFO: 2018/06/13 17:22:59.480262 ->[172.16.20.12:54077|7a:8f:22:1f:0a:17(rowlf)]: connection shutting down due to error: local "7a:8f:22:1f:0a:17(rowlf)" and remote "7a:8f:22:1f:0a:17(rizzo)" peer names collision
INFO: 2018/06/13 17:34:12.668693 ->[172.16.20.13:52541] connection accepted
INFO: 2018/06/13 17:34:12.672113 ->[172.16.20.13:52541|7a:8f:22:1f:0a:17(rowlf)]: connection shutting down due to error: local "7a:8f:22:1f:0a:17(rowlf)" and remote "7a:8f:22:1f:0a:17(fozzie)" peer names collision

第二個被誤解的事情是“對等名稱沖突”錯誤。 這是正常的嗎？

這是來自“rizzo”的日志

kubectl logs weave-net-4sj4n -n kube-system weave

DEBU: 2018/06/13 17:22:58.731864 [kube-peers] Checking peer "7a:8f:22:1f:0a:17" against list &{[{7a:8f:22:1f:0a:17 rowlf}]}
INFO: 2018/06/13 17:22:58.833350 Command line options: map[conn-limit:100 docker-api: host-root:/host http-addr:127.0.0.1:6784 ipalloc-range:10.32.0.0/12 no-dns:true expect-npc:true name:7a:8f:22:1f:0a:17 datapath:datapath db-prefix:/weavedb/weave-net ipalloc-init:consensus=2 metrics-addr:0.0.0.0:6782 nickname:rizzo port:6783]
INFO: 2018/06/13 17:22:58.833525 weave  2.3.0
INFO: 2018/06/13 17:22:59.119956 Bridge type is bridged_fastdp
INFO: 2018/06/13 17:22:59.120025 Communication between peers is unencrypted.
INFO: 2018/06/13 17:22:59.141576 Our name is 7a:8f:22:1f:0a:17(rizzo)
INFO: 2018/06/13 17:22:59.141787 Launch detected - using supplied peer list: [172.16.20.12 172.16.20.11]
INFO: 2018/06/13 17:22:59.141894 Checking for pre-existing addresses on weave bridge
INFO: 2018/06/13 17:22:59.157517 [allocator 7a:8f:22:1f:0a:17] Initialising with persisted data
INFO: 2018/06/13 17:22:59.157884 Sniffing traffic on datapath (via ODP)
INFO: 2018/06/13 17:22:59.158806 ->[172.16.20.11:6783] attempting connection
INFO: 2018/06/13 17:22:59.159081 ->[172.16.20.12:6783] attempting connection
INFO: 2018/06/13 17:22:59.159815 ->[172.16.20.12:42371] connection accepted
INFO: 2018/06/13 17:22:59.161572 ->[172.16.20.12:6783|7a:8f:22:1f:0a:17(rizzo)]: connection shutting down due to error: cannot connect to ourself
INFO: 2018/06/13 17:22:59.161836 ->[172.16.20.12:42371|7a:8f:22:1f:0a:17(rizzo)]: connection shutting down due to error: cannot connect to ourself
INFO: 2018/06/13 17:22:59.265736 Listening for HTTP control messages on 127.0.0.1:6784
INFO: 2018/06/13 17:22:59.266483 Listening for metrics requests on 0.0.0.0:6782
INFO: 2018/06/13 17:22:59.443937 ->[172.16.20.11:6783|7a:8f:22:1f:0a:17(rizzo)]: connection shutting down due to error: local "7a:8f:22:1f:0a:17(rizzo)" and remote "7a:8f:22:1f:0a:17(rowlf)" peer names collision
INFO: 2018/06/13 17:23:00.355761 [kube-peers] Added myself to peer list &{[{7a:8f:22:1f:0a:17 rowlf}]}
DEBU: 2018/06/13 17:23:00.367309 [kube-peers] Nodes that have disappeared: map[]
INFO: 2018/06/13 17:34:12.671287 ->[172.16.20.13:60523] connection accepted
INFO: 2018/06/13 17:34:12.674712 ->[172.16.20.13:60523|7a:8f:22:1f:0a:17(rizzo)]: connection shutting down  due to error: local "7a:8f:22:1f:0a:17(rizzo)" and remote "7a:8f:22:1f:0a:17(fozzie)" peer names collision

我問，因為我現在第四次從頭開始重新安裝所有東西，每次我在從 traefik 連接到另一台主機上的 pod 時遇到一些麻煩。 我責怪網絡，因為這看起來不健康。 你能告訴我到目前為止設置是否正確。 錯誤是正常的還是我必須關心它們？ 最后：我如何請求幫助以及我必須提供哪些信息才能讓像您這樣的人輕松幫助我擺脫這個令人沮喪的境地？

這是我的版本：

Client Version: version.Info{Major:"1", Minor:"10", GitVersion:"v1.10.2", GitCommit:"81753b10df112992bf51bbc2c2f85208aad78335", GitTreeState:"clean", BuildDate:"2018-04-27T09:22:21Z", GoVersion:"go1.9.3", Compiler:"gc", Platform:"linux/arm"}
Server Version: version.Info{Major:"1", Minor:"10", GitVersion:"v1.10.4", GitCommit:"5ca598b4ba5abb89bb773071ce452e33fb66339d", GitTreeState:"clean", BuildDate:"2018-06-06T08:00:59Z", GoVersion:"go1.9.3", Compiler:"gc", Platform:"linux/arm"}

謝謝你。

++++ 更新 ++++ 我像這里提到的那樣重置機器 ID： https : //github.com/weaveworks/weave/issues/2767但這會導致我的機器不斷重啟！

kernel:[ 2257.674153] Internal error: Oops: 80000007 [#1] SMP ARM

Answer 1

最后我在這里找到了解決方案： https : //github.com/weaveworks/weave/issues/3314我們必須禁用fastDP！

Answer 2

我遇到了同樣的問題，禁用 fastDP 對我不起作用，但我發現原因是由於我從同一操作系統映像克隆它們，所有節點的/etc/machine-id值都相同。

我從所有節點中刪除了機器 ID，並使用以下命令生成了新的機器 ID：

sudo rm /etc/machine-id
sudo systemd-machine-id-setup

然后重置我的集群

Answer 3

'sudo systemd-machine-id-setup' 為我生成機器 ID 作為舊機器 ID。 我只是編輯機器 ID 並且它可以工作。