很多 kubelet 錯誤：無法更新容器的統計信息

Question

我有很多 kubelet 報告的錯誤日志：

Sep 07 09:43:51 kubenode-1 kubelet[10320]: I0907 09:43:51.651224   10320 container.go:369] Failed to update stats for container "/docker/01ad0eff434033752c1f39944e9965e38a07081fcbfe26dc35358bb63be18082": failed to read stat from "/sys/class/net/veth2fc2d33/statistics/rx_bytes" for device "veth2fc2d33", continuing to 
Sep 07 09:43:56 kubenode-1 kubelet[10320]: I0907 09:43:56.051022   10320 container.go:369] Failed to update stats for container "/": failed to read stat from "/sys/class/net/calic1976c4e52f/statistics/rx_bytes" for device "calic1976c4e52f", continuing to push stats

我不知道究竟是什么問題。 請在下面找到更多信息

docker info
Containers: 27
Images: 121
Storage Driver: aufs
 Root Dir: /var/lib/docker/aufs
 Backing Filesystem: extfs
 Dirs: 175
 Dirperm1 Supported: true
Execution Driver: native-0.2
Kernel Version: 3.16.0-4-amd64
Operating System: Debian GNU/Linux 8 (jessie)
CPUs: 2
Total Memory: 2.95 GiB
Name: kubenode-1
ID: LXO4:TD3E:ZAL5:AUWE:PN6W:KFZX:S4QR:AX6V:776M:VHVT:7Z3O:O72V
Username: smaine
Registry: [https://index.docker.io/v1/]

kubelet --version
Kubernetes v1.0.3

我在 debian jessie 上運行 kubernetes 集群。

此外，我有這些錯誤：

Sep 07 09:37:19 kubenode-1 kubelet[10320]: W0907 09:37:19.148453   10320 manager.go:1161] No ref for pod '814c1a33de45655e8cff2044485913ab568b9ab858ed2c5aa30d0034b82a6660'
Sep 07 09:37:29 kubenode-1 kubelet[10320]: W0907 09:37:29.265237   10320 manager.go:1161] No ref for pod '06a3e276f8b3dca0c3ea20b5feee4ab9b5ee97ef44aad1aef2f0102d5ddfa40c'
Sep 07 09:37:31 kubenode-1 kubelet[10320]: W0907 09:37:31.065126   10320 manager.go:1161] No ref for pod '686039f754799616a0957d3fcc516bc46b9601a2c6ac304a010397142fc06cd0'
Sep 07 09:37:31 kubenode-1 kubelet[10320]: W0907 09:37:31.075288   10320 kubelet.go:1343] Orphaned volume "c2f1c714-52f8-11e5-b108-5254000ec1d7/influxdb-persistent-storage" found, tearing down volume
Sep 07 09:37:31 kubenode-1 kubelet[10320]: W0907 09:37:31.914188   10320 kubelet.go:1343] Orphaned volume "c2f1c714-52f8-11e5-b108-5254000ec1d7/default-token-v2j1u" found, tearing down volume
Sep 07 09:38:49 kubenode-1 kubelet[10320]: E0907 09:38:49.224758   10320 kubelet.go:682] Image garbage collection failed: failed to find information for the filesystem labeled "docker-images"

我使用 calico 作為網絡覆蓋，但我認為它不相關。 如果您需要更多信息，請告訴我。

問候，

斯馬納

Answer 1

如果您使用 Docker v1.7 或更高版本，您可能會遇到 cadvisor 錯誤，其中網絡統計信息收集不正確。 見https://github.com/kubernetes/kubernetes/issues/13189

該錯誤已在 cadvisor 中修復，kubernetes 將很快包含修復。