【发布时间】:2020-02-11 22:22:02
【问题描述】:
我刚刚将我的第一个集群从 minikube 移到了 AWS EKS。到目前为止,一切都很顺利,除了我认为遇到了一些 DNS 问题,但只是在一个集群节点上。
我在运行 v1.14 的集群中有两个节点,一种类型的 4 个 Pod,另一种类型的 4 个,每个工作 3 个,但每个工作 1 个 - 都在同一个节点上 - 启动然后错误 (CrashLoopBackOff)容器内的脚本出错,因为它无法解析数据库的主机名。删除出错的 pod,甚至删除所有 pod,每次都会导致同一节点上的一个 pod 失败。
数据库在自己的 pod 中并分配了服务,相同类型的其他 pod 没有解析名称或连接问题。数据库 pod 与无法解析主机名的 pod 位于同一节点上。我不确定如何将 pod 迁移到不同的节点,但这可能值得尝试看看问题是否随之而来。 coredns pod 中没有错误。我不确定从哪里开始寻找从这里发现问题的方法,如果有任何帮助或建议,我们将不胜感激。
提供以下配置。如前所述,它们都在 Minikube 上运行,并且它们都在一个节点上运行。
kubectl get pods - 注意年龄,同时删除所有 pod1 并重新创建自己,3 工作正常,4 没有。
NAME READY STATUS RESTARTS AGE
pod1-85f7968f7-2cjwt 1/1 Running 0 34h
pod1-85f7968f7-cbqn6 1/1 Running 0 34h
pod1-85f7968f7-k9xv2 0/1 CrashLoopBackOff 399 34h
pod1-85f7968f7-qwcrz 1/1 Running 0 34h
postgresql-865db94687-cpptb 1/1 Running 0 3d14h
rabbitmq-667cfc4cc-t92pl 1/1 Running 0 34h
pod2-94b9bc6b6-6bzf7 1/1 Running 0 34h
pod2-94b9bc6b6-6nvkr 1/1 Running 0 34h
pod2-94b9bc6b6-jcjtb 0/1 CrashLoopBackOff 140 11h
pod2-94b9bc6b6-t4gfq 1/1 Running 0 34h
postgresql 服务
apiVersion: v1
kind: Service
metadata:
name: postgresql
spec:
ports:
- port: 5432
selector:
app: postgresql
pod1 部署:
apiVersion: apps/v1
kind: Deployment
metadata:
name: pod1
spec:
replicas: 4
selector:
matchLabels:
app: pod1
template:
metadata:
labels:
app: pod1
spec:
containers:
- name: pod1
image: us.gcr.io/gcp-project-8888888/pod1:latest
env:
- name: rabbitmquser
valueFrom:
secretKeyRef:
name: rabbitmq-secrets
key: rmquser
volumeMounts:
- mountPath: /data/files
name: datafiles
volumes:
- name: datafiles
persistentVolumeClaim:
claimName: datafiles-pv-claim
imagePullSecrets:
- name: container-readonly
pod2 部署:
apiVersion: apps/v1
kind: Deployment
metadata:
name: pod2
spec:
replicas: 4
selector:
matchLabels:
app: pod2
template:
metadata:
labels:
app: pod2
spec:
containers:
- name: pod2
image: us.gcr.io/gcp-project-8888888/pod2:latest
env:
- name: rabbitmquser
valueFrom:
secretKeyRef:
name: rabbitmq-secrets
key: rmquser
volumeMounts:
- mountPath: /data/files
name: datafiles
volumes:
- name: datafiles
persistentVolumeClaim:
claimName: datafiles-pv-claim
imagePullSecrets:
- name: container-readonly
CoreDNS 配置映射以将 DNS 转发到外部服务(如果它没有在内部解析)。这是我认为会导致问题的唯一地方 - 但正如所说,它适用于一个节点。
apiVersion: v1
kind: ConfigMap
metadata:
name: coredns
namespace: kube-system
data:
Corefile: |
.:53 {
errors
health
kubernetes cluster.local in-addr.arpa ip6.arpa {
pods insecure
upstream
fallthrough in-addr.arpa ip6.arpa
}
prometheus :9153
proxy . 8.8.8.8
cache 30
loop
reload
loadbalance
}
错误的 Pod 输出。两个 pod 都一样,因为它出现在两者共有的库代码中。如前所述,并非所有 pod 都会出现这种情况,因此问题可能不在于代码。
Error connecting to database (psycopg2.OperationalError) could not translate host name "postgresql" to address: Try again
错误的 Pod1 描述:
Name: xyz-94b9bc6b6-jcjtb
Namespace: default
Priority: 0
Node: ip-192-168-87-230.us-east-2.compute.internal/192.168.87.230
Start Time: Tue, 15 Oct 2019 19:43:11 +1030
Labels: app=pod1
pod-template-hash=94b9bc6b6
Annotations: kubernetes.io/psp: eks.privileged
Status: Running
IP: 192.168.70.63
Controlled By: ReplicaSet/xyz-94b9bc6b6
Containers:
pod1:
Container ID: docker://f7dc735111bd94b7c7b698e69ad302ca19ece6c72b654057627626620b67d6de
Image: us.gcr.io/xyz/xyz:latest
Image ID: docker-pullable://us.gcr.io/xyz/xyz@sha256:20110cf126b35773ef3a8656512c023b1e8fe5c81dd88f19a64c5bfbde89f07e
Port: <none>
Host Port: <none>
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Wed, 16 Oct 2019 07:21:40 +1030
Finished: Wed, 16 Oct 2019 07:21:46 +1030
Ready: False
Restart Count: 139
Environment:
xyz: <set to the key 'xyz' in secret 'xyz-secrets'> Optional: false
Mounts:
/data/xyz from xyz (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-m72kz (ro)
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
xyz:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: xyz-pv-claim
ReadOnly: false
default-token-m72kz:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-m72kz
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning BackOff 2m22s (x3143 over 11h) kubelet, ip-192-168-87-230.us-east-2.compute.internal Back-off restarting failed container
错误的 Pod 2 描述:
Name: xyz-85f7968f7-k9xv2
Namespace: default
Priority: 0
Node: ip-192-168-87-230.us-east-2.compute.internal/192.168.87.230
Start Time: Mon, 14 Oct 2019 21:19:42 +1030
Labels: app=pod2
pod-template-hash=85f7968f7
Annotations: kubernetes.io/psp: eks.privileged
Status: Running
IP: 192.168.84.69
Controlled By: ReplicaSet/pod2-85f7968f7
Containers:
pod2:
Container ID: docker://f7c7379f92f57ea7d381ae189b964527e02218dc64337177d6d7cd6b70990143
Image: us.gcr.io/xyz-217300/xyz:latest
Image ID: docker-pullable://us.gcr.io/xyz-217300/xyz@sha256:b9cecdbc90c5c5f7ff6170ee1eccac83163ac670d9df5febd573c2d84a4d628d
Port: <none>
Host Port: <none>
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Wed, 16 Oct 2019 07:23:35 +1030
Finished: Wed, 16 Oct 2019 07:23:41 +1030
Ready: False
Restart Count: 398
Environment:
xyz: <set to the key 'xyz' in secret 'xyz-secrets'> Optional: false
Mounts:
/data/xyz from xyz (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-m72kz (ro)
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
xyz:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: xyz-pv-claim
ReadOnly: false
default-token-m72kz:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-m72kz
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning BackOff 3m28s (x9208 over 34h) kubelet, ip-192-168-87-230.us-east-2.compute.internal Back-off restarting failed container
【问题讨论】:
-
您必须提供有关您当前配置的更多信息,一些服务,部署 YAML 会非常有帮助。您可以将其添加到您的问题中。也请提供
kubectl describe pod <pod-name> -n <namespace>的错误输出。 -
谢谢@PjoterS - 尽我所能。
标签: kubernetes dns