Why Kubernetes pods sometimes got CrashLoopBackOff? - ruby-on-rails

Question

Using Kubernetes cluster: 3 hosts(1 master and 2 nodes).
Kubernetes version: 1.7
Deployed a Rails app to Kubernetes cluster.
Here is the deployment.yaml file:
apiVersion: v1
kind: Service
metadata:
name: server
labels:
app: server
spec:
ports:
- port: 80
selector:
app: server
tier: backend
type: LoadBalancer
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: server
labels:
app: server
spec:
replicas: 3
template:
metadata:
labels:
app: server
tier: backend
spec:
containers:
- image: 192.168.33.13/myapp/server
name: server
ports:
- containerPort: 3000
name: server
imagePullPolicy: Always
Deploy it:
$ kubectl create -f deployment.yaml
Then check the pods status:
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
server-962161505-kw3jf 0/1 CrashLoopBackOff 6 9m
server-962161505-lxcfb 0/1 CrashLoopBackOff 6 9m
server-962161505-mbnkn 0/1 CrashLoopBackOff 6 9m
At the beginning, its status got Completed but went to CrashLoopBackOff soon. Is there anything wrong in the config yaml file?
(By the way, I don't want to run a entrypoint.sh script here but used a job.yaml file to call k8s Job kind to do it.)
Edit
Result of kubectl describe pod server-962161505-kw3jf:
Name: server-962161505-kw3jf
Namespace: default
Node: node1/192.168.33.11
Start Time: Mon, 13 Nov 2017 17:45:47 +0900
Labels: app=server
pod-template-hash=962161505
tier=backend
Annotations: kubernetes.io/created-by={"kind":"SerializedReference","apiVersion":"v1","reference":{"kind":"ReplicaSet","namespace":"default","name":"server-962161505","uid":"0acadda6-c84f-11e7-84b8-02178ad2db9a","...
Status: Running
IP: 10.42.254.104
Created By: ReplicaSet/server-962161505
Controlled By: ReplicaSet/server-962161505
Containers:
server:
Container ID: docker://29eca3d9a20c60c83314101b036d742c5868c3bf25a39f28c5e4208bcdbfcede
Image: 192.168.33.13/myapp/server
Image ID: docker-pullable://192.168.33.13/myapp/server#sha256:0e056e3ff5b1f1084e0946bc4211d33c6f48bc06dba7e07340c1609bbd5513d6
Port: 3000/TCP
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Completed
Exit Code: 0
Started: Tue, 14 Nov 2017 10:13:12 +0900
Finished: Tue, 14 Nov 2017 10:13:13 +0900
Ready: False
Restart Count: 26
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from default-token-csjqn (ro)
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
Volumes:
default-token-csjqn:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-csjqn
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.alpha.kubernetes.io/notReady:NoExecute for 300s
node.alpha.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulMountVolume 22m kubelet, node1 MountVolume.SetUp succeeded for volume "default-token-csjqn"
Normal SandboxChanged 22m kubelet, node1 Pod sandbox changed, it will be killed and re-created.
Warning Failed 20m (x3 over 21m) kubelet, node1 Failed to pull image "192.168.33.13/myapp/server": rpc error: code = 2 desc = Error response from daemon: {"message":"Get http://192.168.33.13/v2/: dial tcp 192.168.33.13:80: getsockopt: connection refused"}
Normal BackOff 20m (x5 over 21m) kubelet, node1 Back-off pulling image "192.168.33.13/myapp/server"
Normal Pulling 4m (x7 over 21m) kubelet, node1 pulling image "192.168.33.13/myapp/server"
Normal Pulled 4m (x4 over 20m) kubelet, node1 Successfully pulled image "192.168.33.13/myapp/server"
Normal Created 4m (x4 over 20m) kubelet, node1 Created container
Normal Started 4m (x4 over 20m) kubelet, node1 Started container
Warning FailedSync 10s (x99 over 21m) kubelet, node1 Error syncing pod
Warning BackOff 10s (x91 over 20m) kubelet, node1 Back-off restarting failed container