Kubernetes deployment not auto-terminating after successful run of command - kubernetes

Question

I have a Kubernetes cluster in which I've created a deployment to run a pod. Unfortunately after running it the pod does not want to self-terminate, instead it enters a continuous state of restart/CrashLoopBackOff cycle.
The command (on the entry point) runs correctly when first deployed, and I want it to run only one time.
I am programatically deploying the docker image with the entrypoint configured, using the Python K8s API. Here is my deployment YAML:
apiVersion: apps/v1
kind: Deployment
metadata:
name: kio
namespace: kmlflow
labels:
app: kio
name: kio
spec:
replicas: 1
selector:
matchLabels:
app: kio
name: kio
template:
metadata:
labels:
app: kio
name: kio
spec:
containers:
- name: kio-ingester
image: MY_IMAGE
command: ["/opt/bin/kio"]
args: ["some", "args"]
imagePullPolicy: Always
restart: Never
backofflimit: 0
Thanks for any help
Output from kubectl pod is:
Name: ingest-160-779874b676-8pgv5
Namespace: kmlflow
Priority: 0
PriorityClassName: <none>
Node: 02-w540-02.glebe.kinetica.com/172.30.255.205
Start Time: Thu, 11 Oct 2018 13:31:20 -0400
Labels: app=kio
name=kio
pod-template-hash=3354306232
Annotations: <none>
Status: Running
IP: 10.244.0.228
Controlled By: ReplicaSet/ingest-160-779874b676
Containers:
kio-ingester:
Container ID: docker://b67a682d04e69c2dc5c1be7e02bf2e4cf7a12a7557dfbe642dfb531ca4b03f07
Image: kinetica/kinetica-intel
Image ID: docker-pullable://docker.io/kinetica/kinetica-intel#sha256:eefbb6595eb71822300ef97d5cbcdac7ec58f2041f8190d3a2ba9cffd6a0d87c
Port: <none>
Host Port: <none>
Command:
/opt/gpudb/bin/kio
Args:
--source
kinetica://172.30.50.161:9191::dataset_iris
--destination
kinetica://172.30.50.161:9191::iris5000
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Completed
Exit Code: 0
Started: Thu, 11 Oct 2018 13:33:27 -0400
Finished: Thu, 11 Oct 2018 13:33:32 -0400
Ready: False
Restart Count: 4
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from default-token-69wkn (ro)
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
default-token-69wkn:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-69wkn
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
---- ------ ---- ---- -------
Normal Scheduled 2m39s default-scheduler Successfully assigned kmlflow/ingest-160-779874b676-8pgv5 to 02-w540-02.glebe.kinetica.com
Normal Created 89s (x4 over 2m28s) kubelet, 02-w540-02.glebe.kinetica.com Created container
Normal Started 89s (x4 over 2m28s) kubelet, 02-w540-02.glebe.kinetica.com Started container
Warning BackOff 44s (x7 over 2m15s) kubelet, 02-w540-02.glebe.kinetica.com Back-off restarting failed container
Normal Pulling 33s (x5 over 2m28s) kubelet, 02-w540-02.glebe.kinetica.com pulling image "kinetica/kinetica-intel"
Normal Pulled 33s (x5 over 2m28s) kubelet, 02-w540-02.glebe.kinetica.com Successfully pulled image "kinetica/kinetica-intel"
There is no output from Kubectl logs <crashing-pod> because a successful run of the command KIO with the injected parameters does not print anything to standard output.

Answer 1

If you'd like to run your task one time and finish after a successful completion you should consider using Kubernetes Jobs or CronJobs
Something like this:
apiVersion: batch/v1
kind: Job
metadata:
name: kio
namespace: kmlflow
labels:
app: kio
name: kio
spec:
template:
metadata:
labels:
app: kio
name: kio
spec:
containers:
- name: kio-ingester
image: MY_IMAGE
command: ["/opt/bin/kio"]
args: ["some", "args"]
imagePullPolicy: Always
restart: Never
backoffLimit: 4
To delete the jobs automatically if you have Kubernetes 1.12 or later you can use ttlSecondsAfterFinished. Unfortunately, you if you are using Kuberbetes 1.11 or earlier you will have to delete them manually or you can set up a CronJob to do it.