I have an application that has 3 pods and each pod needs a fixed variable name stored in each pod. So if everything is running fine, the three pods would have var1, var2, and var3 stored on the corresponding pods.
If the first pod gets replaced which has var1, how can I determine that the other 2 pods have var2 and var3, and thus know that the new pod should be assigned var1?
Can this be done with Stateful Sets?
I see two ways of doing that:
Using StatefulSets:
For a StatefulSet with N replicas, each Pod in the StatefulSet will be
assigned an integer ordinal, from 0 up through N-1, that is unique
over the Set.
Creating the Pods manually. Example:
apiVersion: v1
kind: Pod
metadata:
name: memory-demo-3
namespace: mem-example
spec:
containers:
- name: memory-demo-3-ctr
image: polinux/stress
If you need your application to be aware of the Pod where it is running on, there is an interesting page in Kubernetes documentation: "Expose Pod Information to Containers Through Environment Variables".
Example:
apiVersion: v1
kind: Pod
metadata:
name: mypod-var1
spec:
containers:
- name: test-container
image: k8s.gcr.io/busybox
env:
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
Using a StatefulSet you can extract this from the pod-name.
env:
- name: podname
valueFrom:
fieldRef:
fieldPath: metadata.name
and then get it from the end of the name. The pods in a StatefulSet will be named <StatfulSetName>-<ordinal>, see pod-identity
Related
I want to run Python code inside a pod. The pod is created by airflow that I don't control.
I want to somehow get the name of the pod I'm running in.
How can it be done?
You can tell kuberenetes to mount an env variable for you:
apiVersion: v1
kind: Pod
metadata:
name: dapi-test-pod
spec:
containers:
- name: test-container
image: gcr.io/google_containers/busybox
command: [ "/bin/sh", "-c", "env" ]
env:
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
and then in python you can access it like:
import os
pod_name = os.environ['MY_POD_NAME']
Or you can just open and read /etc/hostname:
f = open('/etc/hostname')
pod_name = f.read()
f.close()
Exposing Pod and Cluster Vars to Containers
Lets say you need some data about Pod or K8s environment in your application to add Pod informnation as metada tp logs. such as e.g.
Pod IP
Pod Name
Service Account of Pod
NOTE: All Pod information can be made available in the config file.
There are 2 ways to expose Pod fields into a running Container:
Environment Variables
Volume Files
Example of Environment Variable
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-deployment-env
labels:
app: nginx
spec:
replicas: 3
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
ports:
- containerPort: 80
- name: log-sider
image: busybox
command: [ 'sh', '-c' ]
args:
- while true; do
echo sync app logs;
printenv POD_NAME POD_IP POD_SERVICE_ASCCOUNT;
sleep 20;
done;
env:
- name: POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
- name: POD_SERVICE_ASCCOUNT
valueFrom:
fieldRef:
fieldPath: spec.serviceAccountName
Kubernets Docs
Source
one way can be
So in the Kubernetes cluster you are operating in do
kubectl get pods
now see the yaml of all pods by
oc get pods <pod-name> -o yaml
then in that find the container images being used by the pods .
identify the image tag that belongs to your container creation.
That means that when you build you image , image has a name and a tag , which is further pushed to some cloud hub from where your pod will pull the image and start the container .
you need to find the image tag and name in pod yaml using the above commands given .
Try the below :
# List all pods in all namespaces
kubectl get pods --all-namespaces
# List all pods in the current namespace
kubectl get pods -o wide
Then u can see more details using the below :
kubectl describe pod <pod-name>
Also you can refer to the following stackoverflow question and the related answers.
get-current-pod
I'm trying to figure it out, how to change one string inside configmap in Kubernetes. I have pretty simple configmap:
apiVersion: v1
data:
config.cfg: |-
[authentication]
USERNAME=user
PASSWORD=password
[podname]
PODNAME=metadata.podName
kind: ConfigMap
metadata:
name: name_here
And I need to mount the configmap inside a couple of pods. But PODNAME should be matched to current podname. Is it possible in any another way? thanks!
I do not think it could be done with ConfigMap. But you can set environment variables in your pod spec that references a pod fields.
apiVersion: v1
kind: Pod
metadata:
name: test-ref-pod-name
spec:
containers:
- name: test-container
image: busybox
command: [ "sh", "-c"]
args:
- env | grep PODNAME
env:
- name: PODNAME
valueFrom:
fieldRef:
fieldPath: metadata.name
restartPolicy: Never
See official documentation: https://kubernetes.io/docs/tasks/inject-data-application/environment-variable-expose-pod-information/#use-pod-fields-as-values-for-environment-variables
This doesn't answer you question exactly but the pod name normally ends up as the host name inside the pod and can be accessed as a standard environment variable
echo $HOSTNAME
I have a statefulset and I need to know what is the current replica count from inside the pod. To do so, I tried:
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: sample-mariadb
namespace: demo
spec:
replicas: 3
template:
spec:
containers:
- env:
- name: REPLICAS
valueFrom:
fieldRef:
apiVersion: v1
fieldPath: spec.replicas
and got this error:
Warning FailedCreate 4m15s (x17 over 9m43s) statefulset-controller create Pod sample-mariadb-0 in StatefulSet sample-mariadb failed error: Pod "sample-mariadb-0" is invalid: spec.containers[1].env[3].valueFrom.fieldRef.fieldPath: Invalid value: "spec.replicas": error converting fieldPath: field label not supported: spec.replicas
How can I get current replica count from inside the pod?
You can only send the fields that are part of Pod specification. The spec.replicas field is part of StatefulSet specification not the underlying pod's. The template part of a StatefulSet is the pod specification. Hence, you are getting this error.
I have looked for a solution, and the alternatives I could find is:
Setting the same environment variable
- name: REPLICAS
value: "3"
Or you can update the value of the replicas as template variable. Example updating it as template.yaml:
replicas: ${num_replicas}
spec:
- name: REPLICAS
value: ${num_replicas}
export num_replicas=3
cat template.yaml | envsubst > dapi-stateful.yaml
I'm currently using Kubernetes to schedule a DaemonSet on both master and worker nodes.
The DaemonSet definition is the same for both node types (same image, same volumes, etc), the only difference is that when the entrypoint is executed, I need to write a different configuration file (which is generated in Python with some dynamic values) if the node is a master or a worker.
Currently, to overcome this I'm using two different DaemonSet definitions with an env value which tells if the node is a master or not. Here's the yaml file (only relevant parts):
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
name: worker-ds
namespace: kube-system
labels:
k8s-app: worker
spec:
...
spec:
hostNetwork: true
containers:
- name: my-image
...
env:
- name: NODE_IP
valueFrom:
fieldRef:
fieldPath: status.hostIP
- name: IS_MASTER
value: "false"
...
---
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
name: master-ds
namespace: kube-system
labels:
k8s-app: master
spec:
...
spec:
hostNetwork: true
nodeSelector:
node-role.kubernetes.io/master: ""
tolerations:
- key: node-role.kubernetes.io/master
operator: Exists
effect: NoSchedule
containers:
- name: my-image
...
env:
- name: NODE_IP
valueFrom:
fieldRef:
fieldPath: status.hostIP
- name: IS_MASTER
value: "true"
...
However, since the only difference is the IS_MASTER value, I want to collapse both the definitions in a single one that programmatically understands if the current node where the pod is being scheduled is a master or a worker.
Is there any way to know this information about the node programmatically (even reading a configuration file [for example something that only the master has or viceversa] in the node or something like that)?
Thanks in advance.
Unfortunately, there is not a convenient way to access the node information in pod.
If you only want a single DaemonSet definition, you can add a sidecar container to your pod, the sidecar container can access the k8s api, then your main container can get something useful from the sidecar.
By the way, I think your current solution is properly :)
You can tell a node is the master if it has the label node-role.kubernetes.io/master: "".
What you need to do is access that label from your containers which can be done with the Downward Api (Edit: Wrong, only Pod information can be accessed from the Downward Api). You can mount the labels inside your containers using:
volumes:
- name: podinfo
downwardAPI:
items:
- path: "labels"
fieldRef:
fieldPath: metadata.labels
You can then search the content of that file from within the container.
How can I to source the Kubernetes deployment/job name that spawned the current pod from within the pod?
In many cases the hostname of the Pod equals to the name of the Pod (you can access that by the HOSTNAME environment variable). However that's not a reliable method of determining the Pod's identity.
You will want to you use the Downward API which allows you to expose metadata as environment variables and/or files on a volume.
The name and namespace of a Pod can be exposed as environment variables (fields: metadata.name and metadata.namespace) but the information about the creator of a Pod (which is the annotation kubernetes.io/created-by) can only be exposed as a file.
Example:
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: busybox
labels: {app: busybox}
spec:
selector: {matchLabels: {app: busybox}}
template:
metadata: {labels: {app: busybox}}
spec:
containers:
- name: busybox
image: busybox
command:
- "sh"
- "-c"
- |
echo "I am $MY_POD_NAME in the namespace $MY_POD_NAMESPACE"
echo
grep ".*" /etc/podinfo/*
while :; do sleep 3600; done
env:
- name: MY_POD_NAME
valueFrom: {fieldRef: {fieldPath: metadata.name}}
- name: MY_POD_NAMESPACE
valueFrom: {fieldRef: {fieldPath: metadata.namespace}}
volumeMounts:
- name: podinfo
mountPath: /etc/podinfo/
volumes:
- name: podinfo
downwardAPI:
items:
- path: "labels"
fieldRef: {fieldPath: metadata.labels}
- path: "annotations"
fieldRef: {fieldPath: metadata.annotations}
Too see the output:
$ kubectl logs `kubectl get pod -l app=busybox -o name | cut -d / -f2`
Output:
I am busybox-1704453464-m1b9h in the namespace default
/etc/podinfo/annotations:kubernetes.io/config.seen="2017-02-16T16:46:57.831347234Z"
/etc/podinfo/annotations:kubernetes.io/config.source="api"
/etc/podinfo/annotations:kubernetes.io/created-by="{\"kind\":\"SerializedReference\",\"apiVersion\":\"v1\",\"reference\":{\"kind\":\"ReplicaSet\",\"namespace\":\"default\",\"name\":\"busybox-1704453464\",\"uid\":\"87b86370-f467-11e6-8d47-525400247352\",\"apiVersion\":\"extensions\",\"resourceVersion\":\"191157\"}}\n"
/etc/podinfo/annotations:kubernetes.io/limit-ranger="LimitRanger plugin set: cpu request for container busybox"
/etc/podinfo/labels:app="busybox"
/etc/podinfo/labels:pod-template-hash="1704453464"
If you are using the Downwards API to get deployment name from inside the pod, and you want to avoid using the volume mount way - there is one opinionated way to get deployment info, exposed to pod as environment variables.
Template labels specified in a Deployment spec are added as pod labels to each pod of that deployment.
Example : the app label below will be added to all pods of this deployment
...
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx
labels:
app: nginx
spec:
replicas: 3
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
...
It is a commonly followed (again, not necessarily true for your case) convention for deployments, to keep the app label value same as the the deployment name, as shown in the above example. If your deployments follow this convention (mine did), you can expose this label's value (essentially, the name of deployment) as an environment variable to the pod, using the downwards API
Continuing on above example :
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx
labels:
app: nginx
spec:
replicas: 3
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- containerPort: 80
env:
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: metadata.namespace
- name: DEPLOYMENT_NAME
valueFrom:
fieldRef:
fieldPath: metadata.labels['app']
Again, clarifying that this is not a guaranteed solution for your problem as it still does not give the deployment name in env vars. It is just an opinionated way which I found useful and thought would be good to share.
In my case, there were a lot of deployments (>20) and I didn't want to add the deployment name manually as an env variable, for each of the deployment config. As my deployments already followed the above convention, I just copied the bit of yaml specifying NAMESPACE and DEPLOYMENT_NAME variable to each deployment config
references :
https://kubernetes.io/docs/tasks/inject-data-application/downward-api-volume-expose-pod-information/#capabilities-of-the-downward-api