Support for migrating docker-compose solutions to kubernetes

[kafnevod]

Hi
I'm using podman-compose to migrate docker-composite solutions to kubernetes via:

podman kube generate <podId> 

podman-compose uses the _ character as a delimiter when generating volume and container names.
For example, when starting a service stack
hello-python in POD:

# cd examples/hello-python
# podman-compose --in-pod hello-python up -d

we get a pod with container names hello-python_redis_1,``hello-python_web_1:

# podman pod ls --ctr-names
POD ID        NAME              STATUS      CREATED        INFRA ID    NAMES
6bc55962dc29  pod_hello-python  Running     2 minutes ago              hello-python_redis_1,hello-python_web_1

and volume name hello-python_redis

# podman volume ls | grep hello-python
hello-python_redis

After generating the kube- manifest

# podman kube generate pod_hello-python > ./kube.yml

and trying to run it,

# kubectl apply -f ./kube.yml

we receive an error message:

The Pod "podhello-python" is invalid: 
* spec.volumes[1].name: Invalid value: "hello-python_redis-pvc": a lowercase RFC 1123 label must consist of lower case alphanumeric characters or '-', and must start and end with an alphanumeric character (e.g. 'my-name',  or '123-abc', regex used for validation is '[a-z0-9]([-a-z0-9]*[a-z0-9])?')
* spec.containers[0].volumeMounts[0].name: Not found: "hello-python_redis-pvc"
...

This pull request supports the --kubecompat flag, which when specified uses the - separator character when generating names of containers and volumes.
When podman-compose called with a flag --kubecompat

podman-compose --kubecompat --in-pod hello-python up -d

we get a pod with container names hello-python-redis-1, hello-python-web-1:

# podman pod ls --ctr-names
POD ID        NAME              STATUS      CREATED         INFRA ID    NAMES
87c28523698a  pod_hello-python  Running     59 minutes ago              hello-python-redis-1,hello-python-web-1

and volume name hello-python-redis

# podman volume ls | grep hello-python
hello-python-redis

After generating the kube- manifest

# podman kube generate pod_hello-python > ./kube.yml

and trying to run it,

# kubectl apply -f ./kube.yml

we get a correctly deployed pod:

# kubectl apply -f ./kube.yml 
pod/podhello-python created
# kubectl get pod/podhello-python -o wide
NAME              READY   STATUS    RESTARTS   AGE   IP           NODE     NOMINATED NODE   READINESS GATES
podhello-python   2/2     Running   0          47s   10.88.0.21   host-8              

The reason for using the _ character as a delimiter is not obvious to me.
Maybe it's worth using the - character as a default separator?

[kafnevod]

In addition to using the - character as a separator for compatibility with kubernetes, the pull request lowercases container and volume names.

[p12tic]

Seems like useful addition to me.

[p12tic]

@kafnevod Is there a chance that you will use this feature in the future or was this just one off and you don't personally need it anymore?

[kafnevod]

@kafnevod Is there a chance that you will use this feature in the future or was this just one off and you don't personally need it anymore?

@p12tic

I don't need this feature in the future

I found another approach - creating a separate python application podman-compose-to-kube.
This script provides conversion of a docker-compose solution to kubernetes.
Conversion to kubernetes POD and kubernetes Deployment is currently supported.
Conversion to Helm is planned in the near future.