HOWTO deploy ceph-nano on OpenShift 4.x

If you need S3 endpoints for Quay, Kafka etc in a small-scale OpenShift lab environment you can easily deploy ceph-nano to satisfy that need. Working example can be found from here.

Just follow the next steps and you are good to go. Instructions are tested with OCP 4.6 and 4.7.

Original instructions how to run ceph-nano on Kubernetes can be found from the project website.

It's good to be aware cn kube generates a broken yaml and it will take while to fix everything (indendations, doubles and so on)

1) Set Anyuid SCC

Follow next instructions: https://examples.openshift.pub/deploy/scc-anyuid/

2) Don't use default container image - Use:

registry.access.redhat.com/rhceph/rhceph-3-rhel7

3) Change Kubernetes Service type from NodePort to ClusterIP

4) Add one more environment variable

RGW_NAME - value must match with route host name

5) Login to OpenShift

oc login

6) Apply

oc apply -f deploy_ceph-nano.yml -n your-ns

7) Setup your S3-Client - For example s3cmd configuration should look like this:

s3cmd --configure and change following values:

New settings:
  Access Key: userkey
  Secret Key: secretkey
  Default Region: us-east-1
  S3 Endpoint: your.ceph.nano.host.name
  DNS-style bucket+hostname:port template for accessing a bucket: your.ceph.nano.host.name
  Use HTTPS protocol: True

8) Test it (example using s3cmd)

s3cmd mb s3://foobucket
date > test.txt
s3cmd put ./test.txt s3://foobucket
s3cmd ls s3://foobucket
s3cmd get s3://foobucket/test.txt froms3.txt