OpenShift: No DB is getting deployed

[gkarthiks]

So I followed the guide for deploying ArangoDB in the OpenShift Cluster. The CRD and the arango-deployment-operator has been created. When I ran the single-server deployment command, nothing happened in the cluster. But the CLI on the terminal shows its created.

Upon running the kubectl describe ArangoDeployment single-server command, it gives me the below output.

Name:         single-server
Namespace:    data-fabric
Labels:       <none>
Annotations:  kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"database.arangodb.com/v1alpha","kind":"ArangoDeployment","metadata":{"annotations":{},"name":"single-server","namespace":"data-fabric"},...
API Version:  database.arangodb.com/v1alpha
Kind:         ArangoDeployment
Metadata:
  Cluster Name:                   
  Creation Timestamp:             2018-04-27T05:33:06Z
  Deletion Grace Period Seconds:  <nil>
  Deletion Timestamp:             <nil>
  Resource Version:               73456034
  Self Link:                      /apis/database.arangodb.com/v1alpha/namespaces/data-fabric/arangodeployments/single-server
  UID:                            76a59a8a-49dc-11e8-b98d-fa163e3d2f07
Spec:
  Mode:  Single
Events:  <none>

followed by running the below command

kubectl get pods --selector=arango_deployment=single-server

gives No Resources found. Because nothing is created. Also interestingly the logs from the arango-deployment-operator prints the following as attached in the below log file.
arango-deployment-operator-1812060196-27kbj.log

version:
OpenShift: 3.7
Kubernetes: 1.7

[ewoutp]

From the logs, it looks that either the kubernetes cluster is not happy (log shows several connection errors to the apiserver).
The other possibility is that this kubernetes cluster is too old.
We've not tested on 1.7 and we assume a minimum of 1.8

[ewoutp]

We've tried running kube-arangodb on minishift.
It does not work yet. There is a problem with a user permission (when using all default settings) and minishift is indeed based on an "old" kubernetes version.

We'll investigate further and keep this issue open.

[gkarthiks]

@ewoutp Agreed! Because by default, all containers within OpenShift are blocked from “RunAsAny”, basically that they are not allowed to use a root user within the container.

We've not tested on 1.7 and we assume a minimum of 1.8

I will test in OpenShift v3.9 (k8s 1.9) may be.
But still, a quick patch or fix would be much appreciated. We are planning to take it to production in the cluster environment.

[gkarthiks]

Hi @ewoutp do we have any guesstimated date for the OpenShift deployment? That would be more helpful for us.

[gkarthiks]

Any updates on this?

[ewoutp]

We've found some issues with the standard docker image (arangodb/arangodb:).
It is switching to another user. We expect (but have not yet confirmed) that this is also causing issues on openshift.
If you want to try, please use the latest of this image arangodb/arangodb-preview:3.3.

[gkarthiks]

nice, let me try with the latest image @ewoutp

[ewoutp]

Progress is being made, see #144

[ewoutp]

With arangodb 3.3.10 & the latest operator (in master branch) this should be solved.

[gkarthiks]

Thanks, @ewoutp The DB is getting deployed successfully. But couldn't run the foxx UI.

I am trying to deploy the single-server and I am having two services. single-server-init and single-server. So I am creating a route for the single-server. But when I hit that URL. it says

502 Bad Gateway. Am i missing anything here?

getting thie below statement in the logs.

2018-06-09T06:48:17Z [1] ERROR {communication} unable to perform ssl handshake: http request : 336130204

[gkarthiks]

Sorry, it worked with the route and port together. Thanks @ewoutp !