Specifically, users can opt to manage their deployment by specifying a custom docker image. OpenShift will provision a pod using that custom image that performs the actual deployment and can deal with these scenarios.
"Clayton Coleman" <ccoleman redhat com>To:
ejacobs redhat comCc:
"Abhishek Gupta" <abhgupta redhat com>, "Openshift Dev" <dev lists openshift redhat com>Sent:
Monday, February 23, 2015 1:48:17 PMSubject:
Re: "provisioning" external endpoint
Post 3.0, or as an integration. I think it's valuable, and when we design core interfaces they should support this mechanism on top of rock solid fundamentals. The bulk of this work has to happen in Kube first then we can come back and add the provisioning story.
However, pending that, it should be possible to use our APIs to accomplish the same goal.
> On Feb 23, 2015, at 4:39 PM, Erik M Jacobs <ejacobs redhat com> wrote:
> Hi Abhishek,
>> On 02/23/2015 04:31 PM, Abhishek Gupta wrote:
>> Provisioning an "external" endpoint is outside the scope of this
>> feature. In most cases, OpenShift will not have control/access to
>> provision the external endpoint.
> Why do you feel that to be the case? I can think of a large pile of
> organizations that wanted to do exactly this with V2. For example, the
> following user story is pervasive:
> * given an existing database cluster
> * as a user I want to
> * add a database to my application which
> * creates a user and database and etc. on that existing database cluster
>> This feature merely talks about
>> OpenShift being able to configure/create the endpoint 'manually' for an
>> external service.
> I understand what the current feature description does. That's why I
> asked if a different feature, provisioning the endpoint, was on the map
>> *From: *"Erik M Jacobs" <ejacobs redhat com>
>> *To: *"Openshift Dev" <dev lists openshift redhat com>
>> *Sent: *Monday, February 23, 2015 1:22:12 PM
>> *Subject: *"provisioning" external endpoint
>> Hi all,
>> I just saw this:
>> Which is super duper awesome.
>> But, conceptually, I don't think that kubernetes has a "pluggability" to
>> its workflow where we can actually provision that endpoint.
>> For example, this would be great if we wanted to talk to an Amazon RDS
>> instance, but it will only work if that instance already exists.
>> Is there any thought about enabling the provisioning of the endpoint as
>> part of this process? Maybe some modification/enhancement to
>> deploymentconfig or something?
>> Erik M Jacobs, RHCA
>> Principal Technical Marketing Manager, OpenShift Enterprise
>> Red Hat, Inc.
>> Phone: 646.462.3745
>> Email: ejacobs redhat com
>> AOL Instant Messenger: ejacobsatredhat
>> Twitter: @ErikonOpen
>> Freenode: thoraxe
>> dev mailing list
>> dev lists openshift redhat com
> dev mailing list
> dev lists openshift redhat com