search

redhat-developer-tooling / openshift-vagrant

All-in-One OpenShift Enterprise Vagrant setup
37 stars 32 forks source link

Increment the number of nodes #83

Open rafabene opened 8 years ago

rafabene commented 8 years ago

CDK should allow a parameterized way to define more than one node to be created so these nodes can be used by developers to test the application behaviour (in kubernetes and openshift) to see how an application will behave in case of node failure.

hferentschik commented 8 years ago

+1, there was already #58. This seems to be a duplicate. I have this on my todo list for quite some time. As you say, this feature would be very useful to test application behavior on for example node failure. One should be able to monitor how pods gets re-started on different nodes.

That said, having multiple ADB/CDK instances running on the same host might be prohibitive from a performance aspect. Needs to be tried. The other option is to allow somehow to "join" multiple instances across multiple hosts. Not sure whether this is possible though.

pmuir commented 8 years ago

Is this really a CDK use case? Or is it more of a "lab" environment by this point... Maybe the answer is to look at how the CDK and OpenShift Container Lab interact.

On 4 August 2016 at 08:57, Hardy Ferentschik notifications@github.com wrote:

+1, there was already #58 https://github.com/redhat-developer-tooling/openshift-vagrant/issues/58. This seems to be a duplicate. I have this on my todo list for quite some time. As you say, this feature would be very useful to test application behavior on for example node failure. One should be able to monitor how pods gets re-started on different nodes.

That said, having multiple ADB/CDK instances running on the same host might be prohibitive from a performance aspect. Needs to be tried. The other option is to allow somehow to "join" multiple instances across multiple hosts. Not sure whether this is possible though.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/redhat-developer-tooling/openshift-vagrant/issues/83#issuecomment-237480613, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJoQS_CU96BnO6zQeuk54XnQ5ehiwBDks5qcZtngaJpZM4Jaq_P .

hferentschik commented 8 years ago

Is this really a CDK use case?

I think so. In my world the CDK is about the ability to experience/experiment with all possible scenarios of OpenShift development. One big promise of OpenShift is transparent fail-over. As a developer I want to be able to see/test this. Being able to at least try it (even with bad performance), I personally would see as a big feature of the CDK.

Or is it more of a "lab" environment by this point... Maybe the answer is to look at how the CDK and OpenShift Container Lab interact.

What exactly is the "OpenShift Container Lab"?

pmuir commented 8 years ago

On 5 August 2016 at 10:59, Hardy Ferentschik notifications@github.com wrote:

Is this really a CDK use case?

I think so. In my world the CDK is about the ability to experience/experiment with all possible scenarios of OpenShift development. One big promise of OpenShift is transparent fail-over. As a developer I want to be able to see/test this. Being able to at least try it (even with bad performance), I personally would see as a big feature of the CDK.

Absolutely agreed it's a valuable thing to do. Just wondering if it should be part of the CDK (vs something like OpenShift Container Lab.

Or is it more of a "lab" environment by this point... Maybe the answer is to look at how the CDK and OpenShift Container Lab interact.

What exactly is the "OpenShift Container Lab"?

https://blog.openshift.com/introducing-red-hat-openshift-container-platform/

Not sure we have an offering available yet.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/redhat-developer-tooling/openshift-vagrant/issues/83#issuecomment-237811864, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJoQVdpQjjWdWV1uhaPC1eQ5ujPr-N8ks5qcwlngaJpZM4Jaq_P .

rafaeltuelho commented 8 years ago

I think this can be implemented following this approach: https://github.com/thoraxe/vagrant-openshift-multihost

I'm not sure if adding more nodes we need one to be the DNS server. If yes, IMHO, "CDK multinode" ca become too heavy/complicated...