Conjure-up uses double the deployed hardware when deploying openstack-base and kubernetes-core

[canonical-cgregan]

cs:kubernetes-core uses 2 physical machines to deploy the services for kubernetes. Conjure-up bootstraps one node for juju, then 4 more for the services. Not sure why this is the case

[battlemidget]

Thanks will look into this

[simonklb]

I think this could be the issue I'm facing as well, but with OpenStack base for MAAS. Rather than deploying all of the charms onto LXD machines, conjure up spins up one machine per charm. Ended up with ~20 servers running instead of 4... :)

[battlemidget]

@simonklb thanks we'll get this sorted

[mikemccracken]

@canonical-cgregan are you seeing this on a MAAS controller or something else? Curious if it's limited to MAAS - the add-machine handling path is different for MAAS vs. the rest.

[canonical-cgregan]

When using the Maas option in conjure-up is when I see this behavior. On Jan 12, 2017 21:34, "Michael McCracken" <notifications@github.com> wrote: @canonical-cgregan <https://github.com/canonical-cgregan> are you seeing this on a MAAS controller or something else? Curious if it's limited to MAAS - the add-machine handling path is different for MAAS vs. the rest. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#553 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APNP-1reUwNs_G60Yz7N3yGYNx2ZWFzNks5rRoAWgaJpZM4LghgN> .

[mikemccracken]

@canonical-cgregan thanks. I have a fix in progress, it should be committed today and will be available in our daily-git ppa for testing.