Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-30 Thread Hongbin Lu
. Best regards, Hongbin From: Ton Ngo [mailto:t...@us.ibm.com] Sent: March-29-16 4:54 PM To: OpenStack Development Mailing List (not for usage questions) Subject: Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder In multiple occasions in the past, we have had to use

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Ton Ngo
with development, and gate tests if it makes sense. Ton Ngo, From: Yolanda Robla Mota <yolanda.robla-m...@hpe.com> To: <openstack-dev@lists.openstack.org> Date: 03/29/2016 01:35 PM Subject: Re: [openstack-dev] [magnum] Generate atomic images using diski

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Yolanda Robla Mota
So the advantages I can see with diskimage-builder are: - we reuse the same tooling that is present in other openstack projects to generate images, rather than relying on an external image - it improves the control we have on the contents of the image, instead of seeing that as a black box. At

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Adrian Otto
Steve, I will defer to the experts in openstack-infra on this one. As long as the image works without modifications, then I think it would be fine to cache the upstream one. Practically speaking, I do anticipate a point at which we will want to adjust something in the image, and it will be

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Steven Dake (stdake)
Adrian, Makes sense. Do the images have to be built to be mirrored though? Can't they just be put on the mirror sites fro upstream? Thanks -steve On 3/29/16, 11:02 AM, "Adrian Otto" wrote: >Steve, > >I¹m very interested in having an image locally cached in glance

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Adrian Otto
Steve, I’m very interested in having an image locally cached in glance in each of the clouds used by OpenStack infra. The local caching of the glance images will produce much faster gate testing times. I don’t care about how the images are built, but we really do care about the performance

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Steven Dake (stdake)
Yolanda, That is a fantastic objective. Matthieu asked why build our own images if the upstream images work and need no further customization? Regards -steve On 3/29/16, 1:57 AM, "Yolanda Robla Mota" wrote: >Hi >The idea is to build own images using

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Yolanda Robla Mota
Hi The idea is to build own images using diskimage-builder, rather than downloading the image from external sources. By that way, the image can live in our mirrors, and is built using the same pattern as other images used in OpenStack. It also opens the door to customize the images, using

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-29 Thread Mathieu Velten
Hi, We are using the official Fedora Atomic 23 images here (on Mitaka M1 however) and it seems to work fine with at least Kubernetes and Docker Swarm. Any reason to continue building specific Magnum image ? Regards, Mathieu Le mercredi 23 mars 2016 à 12:09 +0100, Yolanda Robla Mota a écrit : >

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-24 Thread Yolanda Robla Mota
Hi I have also been talking with diskimage-builder cores and they suggested we better start with the element on magnum. So I moved the change here: https://review.openstack.org/#/c/296719/ About diskimage-size, I'll dig a bit more on it. It can be for two causes: 1. To generate the

Re: [openstack-dev] [magnum] Generate atomic images using diskimage-builder

2016-03-23 Thread Ton Ngo
Hi Yolanda, Thank you for making a huge improvement from the manual process of building the Fedora Atomic image. Although Atomic does publish a public OpenStack image that is being considered in this patch: https://review.openstack.org/#/c/276232/ in the past we have come to many situations