On 2016-08-08 06:00:14 -0500 (-0500), [email protected] wrote: [...] > One solution we were thinking of is having a VM image that > contains the Docker image we need (e.g. saved using docker save), > thus bypassing the need to do "docker pull". Is that possible? I > realise that so far probably devstack is the only example making a > special tailored image.
It's less that we have a specially tailored image for DevStack, and more that our "standard" images pre-cache any downloads that DevStack claims it may use... basically we cache all of the $IMAGE_URLS list built up in http://git.openstack.org/cgit/openstack-dev/devstack/tree/stackrc as returned by the http://git.openstack.org/cgit/openstack-dev/devstack/tree/tools/image_list.sh script. Keeping the set of images we cache on our nodes trimmed to a limited number and only caching reasonably small images is important, however, since our base images are already quite large (in the neighborhood of 5GiB as compressed qcow2) and we don't want to bloat them further if we can help it. -- Jeremy Stanley __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: [email protected]?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
