It’s not a technical question :) The problem: The fact is CI need more resources, HP will quite and we need to find more ressources, right? As a provider, I can tell you that is not easy to provide ressources and I had to ask invests in our company to provide you what you need. As your needs are huge, the invest is high… But I would like to do more.
The solution: OpenStack CI could be base on available ressources at many providers (which means compute nodes not used or not full). You just need to manage when a spot instance is killed (because a customer pays for a resource) and restart the job elsewhere. The benefits: As providers we can give you more ressources, more quickly, more easily, without asking for a large budget to our companies. — Jean-Daniel Bonnetot http://www.ovh.com @pilgrimstack > Le 17 déc. 2015 à 16:08, Clint Byrum <[email protected]> a écrit : > > Excerpts from Jean-Daniel Bonnetot's message of 2015-12-17 01:24:33 -0800: >> Hi infra, >> >> You probably know that the foundation need more ressources for the CI. >> What do you think about pushing spot instances? >> It could be a great solution for our CI and make a smaller step for >> providers to give ressources « when they can ». >> >> https://review.openstack.org/#/c/104883/ > > Help me map this to infra's needs. Right now generally a quota is given, > and that's as much as a provider can let infra use for free. > > So to take advantage of this, spot instances would need to have either > 0 impact on one's quota, or a separate quota. Yes? The word quota does > not appear in the spec, so I'm curious to hear how that would be managed > in general, but especially for nodepool usage by infra. > > _______________________________________________ > OpenStack-Infra mailing list > [email protected] > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra _______________________________________________ OpenStack-Infra mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
