Thank you! It's exactly that I've been searching. On 08/26/2014 06:35 PM, Simon Pasquier wrote: > Hi, > IIUC you probably want to set||ram_weight_multiplier to a negative > number. From the OpenStack documentation [1]: > > By default, the scheduler spreads instances across all hosts > evenly. Set the |ram_weight_multiplier| option to a negative > number if you prefer stacking instead of spreading. Use a > floating-point value. > > > Simon > > [1] > http://docs.openstack.org/trunk/config-reference/content/section_compute-scheduler.html > > > > On Tue, Aug 26, 2014 at 5:01 PM, George Shuklin > <george.shuklin at gmail.com <mailto:george.shuklin at gmail.com>> wrote: > > Good day. > > I can't find any options for nova scheduler to consolidate new > instances on few hosts instead of 'spreading' them on all > available hosts. > > Simple example: let says we has 10 hosts, each host got 10Gb of > memory for instances. We has flavors of 3Gb and 5Gb of RAM. If we > run 20 new instances, they will consume about 6Gb per host and we > will not able to run new instance with 6Gb of RAM (even we have > 10*4=40 Gb of free memory on computes, none of the hosts has more > than 4Gb of memory). > > Is any nice way to say 'consolidate' to openstack? Thanks! > > _______________________________________________ > OpenStack-operators mailing list > OpenStack-operators at lists.openstack.org > <mailto:OpenStack-operators at lists.openstack.org> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140827/ae265b12/attachment.html>