Hi Conrad, We boot to ephemeral disk by default but our ephemeral disk is Ceph RBD just like out cinder volumes.
Using Ceph for Cinder Volumes and Glance Images storage it is possible to very quickly create new Persistent Volumes from Glance Images becasue on the backend it's just a CoW snapshot operation (even though we use seperate pools for ephemeral disk, persistent volumes, and images). This is also what happens for ephemeral boting which is much faster than copying image to local disk on hypervisor first so we get quick starts and relatively easy live migrations (which we use for maintenance like hypervisor reboots and reinstalls). I don't know how to make it the "default" but ceph definately makes it faster. Other backends I've used basicly mount the raw storage volume download the imaage then 'dd' in into place which is painfully slow. As to why ephemeral rather than volume backed by default it's much easier to boot amny copies of the same thing and be sure they're the same using ephemeral storage and iamges or snapshots. Volume backed instances tend to drift. That said workign in a research lab many of my users go for the more "Pet" like persistent VM workflow. We just manage it with docs and education, though there is always someone who misses the red flashing "ephemeral means it gets deleted when you turn it off" sign and is sad. -Jon On Tue, Aug 01, 2017 at 02:50:45PM +0000, Kimball, Conrad wrote: :In our process of standing up an OpenStack internal cloud we are facing the question of ephemeral storage vs. Cinder volumes for instance root disks. : :As I look at public clouds such as AWS and Azure, the norm is to use persistent volumes for the root disk. AWS started out with images booting onto ephemeral disk, but soon after they released Elastic Block Storage and ever since the clear trend has been to EBS-backed instances, and now when I look at their quick-start list of 33 AMIs, all of them are EBS-backed. And I'm not even sure one can have anything except persistent root disks in Azure VMs. : :Based on this and a number of other factors I think we want our user normal / default behavior to boot onto Cinder-backed volumes instead of onto ephemeral storage. But then I look at OpenStack and its design point appears to be booting images onto ephemeral storage, and while it is possible to boot an image onto a new volume this is clumsy (haven't found a way to make this the default behavior) and we are experiencing performance problems (that admittedly we have not yet run to ground). : :So ... : :* Are other operators routinely booting onto Cinder volumes instead of ephemeral storage? : :* What has been your experience with this; any advice? : :Conrad Kimball :Associate Technical Fellow :Chief Architect, Enterprise Cloud Services :Application Infrastructure Services / Global IT Infrastructure / Information Technology & Data Analytics :[email protected]<mailto:[email protected]> :P.O. Box 3707, Mail Code 7M-TE :Seattle, WA 98124-2207 :Bellevue 33-11 bldg, office 3A6-3.9 :Mobile: 425-591-7802 : :_______________________________________________ :OpenStack-operators mailing list :[email protected] :http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators -- _______________________________________________ OpenStack-operators mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
