I'm off now for the holiday period, so I'm sharing what I'm sitting on. This represents a bit over a week of staring at code, and a couple of days of actual keyboard banging. It doesn't run. Details are in the commit message, including where I think I'm going with this. Read the commit message like a slightly more detailed spec.
https://review.openstack.org/259148 I have a reasonable idea of where I want to go with this, but finding a consumable set of steps to get there is challenging. The current code makes lots of assumptions about storage in lots of places. Consequently, adding to it without regressing is becoming increasingly difficult. Making the wholesale change to libvirt storage pools, supporting multiple layouts in the transition period, would require updating these assumptions everywhere they exist in the code. My initial goal is to centralise these assumptions whilst making minimal/no functional changes. Once all the code which assumes on-disk layout is in one place, we can be more confident in making changes to it, and more easily validate it. We can also more easily see what the current assumptions are. This will also make it easier for people to continue adding storage-related features to the libvirt driver. Matt -- Matthew Booth Red Hat Engineering, Virtualisation Team Phone: +442070094448 (UK)
__________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
