qemu management interface (and libvirt)
- qemu standardizing guest enumeration itself
- fs w/ QMP sockets in well-known location per guest or qemud for enumeration
- QMP C library
- QMP for runtime guest mgmt, but not guest launch, storage pools, networking
- selinux, cgroups, tc, etc..done in libvirt, not in qemu, need to
enable qemu users
- libvirt is adding effectively raw QMP interface
- would be useful to launch w/ qemu directly and a later manage w/ libvirt
- some issues around flexible policy (svirt, cgroups, capabilities, uids)
- namespace is esp problematic if libvirtd and qemu don't have same
namesapce when libvirt connects later
- libvirt needs to support qemu as first class primary usecase
- libvirt already focuses on qemu
- QMP passthru is stopgap, important features should grow to proper API
- libvirt
- NetworkManager integration (netcf), possible...
- NetworkManager still needs to become aware of bridges, etc
- whole mess of options...need to move them out to extensible helper
scripts to simplify network setup
stable tree policy (push vs. pull and call for stable volunteers)
- historically pull based (cherry picking)
- doesn't scale wwell
- push based...submit patches directly to stable tree
- not just add "good for stable" in uptream patch submission
- looking for volunteers to manage stable tree
- Anthony volunteers!! (ok, not really)
block watermark
- concensus around polling
- other details to be worked out on list
- block layer ripe for plugins, good time for collecting requirements
and discussion on list
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
