Dan, One of the technical guys here at Cisco asked me a really good technical question about libvirt upgrades in containers which I was unable to answer. My suspicion is the linux VM system just sorts it out, but I wanted to get your input. Assume libvirt version 1 is running in a container. We kill the container, qemu processes go to the main pid space as detailed in my blog[1] using the —pid=host feature you developed. Now version #2 of qemu is started during an atomic upgrade (pull, stop, start of the container) to libvirt version 2. How does the operating system know to keep a copy of version #1 around until the system reboots? Thanks -steve [1] http://sdake.io/2015/01/28/an-atomic-upgrade-process-for-openstack-compute-nodes/ -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20150213/aad24800/attachment.html>