)]}'
{"nova/virt/libvirt/driver.py":[{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"5163e9143ab143d05dce3b7b9ea6d9d69751315c","unresolved":false,"context_lines":[{"line_number":3479,"context_line":""},{"line_number":3480,"context_line":"        if not guest_cpu_numa_config:"},{"line_number":3481,"context_line":"            # No NUMA topology defined for instance - let the host kernel deal"},{"line_number":3482,"context_line":"            # with the NUMA effects."},{"line_number":3483,"context_line":"            # TODO(ndipanov): Attempt to spread the instance"},{"line_number":3484,"context_line":"            # across NUMA nodes and expose the topology to the"},{"line_number":3485,"context_line":"            # instance as an optimisation"}],"source_content_type":"text/x-python","patch_set":2,"id":"3afb71cf_c92d91a9","line":3482,"updated":"2015-06-09 10:14:57.000000000","message":"So, IIUC, with this change, when no hw_numa_nodes flavor/image property is set, the kernel is in full control. If admins want strict pinning, they can still have option of explicitly setting hw_numa_nods\u003d1 which does what this deleted code block would have done, but using the better logic in the NUMA scheduler filter.","commit_id":"eb253255d74cbe6e630225080c463f1eb16b9932"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"e8d23258036b4536e91032bd65e779cf999b2497","unresolved":false,"context_lines":[{"line_number":3480,"context_line":"        if not guest_cpu_numa_config:"},{"line_number":3481,"context_line":"            # No NUMA topology defined for instance - let the host kernel deal"},{"line_number":3482,"context_line":"            # with the NUMA effects."},{"line_number":3483,"context_line":"            # TODO(ndipanov): Attempt to spread the instance"},{"line_number":3484,"context_line":"            # across NUMA nodes and expose the topology to the"},{"line_number":3485,"context_line":"            # instance as an optimisation"},{"line_number":3486,"context_line":"            return GuestNumaConfig(allowed_cpus, None, None, None)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3afb71cf_fd600b03","line":3483,"updated":"2015-06-09 08:44:23.000000000","message":"Does this TODO is still valid?","commit_id":"eb253255d74cbe6e630225080c463f1eb16b9932"},{"author":{"_account_id":5511,"name":"Nikola Dipanov","email":"ndipanov@redhat.com","username":"ndipanov"},"change_message_id":"4fb4b60dbd118b9092311cb2b226703e2324d7af","unresolved":false,"context_lines":[{"line_number":3480,"context_line":"        if not guest_cpu_numa_config:"},{"line_number":3481,"context_line":"            # No NUMA topology defined for instance - let the host kernel deal"},{"line_number":3482,"context_line":"            # with the NUMA effects."},{"line_number":3483,"context_line":"            # TODO(ndipanov): Attempt to spread the instance"},{"line_number":3484,"context_line":"            # across NUMA nodes and expose the topology to the"},{"line_number":3485,"context_line":"            # instance as an optimisation"},{"line_number":3486,"context_line":"            return GuestNumaConfig(allowed_cpus, None, None, None)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3afb71cf_e8396bd8","line":3483,"in_reply_to":"3afb71cf_fd600b03","updated":"2015-06-09 08:47:56.000000000","message":"yeah sure - not that we will do it any time soon :) + it documents well what has been left out.","commit_id":"eb253255d74cbe6e630225080c463f1eb16b9932"}]}