)]}'
{"nova/virt/hardware.py":[{"author":{"_account_id":5511,"name":"Nikola Dipanov","email":"ndipanov@redhat.com","username":"ndipanov"},"change_message_id":"269e5bde7e30fc0b90af36eb811b7da7f06d6f15","unresolved":false,"context_lines":[{"line_number":769,"context_line":"    # itself on any NUMA cell"},{"line_number":770,"context_line":"    if (instance_cell.memory \u003e host_cell.memory or"},{"line_number":771,"context_line":"            len(instance_cell.cpuset) \u003e len(host_cell.cpuset)):"},{"line_number":772,"context_line":"        return None"},{"line_number":773,"context_line":""},{"line_number":774,"context_line":"    if instance_cell.cpu_pinning_requested:"},{"line_number":775,"context_line":"        new_instance_cell \u003d _numa_fit_instance_cell_with_pinning("}],"source_content_type":"text/x-python","patch_set":1,"id":"1a4dcd0f_c176f9d7","side":"PARENT","line":772,"updated":"2015-08-15 07:35:40.000000000","message":"Yeah this was discussed at the time and the idea was that it\u0027s OK to\nhave overcommit, but an instance larger than a NUMA node should _never_\nland on that NUMA noda, as it would effectively be overcommiting against\nitself.\n\nThis is not how overcommit on host level works - but it should probably get fixed there as it is questionable whether overcommitting an instance against itself makes sense.","commit_id":"05dbd82a10fd9ea1172769c71f2170e359560f59"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"95adbd1e33c3100aa9b2ee186da1fbbce34b6c2e","unresolved":false,"context_lines":[{"line_number":769,"context_line":"    # itself on any NUMA cell"},{"line_number":770,"context_line":"    if (instance_cell.memory \u003e host_cell.memory or"},{"line_number":771,"context_line":"            len(instance_cell.cpuset) \u003e len(host_cell.cpuset)):"},{"line_number":772,"context_line":"        return None"},{"line_number":773,"context_line":""},{"line_number":774,"context_line":"    if instance_cell.cpu_pinning_requested:"},{"line_number":775,"context_line":"        new_instance_cell \u003d _numa_fit_instance_cell_with_pinning("}],"source_content_type":"text/x-python","patch_set":1,"id":"1a4dcd0f_805a13f1","side":"PARENT","line":772,"in_reply_to":"1a4dcd0f_c176f9d7","updated":"2015-08-17 05:07:17.000000000","message":"Okay, I can see the logic behind that.  I also agree that the NUMA/host cases should get fixed up to use the same logic, whichever way we decide to go in the end.","commit_id":"05dbd82a10fd9ea1172769c71f2170e359560f59"}]}