)]}'
{"nova/virt/hardware.py":[{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"9aa52af33edaf0c8eac13efc48bffc128d441d6d","unresolved":false,"context_lines":[{"line_number":2265,"context_line":"        network_metadata \u003d limits.network_metadata"},{"line_number":2266,"context_line":""},{"line_number":2267,"context_line":"    host_cells \u003d host_topology.cells"},{"line_number":2268,"context_line":"    random.shuffle(host_cells)"},{"line_number":2269,"context_line":""},{"line_number":2270,"context_line":"    # If PCI device(s) are not required, prefer host cells that don\u0027t have"},{"line_number":2271,"context_line":"    # devices attached. Presence of a given numa_node in a PCI pool is"}],"source_content_type":"text/x-python","patch_set":1,"id":"3f65232a_e6b6c7fa","line":2268,"range":{"start_line":2268,"start_character":4,"end_line":2268,"end_character":30},"updated":"2020-10-27 10:02:32.000000000","message":"Can you add a comment explaining why we\u0027re doing this?\n\nI\u0027m also not sure whether this is the best approach to take. Would it be better to spread work across cores by sorting from NUMA nodes with the least used cores to the most used cores?","commit_id":"e7f4dbeea69d144cf3babf3a815365cc3365e8cb"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f85f5ea8caac8d1faa797306059021d648d1cf4c","unresolved":false,"context_lines":[{"line_number":2265,"context_line":"        network_metadata \u003d limits.network_metadata"},{"line_number":2266,"context_line":""},{"line_number":2267,"context_line":"    host_cells \u003d host_topology.cells"},{"line_number":2268,"context_line":"    random.shuffle(host_cells)"},{"line_number":2269,"context_line":""},{"line_number":2270,"context_line":"    # If PCI device(s) are not required, prefer host cells that don\u0027t have"},{"line_number":2271,"context_line":"    # devices attached. Presence of a given numa_node in a PCI pool is"}],"source_content_type":"text/x-python","patch_set":1,"id":"3f65232a_fc0db635","line":2268,"range":{"start_line":2268,"start_character":4,"end_line":2268,"end_character":30},"in_reply_to":"3f65232a_798d3830","updated":"2020-10-27 12:02:43.000000000","message":"we should not shuffel the cells we shoudl be sorting them as explained in https://bugs.launchpad.net/nova/+bug/1893121","commit_id":"e7f4dbeea69d144cf3babf3a815365cc3365e8cb"},{"author":{"_account_id":17843,"name":"guolei","email":"guolei_yewu@cmss.chinamobile.com","username":"guolei"},"change_message_id":"2087b3e2299fb45fe254256fea99f9e699b6c2c4","unresolved":false,"context_lines":[{"line_number":2265,"context_line":"        network_metadata \u003d limits.network_metadata"},{"line_number":2266,"context_line":""},{"line_number":2267,"context_line":"    host_cells \u003d host_topology.cells"},{"line_number":2268,"context_line":"    random.shuffle(host_cells)"},{"line_number":2269,"context_line":""},{"line_number":2270,"context_line":"    # If PCI device(s) are not required, prefer host cells that don\u0027t have"},{"line_number":2271,"context_line":"    # devices attached. Presence of a given numa_node in a PCI pool is"}],"source_content_type":"text/x-python","patch_set":1,"id":"3f65232a_798d3830","line":2268,"range":{"start_line":2268,"start_character":4,"end_line":2268,"end_character":30},"in_reply_to":"3f65232a_e6b6c7fa","updated":"2020-10-27 11:58:57.000000000","message":"Ok, i will add more comments on launchpad, in fact, not just cores, memory will limit to numa node0 too.","commit_id":"e7f4dbeea69d144cf3babf3a815365cc3365e8cb"}]}