)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":5441,"name":"Andrew Laski","email":"andrew@lascii.com","username":"alaski"},"change_message_id":"f60ce8a1790c365bac74a3507db6d688644c0be0","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Andrew Laski \u003candrew@lascii.com\u003e"},{"line_number":5,"context_line":"CommitDate: 2016-09-12 12:42:32 +0000"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"libvirt: fix live-migration with serial console check"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"If the graphical consoles VNC and SPICE are disabled and the serial console"},{"line_number":10,"context_line":"is enabled, the live-migration of an instance will result in an ERROR state."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":8,"id":"9a89bdaa_4166aebb","line":7,"updated":"2016-09-13 14:44:21.000000000","message":"There was some discussion on this in the live-migration meeting http://eavesdrop.openstack.org/meetings/nova_live_migration/2016/nova_live_migration.2016-09-13-14.00.log.txt at 14:34:05. This patch is no longer a bug fix due to the revert at https://review.openstack.org/#/c/368732/2, but without this patch we can\u0027t revert the revert in O. Basically, we can\u0027t move the graphic/serial consoles check to pre_live_migration until the cycle after this code merges or live migrations break for \u003ccurrent cycle\u003e to \u003cprevious cycle\u003e.\n\nI would suggest updating the commit message away from being a bug fix and explain the upgrade rationale for this. But leave the bug link I think.","commit_id":"e31ec3a5478448b6a8d6344422247e9c148a2840"}],"nova/virt/libvirt/driver.py":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"83f639bb2a478e7e4d4e4f68fdd6d7d5d51a1b77","unresolved":false,"context_lines":[{"line_number":5415,"context_line":"            dest_check_data)"},{"line_number":5416,"context_line":"        migratable_flag \u003d self._host.is_migratable_xml_flag()"},{"line_number":5417,"context_line":"        if not migratable_flag or not listen_addrs:"},{"line_number":5418,"context_line":"            # In this context want to ensure we do not have to migrate"},{"line_number":5419,"context_line":"            # graphic or serial consoles since we can\u0027t update guest\u0027s"},{"line_number":5420,"context_line":"            # domain XML to make it handle destination host."},{"line_number":5421,"context_line":"            self._check_graphics_addresses_can_live_migrate(listen_addrs)"},{"line_number":5422,"context_line":""},{"line_number":5423,"context_line":"        serial_listen_addr \u003d libvirt_migrate.serial_listen_addr("}],"source_content_type":"text/x-python","patch_set":7,"id":"9a89bdaa_bd3c9630","line":5420,"range":{"start_line":5418,"start_character":0,"end_line":5420,"end_character":60},"updated":"2016-09-05 09:51:56.000000000","message":"That comment confuses me. That sounded that we were unable to update the domain XML for the specific serial information, hence it being prevented.\n\nGiven that you now fill in the necessary bits in L5375, I think that specifical rationale is now covered and we can safely allow serial consoled instances to be live migrated.\n\nMaybe a follow-up patch clarifying that would be helpful.","commit_id":"7326d38cefc30da91607d824f0ce758fdcc3dcff"}]}