)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"694d023c90b95c8d3e7b93ee01dc375843cf8587","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This test attempts to list all instances, then list them again with the"},{"line_number":10,"context_line":"first instance as the marker and ensure that the remaining instances"},{"line_number":11,"context_line":"are returned in the page. Now that we are doing queries to cells in"},{"line_number":12,"context_line":"parallel, consecutive unsorted list queres can return things in"},{"line_number":13,"context_line":"different orders as the cells may reply at different times. The fix"},{"line_number":14,"context_line":"in this patch is to ask for results to be sorted, which is the only way"},{"line_number":15,"context_line":"it makes sense. It worked before purely because we were always scanning"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"9f5c4f37_51da4de9","line":12,"range":{"start_line":12,"start_character":36,"end_line":12,"end_character":42},"updated":"2017-09-21 16:34:28.000000000","message":"queries","commit_id":"6143b1db85f012e682f437040bfc04cc44291293"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"694d023c90b95c8d3e7b93ee01dc375843cf8587","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":5,"id":"9f5c4f37_d1ef5d85","line":19,"updated":"2017-09-21 16:34:28.000000000","message":"I guess this should have:\n\n  Related-Bug: #1689692\n\n?","commit_id":"6143b1db85f012e682f437040bfc04cc44291293"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"68ad22de02fd1474d33c4b911babc83ddf8d35be","unresolved":false,"context_lines":[{"line_number":10,"context_line":"first instance as the marker and ensure that the remaining instances"},{"line_number":11,"context_line":"are returned in the page. Now that we are doing queries to cells in"},{"line_number":12,"context_line":"parallel, consecutive unsorted list queries can return things in"},{"line_number":13,"context_line":"different orders as the cells may reply at different times. The fix"},{"line_number":14,"context_line":"in this patch is to ask for results to be sorted, which is the only way"},{"line_number":15,"context_line":"it makes sense. It worked before purely because we were always scanning"},{"line_number":16,"context_line":"the cells linearly and in the same order."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Related-Bug: #1689692"},{"line_number":19,"context_line":"Change-Id: I3ca2a167c902d565c36a5d5dbba1bf1c214aa20b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":7,"id":"7f515b1d_94ec142a","line":16,"range":{"start_line":13,"start_character":60,"end_line":16,"end_character":41},"updated":"2017-09-25 13:37:37.000000000","message":"Well, by default the API sorts by created_at descending:\n\nhttps://github.com/openstack/nova/blob/ef0f6341f45257db68147798b052b1addcee4cbb/nova/api/openstack/common.py#L142\n\nSo this wasn\u0027t totally random before, and I\u0027m not sure it\u0027s accurate to say \"consecutive unsorted list queries\" is what was happening before this change.\n\n\"we were always scanning the cells linearly and in the same order\" is accurate though, sure. If we had shuffled nova.compute.api.CELLS that could have broken this test.","commit_id":"8faf8b3f1f041b515de1958c426b0a87c41a4061"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"3f9624837c1bec88de58b3334b87c9d3b7a314f8","unresolved":false,"context_lines":[{"line_number":10,"context_line":"first instance as the marker and ensure that the remaining instances"},{"line_number":11,"context_line":"are returned in the page. Now that we are doing queries to cells in"},{"line_number":12,"context_line":"parallel, consecutive unsorted list queries can return things in"},{"line_number":13,"context_line":"different orders as the cells may reply at different times. The fix"},{"line_number":14,"context_line":"in this patch is to ask for results to be sorted, which is the only way"},{"line_number":15,"context_line":"it makes sense. It worked before purely because we were always scanning"},{"line_number":16,"context_line":"the cells linearly and in the same order."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Related-Bug: #1689692"},{"line_number":19,"context_line":"Change-Id: I3ca2a167c902d565c36a5d5dbba1bf1c214aa20b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":7,"id":"7f515b1d_03e7c442","line":16,"range":{"start_line":13,"start_character":60,"end_line":16,"end_character":41},"in_reply_to":"7f515b1d_54e0dcfa","updated":"2017-09-25 15:30:32.000000000","message":"\u003e It wasn\u0027t before\n\nDo you mean your series wasn\u0027t before? Confused about what \u0027it\u0027 is here.\n\n\u003e Since I can\u0027t sort on insert order across cells, you\u0027d get unsorted batches which depended on the order in which the cell threads returned.\n\nBut you can sort on created_at, right? Which is the default in the API. I might just not understand what you\u0027re saying.\n\n\u003e Now that I\u0027m forcing an order in instance_list, we could probably drop this patch since we\u0027re always sorting all the time.\n\nRight, your code is now doing what the API was doing by default before, which is good I think (for backward compatibility and all).\n\n\u003e However, it\u0027s doing a thing that depends on a sorted result, so it does\u0027t seem terrible to make it explicit.\n\nIf it was working based on default sort order via created_at, I think it\u0027s OK to leave it as-is since it hasn\u0027t been a problem.\n\nRegarding API contract, reading:\n\nhttps://developer.openstack.org/api-ref/compute/#list-servers\n\nThe sort_keys parameter says:\n\n\"Sorts by a server attribute. Default attribute is created_at.\"\n\nSo we already say we\u0027re going to sort by created_at by default in our API docs, so I think it\u0027s OK to have a test which somewhat relies on that, even if it wasn\u0027t intentional when I wrote the test.","commit_id":"8faf8b3f1f041b515de1958c426b0a87c41a4061"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"4a0f9abaec99189ad51df23477fd7b8e3b0fc9eb","unresolved":false,"context_lines":[{"line_number":10,"context_line":"first instance as the marker and ensure that the remaining instances"},{"line_number":11,"context_line":"are returned in the page. Now that we are doing queries to cells in"},{"line_number":12,"context_line":"parallel, consecutive unsorted list queries can return things in"},{"line_number":13,"context_line":"different orders as the cells may reply at different times. The fix"},{"line_number":14,"context_line":"in this patch is to ask for results to be sorted, which is the only way"},{"line_number":15,"context_line":"it makes sense. It worked before purely because we were always scanning"},{"line_number":16,"context_line":"the cells linearly and in the same order."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Related-Bug: #1689692"},{"line_number":19,"context_line":"Change-Id: I3ca2a167c902d565c36a5d5dbba1bf1c214aa20b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":7,"id":"7f515b1d_54e0dcfa","line":16,"range":{"start_line":13,"start_character":60,"end_line":16,"end_character":41},"in_reply_to":"7f515b1d_94ec142a","updated":"2017-09-25 13:40:57.000000000","message":"It wasn\u0027t before, and I didn\u0027t realize we had a default because it was buried in that process_sort_whatever() function, so I thought you got insert order when you didn\u0027t ask for anything. Since I can\u0027t sort on insert order across cells, you\u0027d get unsorted batches which depended on the order in which the cell threads returned.\n\nNow that I\u0027m forcing an order in instance_list, we could probably drop this patch since we\u0027re always sorting all the time. However, it\u0027s doing a thing that depends on a sorted result, so it does\u0027t seem terrible to make it explicit.\n\nYour call.","commit_id":"8faf8b3f1f041b515de1958c426b0a87c41a4061"}],"nova/tests/functional/regressions/test_bug_1689692.py":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"68ad22de02fd1474d33c4b911babc83ddf8d35be","unresolved":false,"context_lines":[{"line_number":71,"context_line":"            self.addCleanup(self.api.delete_server, server[\u0027id\u0027])"},{"line_number":72,"context_line":"            self._wait_for_state_change(self.api, server, \u0027ERROR\u0027)"},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"        servers \u003d self.api.get_servers(search_opts\u003ddict(sort_key\u003d\u0027uuid\u0027))"},{"line_number":75,"context_line":"        self.assertEqual(3, len(servers))"},{"line_number":76,"context_line":""},{"line_number":77,"context_line":"        # Take the first server and user that as our marker."}],"source_content_type":"text/x-python","patch_set":7,"id":"7f515b1d_545d7ce7","line":74,"updated":"2017-09-25 13:37:37.000000000","message":"I\u0027m still a bit confused about why the default sort key created_at is wrong here.\n\nhttps://github.com/openstack/nova/blob/ef0f6341f45257db68147798b052b1addcee4cbb/nova/api/openstack/common.py#L142","commit_id":"8faf8b3f1f041b515de1958c426b0a87c41a4061"}]}