)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":38562,"name":"Richard Cruise","email":"rcruise@redhat.com","username":"rcruise"},"change_message_id":"f5a220c36cedffa46d572e0d8a065da43e1d9ebe","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"81e9fa18_2a59abf4","updated":"2026-09-28 09:48:37.000000000","message":"LGTM","commit_id":"5b8de83baf404c791c132dc5df1b399582e9721c"},{"author":{"_account_id":29244,"name":"Gregory Thiemonge","email":"gthiemon@redhat.com","username":"gthiemonge"},"change_message_id":"fb650f99b00404a2e824098f2f3eb020727cdaec","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"39b86ba8_7b0301ce","updated":"2026-10-06 08:46:50.000000000","message":"raiseload seems incompatible with our object model, most of our objects contain child objects, and sometimes pulling an object also pulls the parent.\n\nIn the test `octavia.tests.functional.api.v2.test_l7policy.TestL7Policy.test_get_all_limited`\nget_all on l7policies transforms the l7p into our internal structures, it recursively pulls the listeners (l7policy\u0027s parent), and the listeners recursively fetch the allowed_cidrs (which are not loaded in the DB request).\n\nwith noload, the allowed_cidrs were empty even when set (which might be fine because when getting the l7policies, we don\u0027t need the allowed_cidrs), however that triggers an exception with raiseload.\n\nI asked claude about it, it thinks that we should just remove all the noload/raiseload from the code (except line 265 - it considers that raiseload is valid there).\n\nI think I\u0027ll try to measure the impact of it (counting the number of queries for each get_all call)","commit_id":"5b8de83baf404c791c132dc5df1b399582e9721c"},{"author":{"_account_id":38562,"name":"Richard Cruise","email":"rcruise@redhat.com","username":"rcruise"},"change_message_id":"d9cb5f6fcb982e6430c8ca2ce5995de93bca4be4","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"8debbf46_22c19b93","in_reply_to":"39b86ba8_7b0301ce","updated":"2026-10-06 09:26:12.000000000","message":"It seems the L7 policy get_all call uses a different method to load from the DB to the others. I\u0027ve pushed a change to match it with the others and it should resolve the CI failures: https://review.opendev.org/c/openstack/octavia/+/1009009","commit_id":"5b8de83baf404c791c132dc5df1b399582e9721c"},{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"7d4b7e66856df91b76360d63240f12d0b8dd92ee","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"8b89f2ba_d825236f","in_reply_to":"8debbf46_22c19b93","updated":"2026-10-06 14:25:58.000000000","message":"Maybe we could remove noload entirely though I tend to keep these explicit to avoid including \"stale\" information.\n\nIt seems the problem Richard pointed out is the cause and now functional tests are running with the change. Let me re-order the change so that we can merge these step-by-step. (though we still want to evaluate impact of this change)","commit_id":"5b8de83baf404c791c132dc5df1b399582e9721c"},{"author":{"_account_id":29244,"name":"Gregory Thiemonge","email":"gthiemon@redhat.com","username":"gthiemonge"},"change_message_id":"08fafedb9ef72d44222b13cdad1682649b6ee9bf","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"bbf19a04_8bd9d034","updated":"2026-10-07 07:40:56.000000000","message":"LGTM, Thanks","commit_id":"ac01addff830e82fd7b4fe13aeb679ef28a1dce0"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"0cf0de8337ea9d067857631295863bced5e433cf","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"6c11ff0c_8fbfad62","updated":"2026-10-09 13:21:08.000000000","message":"Same change as we\u0027ve made in nova, cinder etc.","commit_id":"64a5187af0ff55803157759b11a28623c72f02e8"},{"author":{"_account_id":38562,"name":"Richard Cruise","email":"rcruise@redhat.com","username":"rcruise"},"change_message_id":"594ff7e1d154247781fbcaefc03da358fee534e2","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"bf9c0398_46e67e5b","updated":"2026-10-07 09:05:32.000000000","message":"Squashed in the other 2 commits","commit_id":"64a5187af0ff55803157759b11a28623c72f02e8"}],"releasenotes/notes/fix-inconsistent-DB-loading-in-L7-policy-endpoint-42e7ddacd595257c.yaml":[{"author":{"_account_id":29244,"name":"Gregory Thiemonge","email":"gthiemon@redhat.com","username":"gthiemonge"},"change_message_id":"af30f82fbf3abdb6ebff8e26ff1822b061f825f2","unresolved":true,"context_lines":[{"line_number":2,"context_line":"fixes:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    Fixed an issue where some of the get_all_API_list calls were using inconsistent"},{"line_number":5,"context_line":"    DB queries. This was likely due to ordering requirements in the response "},{"line_number":6,"context_line":"    which are no longer relevant since pagination works. Updated to use the get_all_orm"},{"line_number":7,"context_line":"    call to be consistent with other endpoints"},{"line_number":8,"context_line":""}],"source_content_type":"text/x-yaml","patch_set":3,"id":"0194f52a_e0fad2a3","line":5,"range":{"start_line":5,"start_character":76,"end_line":5,"end_character":77},"updated":"2026-10-09 12:11:13.000000000","message":"nit extra space","commit_id":"64a5187af0ff55803157759b11a28623c72f02e8"}]}