)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"979d6c45a64e426e11810c0cc74881fcfe3613bd","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"65b7d9af_cf15738a","updated":"2024-10-03 08:08:18.000000000","message":"check experimental","commit_id":"a356ebc0116cd1a6f102a25fc6ccd838fea68bf5"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"1508c6aba0c3a4613b3d0b228b767cd9264dc57a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"130982fd_6c55d5c0","updated":"2024-10-07 14:27:14.000000000","message":"I tried to understand why it was needed in the first place. I think the original bug report was suggesting that there\u0027s a RPC call somewhere else that couldn\u0027t progress, but I don\u0027t see RPC triggered in this code now or in the past. I think back in the days (~2013) there was no good understanding of how mysql deadlocks are triggered in eventlet environment, something that later was explained by an old mysql python driver used (since then we switched to pymysql). E.g. I recollect this problem being discussed in the context of: https://openstack-dev.openstack.narkive.com/aLhmuBgW/neutron-all-switch-from-mysqldb-to-another-eventlet-aware-mysql-client\n\nIt\u0027s been a while and I forgot details, but tl;dr I think these sleeps are no longer needed assuming we don\u0027t trigger RPC calls that may timeout in the middle of db transaction handling code. I don\u0027t think this is happening here, so it should be safe to merge this. (Gate testing will have to prove it; we are in the good place for this, just starting the new release cycle.)","commit_id":"79ad54f6b7695d8e33d984f7dbf947814d0b2aad"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"a016ab4bcf3cf0dc5e931ac7384d3ea406a14983","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"b46999f2_8acbc01f","in_reply_to":"130982fd_6c55d5c0","updated":"2024-10-11 08:16:10.000000000","message":"Right now we create the RPC updates at the end of the API call. Actually we have recently changed the OVO RPC engine to be synchronous with the API call thread. As you commented, we don\u0027t make any RPC call in the middle of a DB txn.\n\n[1]https://review.opendev.org/c/openstack/neutron/+/926922","commit_id":"79ad54f6b7695d8e33d984f7dbf947814d0b2aad"}]}