)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":6681,"name":"Brent Eagles","email":"beagles@redhat.com","username":"beagles"},"change_message_id":"31a55d98f29268cb1b8bfe786e4998b88d772b14","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"343b34d7_51269a43","updated":"2023-04-06 12:16:02.000000000","message":"A few questions:\n- What happens during upgrade? Is a similar refresh required?\n- I\u0027m guessing we don\u0027t lose functionality because the OVN DBs will be probably not be restarted all at once and the RAFT\u0027ness should make it pretty clean. Like to get confirmation on the expectations from OVN folks.","commit_id":"3780ed21bb7f4a77cf2d124a43e4edc7245cef07"},{"author":{"_account_id":6681,"name":"Brent Eagles","email":"beagles@redhat.com","username":"beagles"},"change_message_id":"adf69ac4080a9cf32882fc84e20e43194ca3a586","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"c20ff786_9808ebc2","updated":"2023-04-11 12:42:33.000000000","message":"thanks for the replies to my questions","commit_id":"3780ed21bb7f4a77cf2d124a43e4edc7245cef07"},{"author":{"_account_id":33080,"name":"Mikolaj Ciecierski","display_name":"Mikolaj Ciecierski","email":"mciecier@redhat.com","username":"mciecierski"},"change_message_id":"f67336622985063a87cd7ef3502e67b0c9f08988","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"c576c6dc_2f686ea2","in_reply_to":"343b34d7_51269a43","updated":"2023-04-06 12:22:51.000000000","message":"1. During upgrade(ffu) this steps is triggered:\nhttps://review.opendev.org/c/openstack/tripleo-heat-templates/+/873602/5/deployment/ovn/ovn-dbs-cluster-ansible.yaml#b357 \nCode for upgrade is in wallaby, because this is a target ffu being worked on currently.However, we can\u0027t compare upgrade with update, because this is path from train to wallaby(migration of dbs from pacemaker to raft).\n\n2. Update of dbs happens in serial, as controller nodes are updated in serial.","commit_id":"3780ed21bb7f4a77cf2d124a43e4edc7245cef07"},{"author":{"_account_id":8297,"name":"Sofer Athlan-Guyot","email":"sathlang@redhat.com","username":"chem"},"change_message_id":"884d21e41e8a860e7f2c509bef691da70763e03f","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"a3f5b917_7a03197d","in_reply_to":"c576c6dc_2f686ea2","updated":"2023-04-07 09:51:11.000000000","message":"To add to the upgrade reply, the control plane is brought down and re-created,\nso beside removing ovndb from pacemaker, no special care is needed to get\nthe new ovndb.\n\nFor the functionality, we have added a control plane test during the\nupdate: we create a vm/fip/storage and destroy it in a loop during the\nentire update. The empiric result was that it doesn\u0027t fail more than\nwhen we don\u0027t update the ovn databases[1].\n\nThis is only an empirical test, and the documentation doesn\u0027t mention\nanything special to be done when updating the cluster. So confirmation\nby OVN people would be nice. But as it stands it\u0027s \"working\".\n\n[1] We have some cuts, but short and they recover quickly, and we have them with or without ovn db update.","commit_id":"3780ed21bb7f4a77cf2d124a43e4edc7245cef07"}]}