)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":13861,"name":"yatin","email":"ykarel@redhat.com","username":"yatinkarel"},"change_message_id":"09177b72ad1f9ef1a050fbe5a96ea4d70d5bb126","unresolved":true,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"[OVN] Reduce the OVN hash ring touch interval"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"In some loaded environments, the maintenance thread in charge of"},{"line_number":10,"context_line":"refreshing the OVN hash ring node delays the execution of the method."},{"line_number":11,"context_line":"While it is not possible to ensure this periodicity in loaded"},{"line_number":12,"context_line":"environments, this patch is reducing the interval in order to trigger"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"1ea06af5_8bff80c1","line":9,"updated":"2024-12-20 06:18:07.000000000","message":"in non wsgi envs we not see issues due to this even running with 8 workers in same CI environments.","commit_id":"10bd1516844c8f9d57174d9e79772554bb0b1896"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"36eb7eb79dccf8c7d11a9830fca777c4fb53e473","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"[OVN] Reduce the OVN hash ring touch interval"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"In some loaded environments, the maintenance thread in charge of"},{"line_number":10,"context_line":"refreshing the OVN hash ring node delays the execution of the method."},{"line_number":11,"context_line":"While it is not possible to ensure this periodicity in loaded"},{"line_number":12,"context_line":"environments, this patch is reducing the interval in order to trigger"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"943fa6a8_5287926e","line":9,"in_reply_to":"1ea06af5_8bff80c1","updated":"2025-01-16 08:27:18.000000000","message":"Done","commit_id":"10bd1516844c8f9d57174d9e79772554bb0b1896"},{"author":{"_account_id":13861,"name":"yatin","email":"ykarel@redhat.com","username":"yatinkarel"},"change_message_id":"09177b72ad1f9ef1a050fbe5a96ea4d70d5bb126","unresolved":true,"context_lines":[{"line_number":12,"context_line":"environments, this patch is reducing the interval in order to trigger"},{"line_number":13,"context_line":"the method more often."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"The OVN hash ring cache timeout is also increased to match the OVN node"},{"line_number":16,"context_line":"timeout."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Closes-Bug: #2091284"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"e5dc87c3_1bef32b1","line":15,"range":{"start_line":15,"start_character":0,"end_line":15,"end_character":3},"updated":"2024-12-20 06:18:07.000000000","message":"wasn\u0027t it was intentional to have cache timeout half of the node timeout? good to add reason if things changed now then before.\nalso considering we never hit it before on high scale and loaded envs, is this change temporary until eventlet removal?","commit_id":"10bd1516844c8f9d57174d9e79772554bb0b1896"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"99a98c8c08a167c0cd63e2ef7233528d874ddd9a","unresolved":false,"context_lines":[{"line_number":12,"context_line":"environments, this patch is reducing the interval in order to trigger"},{"line_number":13,"context_line":"the method more often."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"The OVN hash ring cache timeout is also increased to match the OVN node"},{"line_number":16,"context_line":"timeout."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Closes-Bug: #2091284"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"ec5dd72f_8cca1bf0","line":15,"range":{"start_line":15,"start_character":0,"end_line":15,"end_character":3},"in_reply_to":"e5dc87c3_1bef32b1","updated":"2025-01-15 14:12:58.000000000","message":"You are right: this is happening after the WSGI migration and happens because the thread in charge of refreshing the hashring node is not executed on time, leading to a rotten hashring register. When the hashring is refreshed (per worker), it sees less active nodes. That could lead to an issue dealing with OVN events: if two nodes have non-matching hashring managers, the events can be not be processed.\n\nI\u0027ll update this commit message and leave the bug open until we finish the eventlet deprecation.","commit_id":"10bd1516844c8f9d57174d9e79772554bb0b1896"}],"/PATCHSET_LEVEL":[{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"8a0c7f749c63de881c6855f05b825d23684108f5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"54d8ed72_8a3cb87d","updated":"2025-01-15 22:55:43.000000000","message":"recheck functional was fixed","commit_id":"18308c3faa82e89c9e52521629b7095182e82d25"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"ce924dfe24189b61b85ca665fa8f04b43f99012d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":8,"id":"779888eb_6786ea4d","updated":"2025-01-17 06:48:59.000000000","message":"recheck neutron-functional","commit_id":"a12aa9f55205251c2440d3abcd5276076795e6db"}],"neutron/common/ovn/constants.py":[{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"4ce505ec26966c8e17d8edb418970c3e9557017b","unresolved":true,"context_lines":[{"line_number":304,"context_line":""},{"line_number":305,"context_line":"# Hash Ring constants"},{"line_number":306,"context_line":"# NOTE(ralonsoh): the hash ring timeouts and interval have been"},{"line_number":307,"context_line":"# modified in [1], during the eventlet deprecation. Once finished,"},{"line_number":308,"context_line":"# this patch should be reverted."},{"line_number":309,"context_line":"# [1]https://review.opendev.org/c/openstack/neutron/+/937351"},{"line_number":310,"context_line":"HASH_RING_NODES_TIMEOUT \u003d 120"}],"source_content_type":"text/x-python","patch_set":6,"id":"387d55c3_bf8a19e4","line":307,"updated":"2025-01-15 18:55:40.000000000","message":"nit: The reference is self-evident since we have git history. I\u0027d remove the link.","commit_id":"18308c3faa82e89c9e52521629b7095182e82d25"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"c6ddbb515fb02c08b86588f6f6b1121e747c3413","unresolved":false,"context_lines":[{"line_number":304,"context_line":""},{"line_number":305,"context_line":"# Hash Ring constants"},{"line_number":306,"context_line":"# NOTE(ralonsoh): the hash ring timeouts and interval have been"},{"line_number":307,"context_line":"# modified in [1], during the eventlet deprecation. Once finished,"},{"line_number":308,"context_line":"# this patch should be reverted."},{"line_number":309,"context_line":"# [1]https://review.opendev.org/c/openstack/neutron/+/937351"},{"line_number":310,"context_line":"HASH_RING_NODES_TIMEOUT \u003d 120"}],"source_content_type":"text/x-python","patch_set":6,"id":"f7670fd7_7a86dd30","line":307,"in_reply_to":"387d55c3_bf8a19e4","updated":"2025-01-16 06:37:07.000000000","message":"Done","commit_id":"18308c3faa82e89c9e52521629b7095182e82d25"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"4ce505ec26966c8e17d8edb418970c3e9557017b","unresolved":true,"context_lines":[{"line_number":308,"context_line":"# this patch should be reverted."},{"line_number":309,"context_line":"# [1]https://review.opendev.org/c/openstack/neutron/+/937351"},{"line_number":310,"context_line":"HASH_RING_NODES_TIMEOUT \u003d 120"},{"line_number":311,"context_line":"HASH_RING_TOUCH_INTERVAL \u003d 15"},{"line_number":312,"context_line":"HASH_RING_CACHE_TIMEOUT \u003d 120"},{"line_number":313,"context_line":"HASH_RING_ML2_GROUP \u003d \u0027mechanism_driver\u0027"},{"line_number":314,"context_line":""}],"source_content_type":"text/x-python","patch_set":6,"id":"4e05cc28_034b3054","line":311,"updated":"2025-01-15 18:55:40.000000000","message":"I think these should be at least considered separately. Reducing touch interval seems like maybe a good idea since it would give the maintenance thread, hopefully, a better chance to touch the node. On the other hand, bumping the cache timeout up, AFAIU, means that nodes will learn later (sometimes too late) about recent hash ring state, which may affect the recovery time when a node goes offline while others are oblivious to this fact.","commit_id":"18308c3faa82e89c9e52521629b7095182e82d25"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"c6ddbb515fb02c08b86588f6f6b1121e747c3413","unresolved":false,"context_lines":[{"line_number":308,"context_line":"# this patch should be reverted."},{"line_number":309,"context_line":"# [1]https://review.opendev.org/c/openstack/neutron/+/937351"},{"line_number":310,"context_line":"HASH_RING_NODES_TIMEOUT \u003d 120"},{"line_number":311,"context_line":"HASH_RING_TOUCH_INTERVAL \u003d 15"},{"line_number":312,"context_line":"HASH_RING_CACHE_TIMEOUT \u003d 120"},{"line_number":313,"context_line":"HASH_RING_ML2_GROUP \u003d \u0027mechanism_driver\u0027"},{"line_number":314,"context_line":""}],"source_content_type":"text/x-python","patch_set":6,"id":"0e2e8b60_71de321b","line":311,"in_reply_to":"4e05cc28_034b3054","updated":"2025-01-16 06:37:07.000000000","message":"Done","commit_id":"18308c3faa82e89c9e52521629b7095182e82d25"}]}