)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"496722efa101a7a0584e3437103ec49d06303ede","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":10,"id":"6b08628e_98f1be88","updated":"2026-08-21 17:16:06.000000000","message":"Just a few nits inline","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"f7a6c5008dd4301bd4a10f1024159dcd2bcb8d42","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":13,"id":"9db0b5e2_a24b82d4","updated":"2026-08-27 03:03:58.000000000","message":"Update looks good to me, thanks","commit_id":"9d2c6563398e4f1be4243e4c0f19c0f804623ff2"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"c0c106abe3a1db58116e9908f62c6479c3517cbd","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":13,"id":"a65027e7_978190c4","updated":"2026-08-27 14:20:27.000000000","message":"recheck failure on base patch","commit_id":"9d2c6563398e4f1be4243e4c0f19c0f804623ff2"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"6599386f0eddcbb6340a6980b6feeae979b8d8ec","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":13,"id":"c0447b6c_66edcd4c","updated":"2026-08-28 14:48:53.000000000","message":"recheck requeue","commit_id":"9d2c6563398e4f1be4243e4c0f19c0f804623ff2"}],"doc/source/admin/graceful-shutdown.rst":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"496722efa101a7a0584e3437103ec49d06303ede","unresolved":true,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted ungracefully and can leave instances"},{"line_number":19,"context_line":"     in an unwanted state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":10,"id":"dc7c87d1_782cabb4","line":18,"range":{"start_line":18,"start_character":40,"end_line":18,"end_character":52},"updated":"2026-08-21 17:16:06.000000000","message":"Not critical but I would remove \"ungracefully\" here. You already say \"interrupted\" and that\u0027s enough to explain what is going on. The \"ungracefully\" word sounds a bit strange and also makes it seem even worse. \"Interrupted and can leave instances in an unwanted state\" is plenty, IMHO.","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"23183fdb262e7c2506965c6892cc590f4fae68dd","unresolved":false,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted ungracefully and can leave instances"},{"line_number":19,"context_line":"     in an unwanted state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":10,"id":"a71ac9f9_4f3d911d","line":18,"range":{"start_line":18,"start_character":40,"end_line":18,"end_character":52},"in_reply_to":"dc7c87d1_782cabb4","updated":"2026-08-21 17:26:42.000000000","message":"Done","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"496722efa101a7a0584e3437103ec49d06303ede","unresolved":true,"context_lines":[{"line_number":69,"context_line":"Task tracking"},{"line_number":70,"context_line":"-------------"},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Instead of always sleeping for the full"},{"line_number":73,"context_line":":oslo.config:option:`manager_shutdown_timeout`, the service manager tracks"},{"line_number":74,"context_line":"in-progress tasks and waits only until they finish, or until the timeout is"},{"line_number":75,"context_line":"reached, whichever comes first. Tasks are tracked in two ways:"},{"line_number":76,"context_line":""}],"source_content_type":"text/x-rst","patch_set":10,"id":"989a4cea_1d662a80","line":73,"range":{"start_line":72,"start_character":0,"end_line":73,"end_character":47},"updated":"2026-08-21 17:16:06.000000000","message":"I think this makes less sense as a timeless statement of how it works. It makes sense to you because that\u0027s what we\u0027re doing currently and we\u0027re describing a change. But I think just making this a statement of the current behavior is better - start with \"The service manager...\" and keep the rest of the sentence.","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"23183fdb262e7c2506965c6892cc590f4fae68dd","unresolved":false,"context_lines":[{"line_number":69,"context_line":"Task tracking"},{"line_number":70,"context_line":"-------------"},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Instead of always sleeping for the full"},{"line_number":73,"context_line":":oslo.config:option:`manager_shutdown_timeout`, the service manager tracks"},{"line_number":74,"context_line":"in-progress tasks and waits only until they finish, or until the timeout is"},{"line_number":75,"context_line":"reached, whichever comes first. Tasks are tracked in two ways:"},{"line_number":76,"context_line":""}],"source_content_type":"text/x-rst","patch_set":10,"id":"f92e4136_a9317190","line":73,"range":{"start_line":72,"start_character":0,"end_line":73,"end_character":47},"in_reply_to":"989a4cea_1d662a80","updated":"2026-08-21 17:26:42.000000000","message":"Done","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"285eee61148ba3cdee380e39fb18d80e2671eede","unresolved":true,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted and can leave instances in an unwanted"},{"line_number":19,"context_line":"     state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":11,"id":"2c8c76f6_6a206aae","line":18,"updated":"2026-08-25 14:15:41.000000000","message":"Can the admin do something to minimize the chance to such task interruption happen?  Is there config options to tune?","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"42b805fefc64a889a5b18d567b321623e0d80fa0","unresolved":true,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted and can leave instances in an unwanted"},{"line_number":19,"context_line":"     state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":11,"id":"4370db8c_ec0fe363","line":18,"in_reply_to":"2c8c76f6_6a206aae","updated":"2026-08-25 16:34:55.000000000","message":"Also note that up to 20 seconds of the `manager_shutdown_timeout` is reserved for host cleanup and admin might want to take that into account when they configure their timeout.","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"dc6cc0b0686654bc5eed632b9052f94267f7a928","unresolved":true,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted and can leave instances in an unwanted"},{"line_number":19,"context_line":"     state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":11,"id":"81615b0a_17dcc37d","line":18,"in_reply_to":"4370db8c_ec0fe363","updated":"2026-08-25 20:36:55.000000000","message":"yes, \u0027manager_shutdown_timeout\u0027 is config option which control the total wait time (-20 as melwitt mentioned that is reserved for the cleanup_host) for in-progress tasks.\n\nI explained the configuration needed to control these timeout in later section\u0027Configuration\u0027 L96","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"a3acb48fa214bd561f5810c73b71560a6f7f6213","unresolved":false,"context_lines":[{"line_number":15,"context_line":"     The service manager tracks in-progress tasks and waits only until they"},{"line_number":16,"context_line":"     complete or until :oslo.config:option:`manager_shutdown_timeout` is"},{"line_number":17,"context_line":"     reached, whichever comes first. Operations that do not complete within"},{"line_number":18,"context_line":"     this timeout are still interrupted and can leave instances in an unwanted"},{"line_number":19,"context_line":"     state."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"How graceful shutdown works for nova-compute service"}],"source_content_type":"text/x-rst","patch_set":11,"id":"cd608fee_e6e88438","line":18,"in_reply_to":"81615b0a_17dcc37d","updated":"2026-08-27 08:24:21.000000000","message":"Acknowledged","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"42b805fefc64a889a5b18d567b321623e0d80fa0","unresolved":true,"context_lines":[{"line_number":30,"context_line":"   up to :oslo.config:option:`manager_shutdown_timeout` seconds."},{"line_number":31,"context_line":"#. The secondary RPC server (``compute-alt`` topic) is stopped."},{"line_number":32,"context_line":"#. The service is stopped."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"For ``nova-conductor`` and ``nova-scheduler``, the sequence is the same"},{"line_number":35,"context_line":"except there is only one RPC server and the further requests are handled"},{"line_number":36,"context_line":"by their other workers."}],"source_content_type":"text/x-rst","patch_set":11,"id":"dfb36209_9d1622df","line":33,"updated":"2026-08-25 16:34:55.000000000","message":"I wonder if any other update should be made to this list to reflect the change in ordering [1] from 1) shutdown RPC server 2) graceful shutdown and wait to 1) stop RPC server 2) graceful shutdown and wait 3) RPC server wait?\n\n[1] https://review.opendev.org/c/openstack/nova/+/996583","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"dc6cc0b0686654bc5eed632b9052f94267f7a928","unresolved":true,"context_lines":[{"line_number":30,"context_line":"   up to :oslo.config:option:`manager_shutdown_timeout` seconds."},{"line_number":31,"context_line":"#. The secondary RPC server (``compute-alt`` topic) is stopped."},{"line_number":32,"context_line":"#. The service is stopped."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"For ``nova-conductor`` and ``nova-scheduler``, the sequence is the same"},{"line_number":35,"context_line":"except there is only one RPC server and the further requests are handled"},{"line_number":36,"context_line":"by their other workers."}],"source_content_type":"text/x-rst","patch_set":11,"id":"c7ccecfd_9bf469d5","line":33,"in_reply_to":"dfb36209_9d1622df","updated":"2026-08-25 20:36:55.000000000","message":"stop and wait are very implementation details for admin, i would like to prefer the overall flow to them without worry about difference between stop and wait.","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"42b805fefc64a889a5b18d567b321623e0d80fa0","unresolved":true,"context_lines":[{"line_number":52,"context_line":"Operations handled during shutdown"},{"line_number":53,"context_line":"----------------------------------"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"The following operations use the secondary RPC server so that they will be"},{"line_number":56,"context_line":"allowed to complete during a graceful shutdown:"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"* Live migration"}],"source_content_type":"text/x-rst","patch_set":11,"id":"11ee51c7_f1601263","line":55,"range":{"start_line":55,"start_character":14,"end_line":55,"end_character":24},"updated":"2026-08-25 16:34:55.000000000","message":"It might be worth adding a few more words such as \"operations which require a sequence of multiple RPC calls\" to explain why they need the secondary server.\n\nOtherwise it could sound like the secondary server is a necessary component for completing tasks during graceful shutdown, when it\u0027s not necessary for things like snapshot to still complete during shutdown.\n\nI know the multiple RPC calls are mentioned in the section below but thought it still might help to mention it here as this list is singled out as a list of operations that will be able to complete during shutdown.","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"dc6cc0b0686654bc5eed632b9052f94267f7a928","unresolved":false,"context_lines":[{"line_number":52,"context_line":"Operations handled during shutdown"},{"line_number":53,"context_line":"----------------------------------"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"The following operations use the secondary RPC server so that they will be"},{"line_number":56,"context_line":"allowed to complete during a graceful shutdown:"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"* Live migration"}],"source_content_type":"text/x-rst","patch_set":11,"id":"2e9ee465_504b7ccd","line":55,"range":{"start_line":55,"start_character":14,"end_line":55,"end_character":24},"in_reply_to":"11ee51c7_f1601263","updated":"2026-08-25 20:36:55.000000000","message":"yeah, good idea. done","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"42b805fefc64a889a5b18d567b321623e0d80fa0","unresolved":true,"context_lines":[{"line_number":83,"context_line":"  migration/resize, are tracked separately for the duration of the"},{"line_number":84,"context_line":"  background operation. Live migration and cold migration/resize involve"},{"line_number":85,"context_line":"  multiple RPC calls across the source and destination compute hosts, so"},{"line_number":86,"context_line":"  these operations remain tracked from the first call until the final call"},{"line_number":87,"context_line":"  that completes or reverts them."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"Tasks are always tracked, but are only logged once a graceful shutdown has"},{"line_number":90,"context_line":"been initiated. When shutdown begins, the manager logs the in-progress"}],"source_content_type":"text/x-rst","patch_set":11,"id":"2d8d9b63_00f91baf","line":87,"range":{"start_line":86,"start_character":54,"end_line":87,"end_character":32},"updated":"2026-08-25 16:34:55.000000000","message":"This isn\u0027t accurate, right? Cold migration/resize are tracked until finish_resize (VM state VERIFY_RESIZE) not confirm or revert, IIRC.","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"dc6cc0b0686654bc5eed632b9052f94267f7a928","unresolved":false,"context_lines":[{"line_number":83,"context_line":"  migration/resize, are tracked separately for the duration of the"},{"line_number":84,"context_line":"  background operation. Live migration and cold migration/resize involve"},{"line_number":85,"context_line":"  multiple RPC calls across the source and destination compute hosts, so"},{"line_number":86,"context_line":"  these operations remain tracked from the first call until the final call"},{"line_number":87,"context_line":"  that completes or reverts them."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"Tasks are always tracked, but are only logged once a graceful shutdown has"},{"line_number":90,"context_line":"been initiated. When shutdown begins, the manager logs the in-progress"}],"source_content_type":"text/x-rst","patch_set":11,"id":"f13ad308_5398d619","line":87,"range":{"start_line":86,"start_character":54,"end_line":87,"end_character":32},"in_reply_to":"2d8d9b63_00f91baf","updated":"2026-08-25 20:36:55.000000000","message":"true, I am considering the revert/confirm resize as a separate operation but i can understand the confusion here, let me explicitly call revert/confirm resize here.","commit_id":"56cc728cf8bf81c6385f3d01dc309e907fab9d66"}],"releasenotes/notes/nova-graceful-shutdown-task-tracking-13485c8346512ae4.yaml":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"496722efa101a7a0584e3437103ec49d06303ede","unresolved":true,"context_lines":[{"line_number":4,"context_line":"    Graceful shutdown now uses a task tracking mechanism instead of a fixed"},{"line_number":5,"context_line":"    time based sleep to wait for in-progress tasks to finish. When"},{"line_number":6,"context_line":"    ``nova-compute``, ``nova-conductor`` or ``nova-scheduler`` receives"},{"line_number":7,"context_line":"    ``SIGTERM``, the service manager tracks all in-progress tasks and waits"},{"line_number":8,"context_line":"    only until they complete (or until"},{"line_number":9,"context_line":"    ``[DEFAULT]/manager_shutdown_timeout`` is reached), instead of"},{"line_number":10,"context_line":"    always sleeping for the full timeout."}],"source_content_type":"text/x-yaml","patch_set":10,"id":"aa3ffca7_505c31d3","line":7,"range":{"start_line":7,"start_character":37,"end_line":7,"end_character":43},"updated":"2026-08-21 17:16:06.000000000","message":"The tracking is already done.. I think \"surveys\" or \"logs\" or \"checks\" here would make more sense.","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"},{"author":{"_account_id":8556,"name":"Ghanshyam Maan","display_name":"Ghanshyam Maan","email":"gmaan.os14@gmail.com","username":"ghanshyam"},"change_message_id":"23183fdb262e7c2506965c6892cc590f4fae68dd","unresolved":false,"context_lines":[{"line_number":4,"context_line":"    Graceful shutdown now uses a task tracking mechanism instead of a fixed"},{"line_number":5,"context_line":"    time based sleep to wait for in-progress tasks to finish. When"},{"line_number":6,"context_line":"    ``nova-compute``, ``nova-conductor`` or ``nova-scheduler`` receives"},{"line_number":7,"context_line":"    ``SIGTERM``, the service manager tracks all in-progress tasks and waits"},{"line_number":8,"context_line":"    only until they complete (or until"},{"line_number":9,"context_line":"    ``[DEFAULT]/manager_shutdown_timeout`` is reached), instead of"},{"line_number":10,"context_line":"    always sleeping for the full timeout."}],"source_content_type":"text/x-yaml","patch_set":10,"id":"689ac136_bef01b0e","line":7,"range":{"start_line":7,"start_character":37,"end_line":7,"end_character":43},"in_reply_to":"aa3ffca7_505c31d3","updated":"2026-08-21 17:26:42.000000000","message":"Done","commit_id":"459b3ac0d77e822d01ddea6105de0826102ff1f4"}]}