)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"90f85b4340a8c774e76deb7f6ef84fd7975247b5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":5,"id":"c297d51a_a71ffb31","updated":"2023-03-07 11:20:46.000000000","message":"im not conviced this is the right approch to take.\n\nfor one it will be leaking infromation about the infra state to normal uses\nthey should  not be aware if the nova compute is up or not.\n\nsecond there is a delay between a compute stoping and the staus going to down.\nso this would not prevent it in all case as we coudl send the stop command, it woudl be added to the rabbit queue and then decueued when the service is started.\n\nas i noted on the bug i think i would prefer to adress this by providing a way to automate purging the message queue e.g. as a nova manage command or as a config option to have nova-compute do that on start.\n\ndoing thsi in the api i think is not something we should do.\n\n","commit_id":"fc2b02b97ed75f0e00f0a00072b01531e5c3c61b"},{"author":{"_account_id":16137,"name":"Tobias Urdin","email":"tobias.urdin@binero.com","username":"tobasco"},"change_message_id":"1f1a28752ace7a0b15282e32c921b10523ee7a70","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":5,"id":"56f324d2_ddd8f075","in_reply_to":"435ba5f5_c4d13ac2","updated":"2024-03-01 09:58:33.000000000","message":"Looking at it again, I have the same issue/bug but for me it wasn\u0027t a unclean reboot, the node never came back at all.","commit_id":"fc2b02b97ed75f0e00f0a00072b01531e5c3c61b"},{"author":{"_account_id":16137,"name":"Tobias Urdin","email":"tobias.urdin@binero.com","username":"tobasco"},"change_message_id":"3245ec9d4e17dc661124dbb0d7aed3bb3235ffe3","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":5,"id":"435ba5f5_c4d13ac2","in_reply_to":"6591b261_f8b63428","updated":"2024-03-01 09:48:47.000000000","message":"Another alternative to drop messages when nova-compute service record is deleted? \n\nIn our case we brought back (reinstalled, all nova services deleted) a compute node that had died due to hardware, back with same hostname, when started updated vm_state\u003dstopped on an instance that was previously evacuated when the hardware failed, the current compute node for the instance during _sync_instance_power_state() saw vm_state\u003dstopped in database and issues a stop on the instance when it shouldn\u0027t have.","commit_id":"fc2b02b97ed75f0e00f0a00072b01531e5c3c61b"},{"author":{"_account_id":32087,"name":"zhouzhong","display_name":"zhouzhong","email":"1923001710@qq.com","username":"Yormng","status":"Easystack"},"change_message_id":"8975f8e7903a97464eba6a02ad6852d400e2e8d4","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":5,"id":"6591b261_f8b63428","in_reply_to":"c297d51a_a71ffb31","updated":"2023-12-05 12:28:30.000000000","message":"Purging the message queue may result in the source nova compute being unable to consume messages, and some write operations at the nova API level may need to be rolled back at the nova compute level. If the message queue is cleared, it will not be able to trigger the nova compute rollback, causing resource residue.","commit_id":"fc2b02b97ed75f0e00f0a00072b01531e5c3c61b"}]}