)]}'
{"specs/queens/approved/add-evacuate-action.rst":[{"author":{"_account_id":2394,"name":"Adam Spiers","email":"aspiers@suse.com","username":"adam.spiers"},"change_message_id":"3d114d656c9a2f13414d4318cdfdf5fc7f49c821","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"},{"line_number":22,"context_line":"* add an action to evacuate all the instances of the failure compute node"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_f1bb84c1","line":19,"range":{"start_line":17,"start_character":0,"end_line":19,"end_character":29},"updated":"2017-08-18 12:08:22.000000000","message":"By \"evacuate\" here I assume you mean \"nova evacuate\" which is not real evacuation:\n\nhttps://youtu.be/lddtWUP_IKQ?t\u003d13m7s\n\nAs explained in I2140167273fa51f234a0e8ac726b629c5b84d448 I don\u0027t think Watcher should be reinventing compute HA.  However there are still other good reasons for supporting compute node evacuation (i.e. real evacuation, where you evacuate a node *before* anything goes wrong) via Watcher, since real evacuation can help a lot with optimization, especially with https://blueprints.launchpad.net/watcher/+spec/strategy-to-trigger-power-on-and-power-off-actions\n\nSo I think that a Watcher spec for adding an evacuate action is justified, but it should be for real evacuation, not for running `nova evacuate` in a failure scenario.","commit_id":"ded7ca32ab7fb1e907349893f116fdd7799d3293"},{"author":{"_account_id":2394,"name":"Adam Spiers","email":"aspiers@suse.com","username":"adam.spiers"},"change_message_id":"3d7ebcaeba5bb6ed0af94d2244b9efcb7314c458","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"},{"line_number":22,"context_line":"* add an action to evacuate all the instances of the failure compute node"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7f287b81_9cbc4e6a","line":19,"range":{"start_line":17,"start_character":0,"end_line":19,"end_character":29},"in_reply_to":"7f287b81_880e52e3","updated":"2017-08-23 13:00:58.000000000","message":"\u003e A choice better than no choice. Evacuating the instances is better than we doing nothing when compute node failed.\n\nYes, and that is why there is an existing OpenStack project called Masakari which does exactly this.  Why do you want to reinvent this in Watcher when it already exists elsewhere and is outside the scope of Watcher as declared in Watcher\u0027s mission statement?","commit_id":"ded7ca32ab7fb1e907349893f116fdd7799d3293"},{"author":{"_account_id":24501,"name":"sue","display_name":"suzhengwei","email":"sugar-2008@163.com","username":"suzhengwei"},"change_message_id":"3f5b7e0115f909dd2a9b0581cb6473c7c136bc65","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"},{"line_number":22,"context_line":"* add an action to evacuate all the instances of the failure compute node"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7f287b81_880e52e3","line":19,"range":{"start_line":17,"start_character":0,"end_line":19,"end_character":29},"in_reply_to":"9f436f4f_51133567","updated":"2017-08-23 07:43:50.000000000","message":"Nova evacuate is progress to schduler and rebuild the VM, with the original metadata, such as flavor, image, network, etc. It is risky.\nNo one can foresee all failure and migrate all instances away before the failure. So we will meet this situation.\nA choice better than no choice. Evacuating the instances is better than we doing nothing when compute node failed.","commit_id":"ded7ca32ab7fb1e907349893f116fdd7799d3293"},{"author":{"_account_id":2394,"name":"Adam Spiers","email":"aspiers@suse.com","username":"adam.spiers"},"change_message_id":"1eba33f69463380cdafbc2f6c36b9ad8602f7e6e","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"},{"line_number":22,"context_line":"* add an action to evacuate all the instances of the failure compute node"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_51133567","line":19,"range":{"start_line":17,"start_character":0,"end_line":19,"end_character":29},"in_reply_to":"9f436f4f_e17a72b8","updated":"2017-08-21 21:19:07.000000000","message":"The Youtube link above explains the difference i.e. why nova evacuate is not \"real\" evacuation, so please watch it if you didn\u0027t already, but I will try to explain here briefly too.  The real meaning of evacuation (in English) is to move things away from some kind of container (e.g. people away from a building) until the container is empty.  But if a compute node fails then it is too late to evacuate (migrate) the VMs because they are already dead.  In that scenario you can only do resurrection of the VMs, not evacuation.","commit_id":"ded7ca32ab7fb1e907349893f116fdd7799d3293"},{"author":{"_account_id":24501,"name":"sue","display_name":"suzhengwei","email":"sugar-2008@163.com","username":"suzhengwei"},"change_message_id":"0cad8361a2d0f9bc23e15fd4647124d597756ccd","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"},{"line_number":22,"context_line":"* add an action to evacuate all the instances of the failure compute node"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_e17a72b8","line":19,"range":{"start_line":17,"start_character":0,"end_line":19,"end_character":29},"in_reply_to":"9f436f4f_f1bb84c1","updated":"2017-08-21 06:10:48.000000000","message":"What do you mean about \"real evacuation\"? What diffrence between \u0027real evacuation\u0027 and `nova evacuate`?","commit_id":"ded7ca32ab7fb1e907349893f116fdd7799d3293"}],"specs/queens/approved/workload-evacuate-action.rst":[{"author":{"_account_id":2394,"name":"Adam Spiers","email":"aspiers@suse.com","username":"adam.spiers"},"change_message_id":"b2b678909509ad374021a246107341a1abf4d05f","unresolved":false,"context_lines":[{"line_number":14,"context_line":"Problem description"},{"line_number":15,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Sometimes if the failure has happened to one compute node, it would be"},{"line_number":18,"context_line":"marked as down. Then we need to evacuate the instances which are running"},{"line_number":19,"context_line":"on this failure compute node."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"It needs to do following things:"}],"source_content_type":"text/x-rst","patch_set":2,"id":"7f287b81_2f4aba54","line":18,"range":{"start_line":17,"start_character":0,"end_line":18,"end_character":15},"updated":"2017-08-23 13:07:03.000000000","message":"Sorry, but for reasons already explained in patch set 1 and elsewhere, I think it is a really bad idea to attempt to reimplement compute node HA inside Watcher, and I don\u0027t understand why this is being proposed. I will attend today\u0027s Watcher IRC meeting in case anyone can explain this to me.\n\nAs I already said, I think it *is* a good idea for Watcher to support real evacuation (i.e. live migration), but not nova-evacuate.","commit_id":"67f38b1572845e998919b72364f220ac6e71b31c"}]}