)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"2333cb5644b170f198c72aac60ab36eb2c6719a7","unresolved":false,"context_lines":[{"line_number":9,"context_line":"Provides additional flexibility for specifying how destination hosts are"},{"line_number":10,"context_line":"selected for resize operations."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"blueprint flexible-resize-behavior"},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"Change-Id: Ie231b0bc40b6842cb459be57b55e673650f25edd"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dac4157f_69d26302","line":12,"updated":"2015-04-20 21:20:23.000000000","message":"Seems like this is a duplicate of blueprint configurable-resize-placement.  Can\u0027t you just re-use that since there was ML discussion around the original one back in Grizzly.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c3d018c918016e705cf547be63c0660a302660ef","unresolved":false,"context_lines":[{"line_number":9,"context_line":"Provides additional flexibility for specifying how destination hosts are"},{"line_number":10,"context_line":"selected for resize operations."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"blueprint flexible-resize-behavior"},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"Change-Id: Ie231b0bc40b6842cb459be57b55e673650f25edd"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dac4157f_2c4fd9be","line":12,"in_reply_to":"dac4157f_69d26302","updated":"2015-04-20 21:57:29.000000000","message":"Yeah, that was has another owner and looks like it\u0027s been marked obsolete.  I\u0027d prefer to just use this new one, but if others want to revive the old one, I\u0027m game for that too.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"}],"specs/liberty/approved/flexible-resize-behavior.rst":[{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"dc11ff643b49a5b39fdeb3cc27254970be4d5452","unresolved":false,"context_lines":[{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":38,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":39,"context_line":"  resize operation while my VM is running (e.g., Hyper-V)."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"Project Priority"},{"line_number":42,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_7fbfc3d0","line":39,"updated":"2015-04-29 10:42:15.000000000","message":"Sorry, I am finding this confusing.\n\nWhat happens if there is no space on the hypervisor to perform the resize, do we just fail? We should explain that.\n\nIf you have live-migrate, moving the VM is not going to cause downtime.\n\nIf you want this as part of doing a \"live-resize\" please consider that feature as a whole piece, as looking at this by itself it doesn\u0027t seem to make sense.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"74397901c0ffaaf04cb63da08d698a92292c6491","unresolved":false,"context_lines":[{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":38,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":39,"context_line":"  resize operation while my VM is running (e.g., Hyper-V)."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"Project Priority"},{"line_number":42,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_62a043f5","line":39,"in_reply_to":"5ad0253b_7fbfc3d0","updated":"2015-04-29 12:56:23.000000000","message":"1) Yes, it will fail with NoValidHost if there is no room on the host - I\u0027ve clarified this in the \u0027Proposed change\u0027 section - good comment!\n\n2) Regarding live migrate... that\u0027s not really the focus of this spec.  Rather, it\u0027s to allow an admin to configure the resize behavior to not have an implicit mobility operation invoked; there are operators that I\u0027ve heard been wanting such a capability - which is why I\u0027m proposing this spec - it\u0027s to address some real user feedback I\u0027ve received from real enterprises using OpenStack.  :-)","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"dc11ff643b49a5b39fdeb3cc27254970be4d5452","unresolved":false,"context_lines":[{"line_number":57,"context_line":"This will allow us to deprecate the existing allow_resize_to_same_host option"},{"line_number":58,"context_line":"as it provides a superset of its functionality. In the release(s) in which"},{"line_number":59,"context_line":"these options co-exist (e.g., Liberty), allow_resize_to_same_host will take"},{"line_number":60,"context_line":"precedence over the new option."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Alternatives"},{"line_number":63,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_5f82c701","line":60,"updated":"2015-04-29 10:42:15.000000000","message":"Could we not just have:\n\nforce_resize_to_same_host: True/False\n\nWe just need to say, if force_resize_to_same_host is set, we ignore allow_resize_to_same_host.\n\nThis would stop users having to update their config for the old setting.\n\nIf this were a more general feature thats widely useful, then it might be worth talking about changing an existing setting.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"74397901c0ffaaf04cb63da08d698a92292c6491","unresolved":false,"context_lines":[{"line_number":57,"context_line":"This will allow us to deprecate the existing allow_resize_to_same_host option"},{"line_number":58,"context_line":"as it provides a superset of its functionality. In the release(s) in which"},{"line_number":59,"context_line":"these options co-exist (e.g., Liberty), allow_resize_to_same_host will take"},{"line_number":60,"context_line":"precedence over the new option."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Alternatives"},{"line_number":63,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_e2c1f381","line":60,"in_reply_to":"5ad0253b_5f82c701","updated":"2015-04-29 12:56:23.000000000","message":"We could, but I wanted to avoid having yet another related flag that needs to be reconciled.  The 3-tuple is needed to fully deprecate the allow_resize_to_same_host setting; I personally like providing the ability to clean up the conf settings, but if others feel like we want to leave it around, I can certainly accommodate the True/False setting that you suggest.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"57099be3d77710374f303ded21d0bf974f2b76d7","unresolved":false,"context_lines":[{"line_number":68,"context_line":"  configuration option. This wasn\u0027t selected because this level of granularity"},{"line_number":69,"context_line":"  is likely overkill for what operators are trying to do. That is, this"},{"line_number":70,"context_line":"  behavior is generally considered to be static and thus changing the API"},{"line_number":71,"context_line":"  logic did not seem worthwhile."},{"line_number":72,"context_line":""},{"line_number":73,"context_line":"* We could do nothing and leave the function as is, but there have been"},{"line_number":74,"context_line":"  requests for this functionality for several releases now, so it seems like"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_05cf5980","line":71,"updated":"2015-04-29 13:20:23.000000000","message":"This alternative seems to be a better approach and it is used in other case like migration (cf: we can pass a host). I guess enhance the API to accept a new parameter like \"--same-host:force|forbid|allow\" or a defined host \"--host:ip-my-host\" is better like that we can avoid the problem John speaks about, if the compute node is configured to force resize in same host and there is no space left, the operator looks to be blocked.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"d521b00031996ee12b49a60d4fb1a25be9976d65","unresolved":false,"context_lines":[{"line_number":156,"context_line":""},{"line_number":157,"context_line":"The new nova configuration option will need to be documented as well as the"},{"line_number":158,"context_line":"deprecation of the allow_resize_to_same_host configuration option; we will"},{"line_number":159,"context_line":"target its removal in the M release."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":""},{"line_number":162,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dac4157f_49f67f90","line":159,"updated":"2015-04-20 21:21:59.000000000","message":"If you deprecate in L I think you have to wait for N to remove the old option since you need it in the release notes for L and give it a release cycle to burn in.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"74397901c0ffaaf04cb63da08d698a92292c6491","unresolved":false,"context_lines":[{"line_number":156,"context_line":""},{"line_number":157,"context_line":"The new nova configuration option will need to be documented as well as the"},{"line_number":158,"context_line":"deprecation of the allow_resize_to_same_host configuration option; we will"},{"line_number":159,"context_line":"target its removal in the M release."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":""},{"line_number":162,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_02faafbd","line":159,"in_reply_to":"dac4157f_49f67f90","updated":"2015-04-29 12:56:23.000000000","message":"Done","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c3d018c918016e705cf547be63c0660a302660ef","unresolved":false,"context_lines":[{"line_number":156,"context_line":""},{"line_number":157,"context_line":"The new nova configuration option will need to be documented as well as the"},{"line_number":158,"context_line":"deprecation of the allow_resize_to_same_host configuration option; we will"},{"line_number":159,"context_line":"target its removal in the M release."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":""},{"line_number":162,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dac4157f_4c6b2d1f","line":159,"in_reply_to":"dac4157f_49f67f90","updated":"2015-04-20 21:57:29.000000000","message":"No problem, I can update that in the next rev of this.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"d521b00031996ee12b49a60d4fb1a25be9976d65","unresolved":false,"context_lines":[{"line_number":166,"context_line":"* https://bugs.launchpad.net/nova/+bug/1323578"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Related blueprint(s):"},{"line_number":169,"context_line":"* https://blueprints.launchpad.net/nova/+spec/no-migration-resize"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"Hypervisor support for active resize:"},{"line_number":172,"context_line":"* https://technet.microsoft.com/en-us/library/dn282286.aspx"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dac4157f_c9304f59","line":169,"updated":"2015-04-20 21:21:59.000000000","message":"And https://blueprints.launchpad.net/nova/+spec/configurable-resize-placement","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"74397901c0ffaaf04cb63da08d698a92292c6491","unresolved":false,"context_lines":[{"line_number":166,"context_line":"* https://bugs.launchpad.net/nova/+bug/1323578"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Related blueprint(s):"},{"line_number":169,"context_line":"* https://blueprints.launchpad.net/nova/+spec/no-migration-resize"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"Hypervisor support for active resize:"},{"line_number":172,"context_line":"* https://technet.microsoft.com/en-us/library/dn282286.aspx"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5ad0253b_222b6b32","line":169,"in_reply_to":"dac4157f_c9304f59","updated":"2015-04-29 12:56:23.000000000","message":"Done","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c3d018c918016e705cf547be63c0660a302660ef","unresolved":false,"context_lines":[{"line_number":166,"context_line":"* https://bugs.launchpad.net/nova/+bug/1323578"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Related blueprint(s):"},{"line_number":169,"context_line":"* https://blueprints.launchpad.net/nova/+spec/no-migration-resize"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"Hypervisor support for active resize:"},{"line_number":172,"context_line":"* https://technet.microsoft.com/en-us/library/dn282286.aspx"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dac4157f_cc65fd45","line":169,"in_reply_to":"dac4157f_c9304f59","updated":"2015-04-20 21:57:29.000000000","message":"Good call - I was looking for this one, but couldn\u0027t find it... thanks.","commit_id":"0e03cd52c54d2ff7485161924de8796e62bbdb79"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":25,"context_line":"process with respect to resize operations. Right now it is quite possible that"},{"line_number":26,"context_line":"a resize operation will result in the VM changing host systems. While this"},{"line_number":27,"context_line":"might be acceptable in many situations, there are situations in which the"},{"line_number":28,"context_line":"administrator just wants the VM to stay put. There is an existing nova"},{"line_number":29,"context_line":"configuration option, allow_resize_to_same_host, but that only tells the"},{"line_number":30,"context_line":"scheduler that the VM\u0027s current host can be a candidate, but it by no means"},{"line_number":31,"context_line":"contractually agrees to that. This spec gives the operator that level of"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_874eccbf","line":28,"updated":"2015-05-11 09:45:20.000000000","message":"What is an example of those situations please?","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":25,"context_line":"process with respect to resize operations. Right now it is quite possible that"},{"line_number":26,"context_line":"a resize operation will result in the VM changing host systems. While this"},{"line_number":27,"context_line":"might be acceptable in many situations, there are situations in which the"},{"line_number":28,"context_line":"administrator just wants the VM to stay put. There is an existing nova"},{"line_number":29,"context_line":"configuration option, allow_resize_to_same_host, but that only tells the"},{"line_number":30,"context_line":"scheduler that the VM\u0027s current host can be a candidate, but it by no means"},{"line_number":31,"context_line":"contractually agrees to that. This spec gives the operator that level of"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da1255d7_69dbc5ee","line":28,"in_reply_to":"9af37de9_874eccbf","updated":"2015-05-12 03:49:32.000000000","message":"Added example in the previous description... basically when you want to resize a VM (e.g., change vcpu/memory) while it\u0027s running and not interrupt the workload.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":29,"context_line":"configuration option, allow_resize_to_same_host, but that only tells the"},{"line_number":30,"context_line":"scheduler that the VM\u0027s current host can be a candidate, but it by no means"},{"line_number":31,"context_line":"contractually agrees to that. This spec gives the operator that level of"},{"line_number":32,"context_line":"control."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Use Cases"},{"line_number":35,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_c702943e","line":32,"updated":"2015-05-11 09:45:20.000000000","message":"But what happens when we are unable to resize on the same host. Normally that why the VM is moved.\n\nMaybe what you want, is resize to prefer the same host where possible, and only move to a new host when thats required?","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":29,"context_line":"configuration option, allow_resize_to_same_host, but that only tells the"},{"line_number":30,"context_line":"scheduler that the VM\u0027s current host can be a candidate, but it by no means"},{"line_number":31,"context_line":"contractually agrees to that. This spec gives the operator that level of"},{"line_number":32,"context_line":"control."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Use Cases"},{"line_number":35,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da1255d7_698985ca","line":32,"in_reply_to":"9af37de9_c702943e","updated":"2015-05-12 03:49:32.000000000","message":"See below for the NVH explanation... what you\u0027re suggesting is also a possibility, but not the operators I\u0027ve talked to want.  They don\u0027t want the workload to migrate at all... either resize on the same host while the VM is active, or don\u0027t allow it at all (this is very common in enterprise environments with sensitive workloads).","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Use Cases"},{"line_number":35,"context_line":"----------"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":38,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":39,"context_line":"  resize operation while my VM is running (e.g., Hyper-V)."},{"line_number":40,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_075adc83","line":37,"updated":"2015-05-11 09:45:20.000000000","message":"But resize is a user operation.\nMigrate is an admin operation.\n\nDo you mean the cloud administrator doesn\u0027t want to allow the resize requested by users to move to a new hypervisor?\n\nIf thats the case, what happens where there is not enough space? Do we just fail the resize operation?","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Use Cases"},{"line_number":35,"context_line":"----------"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":38,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":39,"context_line":"  resize operation while my VM is running (e.g., Hyper-V)."},{"line_number":40,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a0c5dc1_304e4ec2","line":37,"in_reply_to":"9af37de9_075adc83","updated":"2015-05-12 03:49:32.000000000","message":"Yes, the operation would fail with a NVH exception... it\u0027s a configuration option (and not the default :-).\n\nWhile this may seem \u0027bad\u0027, that is exactly what the operators I\u0027ve spoke with want... attempt to resize on the same host and if there isn\u0027t enough room, don\u0027t resize at all.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":54,"context_line":"  default behavior to maintain backwards compatability"},{"line_number":55,"context_line":"* allow: allow (albeit not guarantee) the VM to stay on its source host"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"This will allow us to deprecate the existing allow_resize_to_same_host option"},{"line_number":58,"context_line":"as it provides a superset of its functionality. In the release(s) in which"},{"line_number":59,"context_line":"these options co-exist (e.g., Liberty), allow_resize_to_same_host will take"},{"line_number":60,"context_line":"precedence over the new option."}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_a75b1054","line":57,"updated":"2015-05-11 09:45:20.000000000","message":"Why deprecate the old config variable.\n\ncould we not add an additional flag for this, to save all the upgrade pain for existing users?","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":54,"context_line":"  default behavior to maintain backwards compatability"},{"line_number":55,"context_line":"* allow: allow (albeit not guarantee) the VM to stay on its source host"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"This will allow us to deprecate the existing allow_resize_to_same_host option"},{"line_number":58,"context_line":"as it provides a superset of its functionality. In the release(s) in which"},{"line_number":59,"context_line":"these options co-exist (e.g., Liberty), allow_resize_to_same_host will take"},{"line_number":60,"context_line":"precedence over the new option."}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a0c5dc1_30fc0e3e","line":57,"in_reply_to":"9af37de9_a75b1054","updated":"2015-05-12 03:49:32.000000000","message":"Great question... the existing parameter is a BoolOpt and the new proposed value has more of an enum feel... I think after its deprecation cycle gets over, I think people would welcome the new name as it\u0027s a little more explicit.  That\u0027s just my $0.02 though.  :-)","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":63,"context_line":"there doesn\u0027t happen to be enough room on the host to accommodate the request,"},{"line_number":64,"context_line":"the resize will fail with the standard NoValidHost error--which is expected"},{"line_number":65,"context_line":"behavior from the user as they have explicitly told us that they do not want"},{"line_number":66,"context_line":"the VM to have a mobility operation implicitly invoked."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"Alternatives"},{"line_number":69,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_a7ad5033","line":66,"updated":"2015-05-11 09:45:20.000000000","message":"Thats a very poor error to fail with, and its not really the standard error in this case, we should be explicit about that.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":63,"context_line":"there doesn\u0027t happen to be enough room on the host to accommodate the request,"},{"line_number":64,"context_line":"the resize will fail with the standard NoValidHost error--which is expected"},{"line_number":65,"context_line":"behavior from the user as they have explicitly told us that they do not want"},{"line_number":66,"context_line":"the VM to have a mobility operation implicitly invoked."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"Alternatives"},{"line_number":69,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a0c5dc1_50ad4230","line":66,"in_reply_to":"9af37de9_a7ad5033","updated":"2015-05-12 03:49:32.000000000","message":"I think the NVH error follow the existing practices in terms of not finding a valid host - unless perhaps I\u0027m overlooking something?  We could easily change this if you had something else in mind, but I didn\u0027t want to break any scheduler contracts.\n\nAlso, this is not the default resize behavior (i.e., forcing to the same host), so someone would have to willingly configure this.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"034520e33720dea4f7f018ba2723b1105ceda04a","unresolved":false,"context_lines":[{"line_number":74,"context_line":"  configuration option. This wasn\u0027t selected because this level of granularity"},{"line_number":75,"context_line":"  is likely overkill for what operators are trying to do. That is, this"},{"line_number":76,"context_line":"  behavior is generally considered to be static and thus changing the API"},{"line_number":77,"context_line":"  logic did not seem worthwhile."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"* We could do nothing and leave the function as is, but there have been"},{"line_number":80,"context_line":"  requests for this functionality for several releases now, so it seems like"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9af37de9_276840f5","line":77,"updated":"2015-05-11 09:45:20.000000000","message":"Please can you make it much clearer what operators are wanting to do, it doesn\u0027t seem very clear in the use cases.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"2aabcbaf094668d4408a3a0287a4b77f821db0a7","unresolved":false,"context_lines":[{"line_number":74,"context_line":"  configuration option. This wasn\u0027t selected because this level of granularity"},{"line_number":75,"context_line":"  is likely overkill for what operators are trying to do. That is, this"},{"line_number":76,"context_line":"  behavior is generally considered to be static and thus changing the API"},{"line_number":77,"context_line":"  logic did not seem worthwhile."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"* We could do nothing and leave the function as is, but there have been"},{"line_number":80,"context_line":"  requests for this functionality for several releases now, so it seems like"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a0c5dc1_30aace1a","line":77,"in_reply_to":"9af37de9_276840f5","updated":"2015-05-12 03:49:32.000000000","message":"Done... I updated the initial description with the operators\u0027 goals.","commit_id":"daa12bec98be1f73510b3aa5b56134fd316dd433"},{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"7c199d320c7daa45dc022387fc590a26e2684ef3","unresolved":false,"context_lines":[{"line_number":18,"context_line":"or if the user doesn\u0027t doesn\u0027t want the VM to implicitly \u0027move\u0027 just for a"},{"line_number":19,"context_line":"resize operation)."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"That is, some hypervisors (such as PowerVM and Hyper-V) can allow attribytes"},{"line_number":22,"context_line":"of a VM (e.g., vcpus/memory) to be dynamically changed without powering off"},{"line_number":23,"context_line":"or migrating the hypervisor.  In such cases, cloud administrators do not"},{"line_number":24,"context_line":"want an implicit migration to occur when a resize operation occurs so that"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a0c5dc1_9fce1e82","line":21,"updated":"2015-05-12 10:03:35.000000000","message":"PowerVM is not in community support list any more ...\nagree this is a case which should be helpful","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"d4ec8c0ab9c07bd50167e315012658de9497c563","unresolved":false,"context_lines":[{"line_number":18,"context_line":"or if the user doesn\u0027t doesn\u0027t want the VM to implicitly \u0027move\u0027 just for a"},{"line_number":19,"context_line":"resize operation)."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"That is, some hypervisors (such as PowerVM and Hyper-V) can allow attribytes"},{"line_number":22,"context_line":"of a VM (e.g., vcpus/memory) to be dynamically changed without powering off"},{"line_number":23,"context_line":"or migrating the hypervisor.  In such cases, cloud administrators do not"},{"line_number":24,"context_line":"want an implicit migration to occur when a resize operation occurs so that"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7a016987_f84a7dab","line":21,"in_reply_to":"7a016987_12af7882","updated":"2015-05-17 04:47:34.000000000","message":"Sounds good.  And Dan, I was trying to coin a new computer science term... attriBYTES sounds so much cooler.  :-)\nBut yes, thanks for noticing the typo!","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"4b389b4e04ee40e748de491580393bf690aefb5c","unresolved":false,"context_lines":[{"line_number":18,"context_line":"or if the user doesn\u0027t doesn\u0027t want the VM to implicitly \u0027move\u0027 just for a"},{"line_number":19,"context_line":"resize operation)."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"That is, some hypervisors (such as PowerVM and Hyper-V) can allow attribytes"},{"line_number":22,"context_line":"of a VM (e.g., vcpus/memory) to be dynamically changed without powering off"},{"line_number":23,"context_line":"or migrating the hypervisor.  In such cases, cloud administrators do not"},{"line_number":24,"context_line":"want an implicit migration to occur when a resize operation occurs so that"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7a016987_12af7882","line":21,"in_reply_to":"9a0c5dc1_9fce1e82","updated":"2015-05-16 22:10:24.000000000","message":"Yeah, I think PowerVM should be removed as an example here. Most hypervisors allow hot-add of virtual hardware, so no need to call out specifics here, IMHO.\n\nAlso, \"attributes\" is misspelled.","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"7c199d320c7daa45dc022387fc590a26e2684ef3","unresolved":false,"context_lines":[{"line_number":40,"context_line":"Use Cases"},{"line_number":41,"context_line":"----------"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":44,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":45,"context_line":"  resize operation while my VM is running (e.g., Hyper-V or PowerVM) to"},{"line_number":46,"context_line":"  change my memory/vcpus/disk without having to \u0027cold migrate\u0027 it (i.e., \"I"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a0c5dc1_9fca9e4e","line":43,"updated":"2015-05-12 10:03:35.000000000","message":"as admin, any real case that he need to do so? \nwhy admin need to change user\u0027s disk / vcpu/ memory? \njust curious","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"d4ec8c0ab9c07bd50167e315012658de9497c563","unresolved":false,"context_lines":[{"line_number":40,"context_line":"Use Cases"},{"line_number":41,"context_line":"----------"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"* As a cloud administrator, I want to ensure that a resize operation does not"},{"line_number":44,"context_line":"  move my VM, so that I can take advantage of hypervisors that can perform a"},{"line_number":45,"context_line":"  resize operation while my VM is running (e.g., Hyper-V or PowerVM) to"},{"line_number":46,"context_line":"  change my memory/vcpus/disk without having to \u0027cold migrate\u0027 it (i.e., \"I"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7a016987_d8c8e124","line":43,"in_reply_to":"9a0c5dc1_9fca9e4e","updated":"2015-05-17 04:47:34.000000000","message":"e.g., I deploy a VM and it has 2 vcpus initially and the workload is running more CPU intensive tasks than originally planned...admin wants to give it a little more computing power without disrupting the active workloads; same applies to memory.","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"7c199d320c7daa45dc022387fc590a26e2684ef3","unresolved":false,"context_lines":[{"line_number":57,"context_line":"This change will introduce a new nova configuration option, which will be named"},{"line_number":58,"context_line":"resize_to_same_host; its acceptable values are:"},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"* force: force the VM to stay on its source host"},{"line_number":61,"context_line":"* forbid: forbid the VM from staying on its source host; this will be the"},{"line_number":62,"context_line":"  default behavior to maintain backwards compatability"},{"line_number":63,"context_line":"* allow: allow (albeit not guarantee) the VM to stay on its source host"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a0c5dc1_ff1cfa78","line":60,"updated":"2015-05-12 10:03:35.000000000","message":"will this config be put into api layer or compute layer? \nif you put into api layer , this means all resize will be \ndone to the same host\n\nif you put into compute layer, it can\u0027t be detected until scheduler already find the host\n\nor put into scheduler layer?\n\nneed more info here, thanks","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"d4ec8c0ab9c07bd50167e315012658de9497c563","unresolved":false,"context_lines":[{"line_number":57,"context_line":"This change will introduce a new nova configuration option, which will be named"},{"line_number":58,"context_line":"resize_to_same_host; its acceptable values are:"},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"* force: force the VM to stay on its source host"},{"line_number":61,"context_line":"* forbid: forbid the VM from staying on its source host; this will be the"},{"line_number":62,"context_line":"  default behavior to maintain backwards compatability"},{"line_number":63,"context_line":"* allow: allow (albeit not guarantee) the VM to stay on its source host"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7a016987_b8cbe51d","line":60,"in_reply_to":"9a0c5dc1_ff1cfa78","updated":"2015-05-17 04:47:34.000000000","message":"The idea is that this would simply be as a config option for now (see alternatives below...API was considered, but seemed overkill relative to what people were really asking for).\n\nIn terms of the scheduler... the way this was planning to be implemented was to actually just pass force_hosts :\u003d \u003ccurrent host\u003e to the scheduler when making the decision - so the scheduler need not be concerned with the details \u0027above\u0027 it.","commit_id":"8d7720fd3b81c993ba55fa2248e752811817be6b"}]}