)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":1207,"name":"Duncan Thomas","email":"duncan.thomas@gmail.com","username":"duncan-thomas"},"change_message_id":"0002add78d7979364b3554e33596c68746048f51","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Add new attach/detach APIs"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Read the spec or the bp for more details."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Change-Id: Ibddabbc9b2e61b2c1cab878881c4d37e3a13a6e7"},{"line_number":12,"context_line":"Implements: blueprint add-new-attach-apis"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"fa7ab95a_134d376c","line":9,"range":{"start_line":9,"start_character":0,"end_line":9,"end_character":4},"updated":"2016-08-28 14:39:16.000000000","message":"If you could just paste in the following, that would be great :-) :\n\nThe current attach API carries a lot of cruft, and has more complexity than actually needed. This spec proposes a new API that is clearer in intent, race free, and better specified.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"}],"specs/newton/add-new-attach-apis.rst":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"21d478f9431d411caa303931b45baaf5faaacb8e","unresolved":false,"context_lines":[{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":8,"context_line":"Introduce new volume attach APIs"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":10,"context_line":"https://blueprints.launchpad.net/cinder/+spec/example"},{"line_number":11,"context_line":""},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Problem description"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_6d0293b5","line":10,"updated":"2016-10-10 17:34:19.000000000","message":"https://blueprints.launchpad.net/cinder/+spec/add-new-attach-apis","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"ecf65107f7c3ad81c765ad94f5b009bac7a7be6d","unresolved":false,"context_lines":[{"line_number":48,"context_line":""},{"line_number":49,"context_line":"  Condense the operations we currently have in initialize_connection and"},{"line_number":50,"context_line":"  attach into a single call which creates an attachment record in the DB"},{"line_number":51,"context_line":"  for the volume."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"On the detach side we do a similar combination:"},{"line_number":54,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_2af85f7c","line":51,"updated":"2016-10-04 15:22:24.000000000","message":"We\u0027ve discussed that Nova could have a call to ACK that they\u0027ve called Brick to attach the volume. Will we have a call to Cinder after that, or included in that, to set the DB state from attaching-\u003e attached and attaching-\u003e in-use?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"21d478f9431d411caa303931b45baaf5faaacb8e","unresolved":false,"context_lines":[{"line_number":48,"context_line":""},{"line_number":49,"context_line":"  Condense the operations we currently have in initialize_connection and"},{"line_number":50,"context_line":"  attach into a single call which creates an attachment record in the DB"},{"line_number":51,"context_line":"  for the volume."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"On the detach side we do a similar combination:"},{"line_number":54,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_2d424bae","line":51,"in_reply_to":"1a95cdbc_2af85f7c","updated":"2016-10-10 17:34:19.000000000","message":"So today the flow is:\n\n1. nova calls initialize_connection to get the connection info\n2. nova calls brick to connect the volume (using the connection_info from #1).\n3. nova calls os-attach (cinder) to change the volume status to \u0027in-use\u0027.\n\nWouldn\u0027t introducing a new ACK call to cinder that nova has called brick...not new? Since that\u0027s basically what we do with os-attach today.\n\nAlso, not all virt drivers in nova use os-brick, but I think what you mean is just the nova virt driver connects the volume (however it does that).","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":49,"context_line":"  Condense the operations we currently have in initialize_connection and"},{"line_number":50,"context_line":"  attach into a single call which creates an attachment record in the DB"},{"line_number":51,"context_line":"  for the volume."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"On the detach side we do a similar combination:"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"* unreserve"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_a64847b4","line":52,"updated":"2016-08-30 16:40:54.000000000","message":"Where do we set the Cinder DB to something other than \"attaching\"? i.e., do we still call attach_volume()?\nI think it\u0027s a great step to just do a wrapper and preserve the existing semantics, but I favor the idea of just giving the caller an attachment and letting them do with it as they please. Cinder should not care. In that case, perhaps we\u0027d want a new state \"exported\" to reflect what happens after we\u0027ve create_attachment\u0027ed.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"On the detach side we do a similar combination:"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"* unreserve"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"  Same unreserve that we have today, and simply changes the DB"},{"line_number":58,"context_line":"  note that this isn\u0027t used in most detach scenarios currently, instead only"},{"line_number":59,"context_line":"  check_detach is used"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"* remove_attachment"},{"line_number":62,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_64698c8b","line":59,"range":{"start_line":55,"start_character":0,"end_line":59,"end_character":22},"updated":"2016-09-20 10:51:08.000000000","message":"So why do we have this, is this to undo the reserve in a rollback situation? Or is this to ensure you are the only person thinking about calling remove_attachement?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"21d478f9431d411caa303931b45baaf5faaacb8e","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"On the detach side we do a similar combination:"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"* unreserve"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"  Same unreserve that we have today, and simply changes the DB"},{"line_number":58,"context_line":"  note that this isn\u0027t used in most detach scenarios currently, instead only"},{"line_number":59,"context_line":"  check_detach is used"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"* remove_attachment"},{"line_number":62,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_8d48bfb4","line":59,"range":{"start_line":55,"start_character":0,"end_line":59,"end_character":22},"in_reply_to":"7a8ec9b2_64698c8b","updated":"2016-10-10 17:34:19.000000000","message":"Nova only uses unreserve today on a rollback from a failed attach.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":63,"context_line":"  This would replace terminate_connection and detach calls.  Rather than try"},{"line_number":64,"context_line":"  and guess what the consumer of the volume means when they send terminate"},{"line_number":65,"context_line":"  or detach we now make an explicit call and leave it up to the caller to"},{"line_number":66,"context_line":"  decide what should happen."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"  In the case of things like live-migration a new attachment record"},{"line_number":69,"context_line":"  should be created and used for the transition and then the old one deleted"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_4493b0fa","line":66,"updated":"2016-09-20 10:51:08.000000000","message":"Can we include the arguments here? I think we just need to specify the volume uuid and the host? Or do we just pass the attachment id?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"21d478f9431d411caa303931b45baaf5faaacb8e","unresolved":false,"context_lines":[{"line_number":63,"context_line":"  This would replace terminate_connection and detach calls.  Rather than try"},{"line_number":64,"context_line":"  and guess what the consumer of the volume means when they send terminate"},{"line_number":65,"context_line":"  or detach we now make an explicit call and leave it up to the caller to"},{"line_number":66,"context_line":"  decide what should happen."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"  In the case of things like live-migration a new attachment record"},{"line_number":69,"context_line":"  should be created and used for the transition and then the old one deleted"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_681c416a","line":66,"in_reply_to":"7a8ec9b2_4493b0fa","updated":"2016-10-10 17:34:19.000000000","message":"terminate_connection also takes a connector blob today.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":69,"context_line":"  should be created and used for the transition and then the old one deleted"},{"line_number":70,"context_line":"  after the migration is completed.  We should also provide an"},{"line_number":71,"context_line":"  update_attachment method, to use to indicate things like shelved instances,"},{"line_number":72,"context_line":"  live migration etc."},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"The initial introduction of this would be rather simple, and we would just wrap"},{"line_number":75,"context_line":"the existing initialize/attach and terminate/detach into helper API methods."}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_c411805f","line":72,"updated":"2016-09-20 10:51:08.000000000","message":"How do we decide when we need to disconnect from a shared volume attachment on a specific host? I think we were going to get more information back about that, or should we have already done that before calling Cinder? I am confused I guess.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":76,"context_line":"This way we can coordinate all of the connector and attachment record info"},{"line_number":77,"context_line":"rather easily.  This will solve a number of problems without having to go and"},{"line_number":78,"context_line":"modify over 80 Cinder drivers and allow us to over time start improving and"},{"line_number":79,"context_line":"rewriting how we actually deal with connections on the Cinder side."},{"line_number":80,"context_line":""},{"line_number":81,"context_line":"Most importantly it will give us some flexibility to move forward on things"},{"line_number":82,"context_line":"like multi-attach."}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_66ed098e","line":79,"updated":"2016-09-20 10:51:08.000000000","message":"Seems like a good way to start his, no impact on the drivers to start with.\n\nI am guessing we need to also store the connection info to implement the new API?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"ecf65107f7c3ad81c765ad94f5b009bac7a7be6d","unresolved":false,"context_lines":[{"line_number":79,"context_line":"rewriting how we actually deal with connections on the Cinder side."},{"line_number":80,"context_line":""},{"line_number":81,"context_line":"Most importantly it will give us some flexibility to move forward on things"},{"line_number":82,"context_line":"like multi-attach."},{"line_number":83,"context_line":""},{"line_number":84,"context_line":"Alternatives"},{"line_number":85,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_4aff6b34","line":82,"updated":"2016-10-04 15:22:24.000000000","message":"Can we discuss multi-attach in this spec? It seems that one of the reasons for a new API is multi-attach, and it should be designed in from the start.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":95,"context_line":""},{"line_number":96,"context_line":"Data model impact"},{"line_number":97,"context_line":"-----------------"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"Changes which require modifications to the data model often have a wider impact"},{"line_number":100,"context_line":"on the system.  The community often has strong opinions on how the data model"},{"line_number":101,"context_line":"should be evolved, from both a functional and performance perspective. It is"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_8d80c68c","line":98,"updated":"2016-08-30 16:40:54.000000000","message":"I think this is None, right?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":95,"context_line":""},{"line_number":96,"context_line":"Data model impact"},{"line_number":97,"context_line":"-----------------"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"Changes which require modifications to the data model often have a wider impact"},{"line_number":100,"context_line":"on the system.  The community often has strong opinions on how the data model"},{"line_number":101,"context_line":"should be evolved, from both a functional and performance perspective. It is"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_c5a41a0a","line":98,"in_reply_to":"fa7ab95a_815c0d4c","updated":"2016-09-20 10:51:08.000000000","message":"I see this as cinder state that is semi-delegated to Nova, in a way that causes races with internal cinder operations. But I would say that, I work on Nova.\n\nI think what we were talking about was storing the connection_info along side the attachment record. So cinder knows exactly what is happening to each volume.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"4c7ddd57be62742b1e828d700be0887b58f5c5ac","unresolved":false,"context_lines":[{"line_number":95,"context_line":""},{"line_number":96,"context_line":"Data model impact"},{"line_number":97,"context_line":"-----------------"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"Changes which require modifications to the data model often have a wider impact"},{"line_number":100,"context_line":"on the system.  The community often has strong opinions on how the data model"},{"line_number":101,"context_line":"should be evolved, from both a functional and performance perspective. It is"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_815c0d4c","line":98,"in_reply_to":"fa7ab95a_8d80c68c","updated":"2016-08-30 16:54:23.000000000","message":"Or, ideally, we\u0027d at a new state \"exported\" to help get us out of the business of managing Nova attach state...","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":115,"context_line":""},{"line_number":116,"context_line":"REST API impact"},{"line_number":117,"context_line":"---------------"},{"line_number":118,"context_line":""},{"line_number":119,"context_line":"Each API method which is either added or changed should have the following"},{"line_number":120,"context_line":""},{"line_number":121,"context_line":"* Specification for the method"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_ad290a9d","line":118,"updated":"2016-08-30 16:40:54.000000000","message":"TODO: put stuff here","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":197,"context_line":"we will need an attachment ID that the Consumer should store and reference"},{"line_number":198,"context_line":"going forward when they are done with the attachment.  This means it\u0027s up to"},{"line_number":199,"context_line":"the consumer to decide when they are truly done and they want to destroy the"},{"line_number":200,"context_line":"attachment as opposed to now where we try and guess for them based on clues."},{"line_number":201,"context_line":""},{"line_number":202,"context_line":""},{"line_number":203,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_3022b1a3","line":200,"updated":"2016-08-30 16:40:54.000000000","message":"+1000\nBut I think it\u0027s important that a Cinder admin has a way to cleanup after badly behaving clients. Which likely means storing the attachment_id and having a way for the admin to find this info for a given volume...","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"21d478f9431d411caa303931b45baaf5faaacb8e","unresolved":false,"context_lines":[{"line_number":197,"context_line":"we will need an attachment ID that the Consumer should store and reference"},{"line_number":198,"context_line":"going forward when they are done with the attachment.  This means it\u0027s up to"},{"line_number":199,"context_line":"the consumer to decide when they are truly done and they want to destroy the"},{"line_number":200,"context_line":"attachment as opposed to now where we try and guess for them based on clues."},{"line_number":201,"context_line":""},{"line_number":202,"context_line":""},{"line_number":203,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a95cdbc_a89d69bd","line":200,"in_reply_to":"7a8ec9b2_c5c97aae","updated":"2016-10-10 17:34:19.000000000","message":"Doesn\u0027t nova just remove the attachment when we (1) swap volume, (2) detach a volume, (3) delete the instance, etc?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":197,"context_line":"we will need an attachment ID that the Consumer should store and reference"},{"line_number":198,"context_line":"going forward when they are done with the attachment.  This means it\u0027s up to"},{"line_number":199,"context_line":"the consumer to decide when they are truly done and they want to destroy the"},{"line_number":200,"context_line":"attachment as opposed to now where we try and guess for them based on clues."},{"line_number":201,"context_line":""},{"line_number":202,"context_line":""},{"line_number":203,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_c5c97aae","line":200,"in_reply_to":"fa7ab95a_3022b1a3","updated":"2016-09-20 10:51:08.000000000","message":"Ah, so thats interesting.\n\nHow does nova know when to remove the attachment?\n\nIs the idea of unreserve an attachment, to stop someone else starting to use the attachment while we are in the process of removing the attachment?","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":206,"context_line":""},{"line_number":207,"context_line":"Primary assignee:"},{"line_number":208,"context_line":"  john-griffith"},{"line_number":209,"context_line":""},{"line_number":210,"context_line":"Work Items"},{"line_number":211,"context_line":"----------"},{"line_number":212,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_b04fa132","line":209,"updated":"2016-08-30 16:40:54.000000000","message":"API-microversion stuff:\n  Scott DAngelo (scottda)","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":212,"context_line":""},{"line_number":213,"context_line":"1. Implement API calls and wrappers in Cinder"},{"line_number":214,"context_line":""},{"line_number":215,"context_line":"2. Implement cinderclient calls to expose them"},{"line_number":216,"context_line":""},{"line_number":217,"context_line":"3. Implement changes in Nova"},{"line_number":218,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_106dadcb","line":215,"updated":"2016-08-30 16:40:54.000000000","message":"and python-brick-cinderclient-ext","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":1207,"name":"Duncan Thomas","email":"duncan.thomas@gmail.com","username":"duncan-thomas"},"change_message_id":"bdd900eac7968200ee7b4e9b4f459f16b1bbb487","unresolved":false,"context_lines":[{"line_number":223,"context_line":"Testing"},{"line_number":224,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":225,"context_line":""},{"line_number":226,"context_line":"For now we won\u0027t add new Tempest tests, the existing tests should test"},{"line_number":227,"context_line":"the initial changes sufficiently assuming we test a version of Nova"},{"line_number":228,"context_line":"incorporates them and uses the corrrect versioning."},{"line_number":229,"context_line":""},{"line_number":230,"context_line":"Documentation Impact"},{"line_number":231,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_930a2720","line":228,"range":{"start_line":226,"start_character":0,"end_line":228,"end_character":51},"updated":"2016-08-28 14:31:20.000000000","message":"Not sure I agree here. Certainly get the things written, but the existing tempest tests are part of the problem - they make assumptions about the API that aren\u0027t necessarily true. I\u0027d see new, good quality tempest tests as part of the (eventual) delivery since they can document intent well, and also not end up being written by somebody with less understanding of the API and therefore writing weak tests that don\u0027t properly test the intent of the API.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":223,"context_line":"Testing"},{"line_number":224,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":225,"context_line":""},{"line_number":226,"context_line":"For now we won\u0027t add new Tempest tests, the existing tests should test"},{"line_number":227,"context_line":"the initial changes sufficiently assuming we test a version of Nova"},{"line_number":228,"context_line":"incorporates them and uses the corrrect versioning."},{"line_number":229,"context_line":""},{"line_number":230,"context_line":"Documentation Impact"},{"line_number":231,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_4549eacf","line":228,"range":{"start_line":226,"start_character":0,"end_line":228,"end_character":51},"in_reply_to":"fa7ab95a_930a2720","updated":"2016-09-20 10:51:08.000000000","message":"Yeah, some good functional tests (maybe tempest) seem like a good way to move forward with this.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"560cf015c1ae24647bf6016c765f588c9f7d57b5","unresolved":false,"context_lines":[{"line_number":231,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":232,"context_line":""},{"line_number":233,"context_line":"Will need to update any and all documents about how attach/detach of"},{"line_number":234,"context_line":"volumes in Cinder works."},{"line_number":235,"context_line":""},{"line_number":236,"context_line":"References"},{"line_number":237,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa7ab95a_107bed7b","line":234,"updated":"2016-08-30 16:40:54.000000000","message":"Should add to Nova docs as well.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"d111da1df46c2ce9ac2027efa9a576b9065a8ee6","unresolved":false,"context_lines":[{"line_number":231,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":232,"context_line":""},{"line_number":233,"context_line":"Will need to update any and all documents about how attach/detach of"},{"line_number":234,"context_line":"volumes in Cinder works."},{"line_number":235,"context_line":""},{"line_number":236,"context_line":"References"},{"line_number":237,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a8ec9b2_05260216","line":234,"in_reply_to":"fa7ab95a_107bed7b","updated":"2016-09-20 10:51:08.000000000","message":"Yeah, I am thinking about a Nova spec for the Nova side of things.","commit_id":"a0046d866063841816ed3c248202ef74764e8d03"}],"specs/ocata/add-new-attach-apis.rst":[{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"cf24ba6847c1ef478d19501689bd7545dc2cadc4","unresolved":false,"context_lines":[{"line_number":48,"context_line":"  future attachment, but don\u0027t actually have the connection info yet. This"},{"line_number":49,"context_line":"  would introduce a new volume_state, rather than `attaching` we set the volume"},{"line_number":50,"context_line":"  status to reserved, create an empty attachment record, and return the"},{"line_number":51,"context_line":"  ID of the newly created attachment object to the caller."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  Args:"},{"line_number":54,"context_line":"      volume_id"}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_44eb9ab0","line":51,"updated":"2016-11-15 23:58:14.000000000","message":"There is a cognitive dissonance to this. The API method is called attachment_reserve() but it\u0027s not the attachment that is being reserved at all; it\u0027s the volume. The attachment is created and returned and the API method has a side-effect (!) of setting a volume to \u0027reserved\u0027 status.\n\nHonestly, I\u0027m finding it tough to visualize what the use case is for this. Why does any caller need to get an attachment to any volume before the volume is ready to attach to?","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"6827f9479c7cce670e0d40a926c154b3cd258395","unresolved":false,"context_lines":[{"line_number":48,"context_line":"  future attachment, but don\u0027t actually have the connection info yet. This"},{"line_number":49,"context_line":"  would introduce a new volume_state, rather than `attaching` we set the volume"},{"line_number":50,"context_line":"  status to reserved, create an empty attachment record, and return the"},{"line_number":51,"context_line":"  ID of the newly created attachment object to the caller."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  Args:"},{"line_number":54,"context_line":"      volume_id"}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_87a41f4b","line":51,"in_reply_to":"7a77a97e_44eb9ab0","updated":"2016-11-20 14:35:41.000000000","message":"Good catch. Added a TODO note in the next version of the spec to discuss this.","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":5997,"name":"Walt","display_name":"Hemna","email":"waboring@hemna.com","username":"walter-boring","status":"SAP"},"change_message_id":"e473d68127afd5eaeed92c4bf0fdfb5769c585fa","unresolved":false,"context_lines":[{"line_number":54,"context_line":"      volume_id"},{"line_number":55,"context_line":"      instance_uuid"},{"line_number":56,"context_line":"  Returns:"},{"line_number":57,"context_line":"      Attachment ID"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":""},{"line_number":60,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a629dbe_77ba0370","line":57,"range":{"start_line":57,"start_character":17,"end_line":57,"end_character":19},"updated":"2016-11-08 16:29:25.000000000","message":"yes!","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"cf24ba6847c1ef478d19501689bd7545dc2cadc4","unresolved":false,"context_lines":[{"line_number":54,"context_line":"      volume_id"},{"line_number":55,"context_line":"      instance_uuid"},{"line_number":56,"context_line":"  Returns:"},{"line_number":57,"context_line":"      Attachment ID"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":""},{"line_number":60,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_442aba2d","line":57,"range":{"start_line":57,"start_character":17,"end_line":57,"end_character":19},"in_reply_to":"9a629dbe_77ba0370","updated":"2016-11-15 23:58:14.000000000","message":"I presume the attachment ID is a UUID?","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"cf24ba6847c1ef478d19501689bd7545dc2cadc4","unresolved":false,"context_lines":[{"line_number":72,"context_line":"  Args:"},{"line_number":73,"context_line":"      connector (mandatory)"},{"line_number":74,"context_line":"      volume_id (exclusive option)"},{"line_number":75,"context_line":"      attachemnt_id (exusive option)"},{"line_number":76,"context_line":"  Returns:"},{"line_number":77,"context_line":"      connection_info"},{"line_number":78,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_84ca126d","line":75,"updated":"2016-11-15 23:58:14.000000000","message":"s/attachemnt_id/attachment_id\ns/exusive/exclusive/","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"6827f9479c7cce670e0d40a926c154b3cd258395","unresolved":false,"context_lines":[{"line_number":72,"context_line":"  Args:"},{"line_number":73,"context_line":"      connector (mandatory)"},{"line_number":74,"context_line":"      volume_id (exclusive option)"},{"line_number":75,"context_line":"      attachemnt_id (exusive option)"},{"line_number":76,"context_line":"  Returns:"},{"line_number":77,"context_line":"      connection_info"},{"line_number":78,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_671c636a","line":75,"in_reply_to":"7a77a97e_84ca126d","updated":"2016-11-20 14:35:41.000000000","message":"Done","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"cf24ba6847c1ef478d19501689bd7545dc2cadc4","unresolved":false,"context_lines":[{"line_number":116,"context_line":"Deal with the split brain issue we have between Nova and Cinder."},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"It is possible that we could rewrite the existing methods and enforce stricter"},{"line_number":119,"context_line":"interface usage and we\u0027d probably be ok, but the code has become pertty"},{"line_number":120,"context_line":"convoluted and ugly to deal with and it might be easier to wrap the existing"},{"line_number":121,"context_line":"calls up into a managing wrapper and then over time iterate over the internal"},{"line_number":122,"context_line":"pieces."}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_e4602638","line":119,"updated":"2016-11-15 23:58:14.000000000","message":"s/pertty/pretty/","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"6827f9479c7cce670e0d40a926c154b3cd258395","unresolved":false,"context_lines":[{"line_number":116,"context_line":"Deal with the split brain issue we have between Nova and Cinder."},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"It is possible that we could rewrite the existing methods and enforce stricter"},{"line_number":119,"context_line":"interface usage and we\u0027d probably be ok, but the code has become pertty"},{"line_number":120,"context_line":"convoluted and ugly to deal with and it might be easier to wrap the existing"},{"line_number":121,"context_line":"calls up into a managing wrapper and then over time iterate over the internal"},{"line_number":122,"context_line":"pieces."}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_47112751","line":119,"in_reply_to":"7a77a97e_e4602638","updated":"2016-11-20 14:35:41.000000000","message":"Done","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":5997,"name":"Walt","display_name":"Hemna","email":"waboring@hemna.com","username":"walter-boring","status":"SAP"},"change_message_id":"e473d68127afd5eaeed92c4bf0fdfb5769c585fa","unresolved":false,"context_lines":[{"line_number":125,"context_line":"-----------------"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"We need to store the connector when making attachments, to do this we\u0027ll add an"},{"line_number":128,"context_line":"attachment_specs table that will slurp in the k/v entries in the connector."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"REST API impact"},{"line_number":131,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a629dbe_b7081b2f","line":128,"updated":"2016-11-08 16:29:25.000000000","message":"can we just add a connector column to the volume_attachment table and store the connector dict for each attachment along with the attachment record?","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":2243,"name":"John Griffith","email":"john.griffith8@gmail.com","username":"john-griffith"},"change_message_id":"03bb5cfebd70cd5730fd326fbe6199729e2ea7c4","unresolved":false,"context_lines":[{"line_number":125,"context_line":"-----------------"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"We need to store the connector when making attachments, to do this we\u0027ll add an"},{"line_number":128,"context_line":"attachment_specs table that will slurp in the k/v entries in the connector."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"REST API impact"},{"line_number":131,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9a629dbe_e65ade82","line":128,"in_reply_to":"9a629dbe_b7081b2f","updated":"2016-11-08 19:54:23.000000000","message":"I can change it to do that if there\u0027s an objection to the back reference for some reason.  It just seems to be that having the backref to the attachment_specs serves the same purpose and I don\u0027t have to fight with things like different connectors having different key-names, not to mention the problem of how do you shove a connector into a DB string (ie too many chars)?\n\nYou can see what I\u0027m doing here in the AttachmentSpecs table:  https://review.openstack.org/#/c/387712/2/cinder/db/sqlalchemy/models.py","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"6827f9479c7cce670e0d40a926c154b3cd258395","unresolved":false,"context_lines":[{"line_number":125,"context_line":"-----------------"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"We need to store the connector when making attachments, to do this we\u0027ll add an"},{"line_number":128,"context_line":"attachment_specs table that will slurp in the k/v entries in the connector."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"REST API impact"},{"line_number":131,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"7a77a97e_47ae872e","line":128,"in_reply_to":"9a629dbe_e65ade82","updated":"2016-11-20 14:35:41.000000000","message":"Added a TODO note in the next version of the spec.","commit_id":"8f7b7bb41da674b018143f915a72dfdbe44ae897"},{"author":{"_account_id":2243,"name":"John Griffith","email":"john.griffith8@gmail.com","username":"john-griffith"},"change_message_id":"7f332229eadb9565df97ce553dca5ece10955a6c","unresolved":false,"context_lines":[{"line_number":44,"context_line":""},{"line_number":45,"context_line":"* attachment_reserve"},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"  TODO: Technically we reserve the volume and not the attachment here, but the"},{"line_number":48,"context_line":"  API proposal currently does not match this concept."},{"line_number":49,"context_line":""},{"line_number":50,"context_line":"  This call would be used in scenarios when we want to reserve a volume for a"},{"line_number":51,"context_line":"  future attachment, but don\u0027t actually have the connection info yet. This"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a77a97e_0a43b613","line":48,"range":{"start_line":47,"start_character":0,"end_line":48,"end_character":53},"updated":"2016-11-20 15:07:18.000000000","message":"Yeah, I can see where this might be misleading.  The action here actually is in fact to *reserve* an attachment, and while doing so it also sets the volume-status to reserved as well.  \n\nWe might be able to get away with just acting on the volume only, but I think having the AttachmentUUID to track through the whole process is going to be useful some day.\n\n@jay\nThe use-case was strictly to give Nova back a single reference UUID that could be stored and used for the remainder of the process.  As we move along maybe this doesn\u0027t make as much sense as originally; we should chat with Matt R and John Garbutt and see if this still feels like the right thing.  There\u0027s been a lot of back and forth on the topic of the reserve call in particular so it\u0027s possible that we can do better (or even that what we had here previously was better).","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"44758bbd3b3d82ecd7d9751d275ea922bfd68065","unresolved":false,"context_lines":[{"line_number":44,"context_line":""},{"line_number":45,"context_line":"* attachment_reserve"},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"  TODO: Technically we reserve the volume and not the attachment here, but the"},{"line_number":48,"context_line":"  API proposal currently does not match this concept."},{"line_number":49,"context_line":""},{"line_number":50,"context_line":"  This call would be used in scenarios when we want to reserve a volume for a"},{"line_number":51,"context_line":"  future attachment, but don\u0027t actually have the connection info yet. This"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a77a97e_4a75aec8","line":48,"range":{"start_line":47,"start_character":0,"end_line":48,"end_character":53},"in_reply_to":"7a77a97e_0a43b613","updated":"2016-11-20 16:05:24.000000000","message":"It is a desire in Nova currently to try to fail early in the sense of returning an error right after the API call if a volume cannot be attached for some reason. As Nova does not have the connector on the API level it cannot call attachment-create as connector is mandatory there. If we can step away from this concept then we can remove \u0027reserve\u0027 or anything equal that does the necessary checks although we might still want to lock the volume anyway...\n\nAssuming we will keep this I think we were/are in an agreement that the attachment_id would be useful so that we can track the whole process from step 0. My vote is to keep creating it.","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"6e060c2b2c5a0ea0b054481020e9eb1c5c5e5f8f","unresolved":false,"context_lines":[{"line_number":44,"context_line":""},{"line_number":45,"context_line":"* attachment_reserve"},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"  TODO: Technically we reserve the volume and not the attachment here, but the"},{"line_number":48,"context_line":"  API proposal currently does not match this concept."},{"line_number":49,"context_line":""},{"line_number":50,"context_line":"  This call would be used in scenarios when we want to reserve a volume for a"},{"line_number":51,"context_line":"  future attachment, but don\u0027t actually have the connection info yet. This"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a77a97e_b43575dc","line":48,"range":{"start_line":47,"start_character":0,"end_line":48,"end_character":53},"in_reply_to":"7a77a97e_4a75aec8","updated":"2016-11-22 14:20:02.000000000","message":"We discussed this on the Nov. 21. Cinder-Nova meeting. The meeting log can be found here: http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2016/cinder_nova_api_changes.2016-11-21-17.01.log.html\n\nThe conclusion at the end of the meeting was to either have reserve without creating the attachment and have creation separately or have the ability to create the empty attachment and then update it if needed. In Nova\u0027s use cases it seems better now to rather update the attachment then create a new one every time, e.g. in case of retries.\n\nDo we only need to deal with the naming now and have the use cases touched during the meeting covered or we need to look into the functionality of these calls as well? Like would it make sense to have attachment_create that creates an empty or non-empty attachment based on having the connector or not and have attachment-update that can update it? (was a quick thought to investigate how to have the simple atomic calls, not good if we don\u0027t want to rely on optional parameters at all)","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":2243,"name":"John Griffith","email":"john.griffith8@gmail.com","username":"john-griffith"},"change_message_id":"7f332229eadb9565df97ce553dca5ece10955a6c","unresolved":false,"context_lines":[{"line_number":61,"context_line":"      volume_id"},{"line_number":62,"context_line":"      instance_uuid"},{"line_number":63,"context_line":"  Returns:"},{"line_number":64,"context_line":"      Attachment ID"},{"line_number":65,"context_line":""},{"line_number":66,"context_line":""},{"line_number":67,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a77a97e_aa6002a6","line":64,"updated":"2016-11-20 15:07:18.000000000","message":"Yep, attachment-uuid, I\u0027ll make sure to get those inconsistencies addressed.","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"44758bbd3b3d82ecd7d9751d275ea922bfd68065","unresolved":false,"context_lines":[{"line_number":61,"context_line":"      volume_id"},{"line_number":62,"context_line":"      instance_uuid"},{"line_number":63,"context_line":"  Returns:"},{"line_number":64,"context_line":"      Attachment ID"},{"line_number":65,"context_line":""},{"line_number":66,"context_line":""},{"line_number":67,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a77a97e_8a7b46bd","line":64,"in_reply_to":"7a77a97e_aa6002a6","updated":"2016-11-20 16:05:24.000000000","message":"Done :)","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":23083,"name":"TommyLike","email":"tommylikehu@gmail.com","username":"TommyLike"},"change_message_id":"011d0faae7c9f8e7ab64e6964e5d1489b1868bbd","unresolved":false,"context_lines":[{"line_number":130,"context_line":"Alternatives"},{"line_number":131,"context_line":"------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"Keep trying to hack our way around the existing attach/detach flow we have."},{"line_number":134,"context_line":"Deal with the split brain issue we have between Nova and Cinder."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":"It is possible that we could rewrite the existing methods and enforce stricter"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_72277612","line":133,"range":{"start_line":133,"start_character":62,"end_line":133,"end_character":66},"updated":"2016-11-24 08:08:18.000000000","message":"s/flow/flows","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"54fe6f96d0021117b0dddfe5d4bcf9b1c9a9b0d0","unresolved":false,"context_lines":[{"line_number":130,"context_line":"Alternatives"},{"line_number":131,"context_line":"------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"Keep trying to hack our way around the existing attach/detach flow we have."},{"line_number":134,"context_line":"Deal with the split brain issue we have between Nova and Cinder."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":"It is possible that we could rewrite the existing methods and enforce stricter"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_3cb5d0ed","line":133,"range":{"start_line":133,"start_character":62,"end_line":133,"end_character":66},"in_reply_to":"5a74a57a_72277612","updated":"2016-11-24 18:44:44.000000000","message":"Done","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":23083,"name":"TommyLike","email":"tommylikehu@gmail.com","username":"TommyLike"},"change_message_id":"011d0faae7c9f8e7ab64e6964e5d1489b1868bbd","unresolved":false,"context_lines":[{"line_number":185,"context_line":"Condense initialize_connection and attach down into a single create_attachment"},{"line_number":186,"context_line":"call."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"The intent here is that Cinder  shouldn\u0027t know or care so much about what a"},{"line_number":189,"context_line":"consumer is doing with a volume or attachment.  We should only care that they"},{"line_number":190,"context_line":"desire to attach or detach and that\u0027s about it.  In order to do this properly"},{"line_number":191,"context_line":"we will need an attachment ID that the Consumer should store and reference"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_665b372f","line":188,"range":{"start_line":188,"start_character":30,"end_line":188,"end_character":31},"updated":"2016-11-24 08:08:18.000000000","message":"remove this","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"54fe6f96d0021117b0dddfe5d4bcf9b1c9a9b0d0","unresolved":false,"context_lines":[{"line_number":185,"context_line":"Condense initialize_connection and attach down into a single create_attachment"},{"line_number":186,"context_line":"call."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"The intent here is that Cinder  shouldn\u0027t know or care so much about what a"},{"line_number":189,"context_line":"consumer is doing with a volume or attachment.  We should only care that they"},{"line_number":190,"context_line":"desire to attach or detach and that\u0027s about it.  In order to do this properly"},{"line_number":191,"context_line":"we will need an attachment ID that the Consumer should store and reference"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_9cca646d","line":188,"range":{"start_line":188,"start_character":30,"end_line":188,"end_character":31},"in_reply_to":"5a74a57a_665b372f","updated":"2016-11-24 18:44:44.000000000","message":"Done","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":23083,"name":"TommyLike","email":"tommylikehu@gmail.com","username":"TommyLike"},"change_message_id":"011d0faae7c9f8e7ab64e6964e5d1489b1868bbd","unresolved":false,"context_lines":[{"line_number":188,"context_line":"The intent here is that Cinder  shouldn\u0027t know or care so much about what a"},{"line_number":189,"context_line":"consumer is doing with a volume or attachment.  We should only care that they"},{"line_number":190,"context_line":"desire to attach or detach and that\u0027s about it.  In order to do this properly"},{"line_number":191,"context_line":"we will need an attachment ID that the Consumer should store and reference"},{"line_number":192,"context_line":"going forward when they are done with the attachment.  This means it\u0027s up to"},{"line_number":193,"context_line":"the consumer to decide when they are truly done and they want to destroy the"},{"line_number":194,"context_line":"attachment as opposed to now where we try and guess for them based on clues."}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_a6646fe5","line":191,"range":{"start_line":191,"start_character":38,"end_line":191,"end_character":47},"updated":"2016-11-24 08:08:18.000000000","message":"s/Consumer/consumer","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"54fe6f96d0021117b0dddfe5d4bcf9b1c9a9b0d0","unresolved":false,"context_lines":[{"line_number":188,"context_line":"The intent here is that Cinder  shouldn\u0027t know or care so much about what a"},{"line_number":189,"context_line":"consumer is doing with a volume or attachment.  We should only care that they"},{"line_number":190,"context_line":"desire to attach or detach and that\u0027s about it.  In order to do this properly"},{"line_number":191,"context_line":"we will need an attachment ID that the Consumer should store and reference"},{"line_number":192,"context_line":"going forward when they are done with the attachment.  This means it\u0027s up to"},{"line_number":193,"context_line":"the consumer to decide when they are truly done and they want to destroy the"},{"line_number":194,"context_line":"attachment as opposed to now where we try and guess for them based on clues."}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_1ca574b4","line":191,"range":{"start_line":191,"start_character":38,"end_line":191,"end_character":47},"in_reply_to":"5a74a57a_a6646fe5","updated":"2016-11-24 18:44:44.000000000","message":"Done","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":23083,"name":"TommyLike","email":"tommylikehu@gmail.com","username":"TommyLike"},"change_message_id":"011d0faae7c9f8e7ab64e6964e5d1489b1868bbd","unresolved":false,"context_lines":[{"line_number":215,"context_line":""},{"line_number":216,"context_line":""},{"line_number":217,"context_line":"Testing"},{"line_number":218,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":219,"context_line":""},{"line_number":220,"context_line":"For now we won\u0027t add new Tempest tests, the existing tests should test"},{"line_number":221,"context_line":"the initial changes sufficiently assuming we test a version of Nova"},{"line_number":222,"context_line":"incorporates them and uses the corrrect versioning."},{"line_number":223,"context_line":""},{"line_number":224,"context_line":"Documentation Impact"},{"line_number":225,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_66f7b7f5","line":222,"range":{"start_line":218,"start_character":8,"end_line":222,"end_character":51},"updated":"2016-11-24 08:08:18.000000000","message":"We need testcases focus on these new APIs","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"54fe6f96d0021117b0dddfe5d4bcf9b1c9a9b0d0","unresolved":false,"context_lines":[{"line_number":215,"context_line":""},{"line_number":216,"context_line":""},{"line_number":217,"context_line":"Testing"},{"line_number":218,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":219,"context_line":""},{"line_number":220,"context_line":"For now we won\u0027t add new Tempest tests, the existing tests should test"},{"line_number":221,"context_line":"the initial changes sufficiently assuming we test a version of Nova"},{"line_number":222,"context_line":"incorporates them and uses the corrrect versioning."},{"line_number":223,"context_line":""},{"line_number":224,"context_line":"Documentation Impact"},{"line_number":225,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"5a74a57a_9f26cdb8","line":222,"range":{"start_line":218,"start_character":8,"end_line":222,"end_character":51},"in_reply_to":"5a74a57a_66f7b7f5","updated":"2016-11-24 18:44:44.000000000","message":"Done","commit_id":"ea50bdd1f0a982b4d3fdef2ed4ea13578f759010"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"afeff2ea7c6dcce501b87490af1b13c2dd2d9375","unresolved":false,"context_lines":[{"line_number":63,"context_line":"      volume_id"},{"line_number":64,"context_line":"      instance_uuid"},{"line_number":65,"context_line":"  Returns:"},{"line_number":66,"context_line":"      attachment_id"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":""},{"line_number":69,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"5a74a57a_8a29b28c","line":66,"updated":"2016-11-28 17:43:49.000000000","message":"given this creates an attachment, can\u0027t we just call attachment_create with an empty connector (and host_uuid)?","commit_id":"3b9a1162b57d39a13423a52f806e36b0e660e85d"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"afeff2ea7c6dcce501b87490af1b13c2dd2d9375","unresolved":false,"context_lines":[{"line_number":92,"context_line":"  Args:"},{"line_number":93,"context_line":"      connector (mandatory)"},{"line_number":94,"context_line":"      volume_id (exclusive option)"},{"line_number":95,"context_line":"      attachment_id (exclusive option)"},{"line_number":96,"context_line":"  Returns:"},{"line_number":97,"context_line":"      connection_info"},{"line_number":98,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"5a74a57a_ca564a05","line":95,"updated":"2016-11-28 17:43:49.000000000","message":"don\u0027t we pass the host_uuid here too? What about instance_uuid as well?\n\n(So with live-migrate we can have two attachments, different host_uuid, but the same volume uuid)","commit_id":"3b9a1162b57d39a13423a52f806e36b0e660e85d"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"afeff2ea7c6dcce501b87490af1b13c2dd2d9375","unresolved":false,"context_lines":[{"line_number":111,"context_line":"      {id: \u003cuuid-of-the-attachment-object\u003e,"},{"line_number":112,"context_line":"       project_id: \u003cproject-id-associated-with-attachment\u003e,"},{"line_number":113,"context_line":"       volume_id: \u003cuuid-of-volume-associated-with-this-attachment-record\u003e,"},{"line_number":114,"context_line":"       instance_uuid: \u003cuuid-of-instance-attachment-is-used-for\u003e,"},{"line_number":115,"context_line":"       attached_host: \u003chostname-of-compute-node-attached-to\u003e,"},{"line_number":116,"context_line":"       mountpoint: \u003cmountpoint-inside-instance-that-is-expected-to-be-used\u003e,"},{"line_number":117,"context_line":"       attach_time: \u003ctime-we-created-the-attachment\u003e,"}],"source_content_type":"text/x-rst","patch_set":4,"id":"5a74a57a_2a409e33","line":114,"updated":"2016-11-28 17:43:49.000000000","message":"what about host_uuid?","commit_id":"3b9a1162b57d39a13423a52f806e36b0e660e85d"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"afeff2ea7c6dcce501b87490af1b13c2dd2d9375","unresolved":false,"context_lines":[{"line_number":118,"context_line":"       detach_time: \u003ctime-we-detached\u003e,"},{"line_number":119,"context_line":"       attach_status: \u003cstatus-of-the-attachment\u003e,"},{"line_number":120,"context_line":"       attach_mode: \u003cro|rw\u003e,"},{"line_number":121,"context_line":"       shared_connections_by_attachment: [\u003cattachment_id\u003e, ...]}"},{"line_number":122,"context_line":""},{"line_number":123,"context_line":"* attachment_remove"},{"line_number":124,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"5a74a57a_0fc3f489","line":121,"updated":"2016-11-28 17:43:49.000000000","message":"don\u0027t we need to return the connection_info an connector here, so os-brick can use it?","commit_id":"3b9a1162b57d39a13423a52f806e36b0e660e85d"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"6befdb302635db165cfb1f974604e8aa6617d01c","unresolved":false,"context_lines":[{"line_number":77,"context_line":"  Args:"},{"line_number":78,"context_line":"      connector (mandatory)"},{"line_number":79,"context_line":"      volume_id (exclusive option)"},{"line_number":80,"context_line":"      attachment_id (exclusive option)"},{"line_number":81,"context_line":"  Returns:"},{"line_number":82,"context_line":"      connection_info"},{"line_number":83,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a71b18c_c0eb1daf","line":80,"updated":"2016-12-08 14:15:05.000000000","message":"Isn\u0027t attachment_id mandatory to update?","commit_id":"b71e3e196766c1f213542f31b8f99e23ada2830d"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"cc8532a1c91687d8d46557eb2d7419b1be17051b","unresolved":false,"context_lines":[{"line_number":77,"context_line":"  Args:"},{"line_number":78,"context_line":"      connector (mandatory)"},{"line_number":79,"context_line":"      volume_id (exclusive option)"},{"line_number":80,"context_line":"      attachment_id (exclusive option)"},{"line_number":81,"context_line":"  Returns:"},{"line_number":82,"context_line":"      connection_info"},{"line_number":83,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a71b18c_161d5b66","line":80,"in_reply_to":"3a71b18c_c0eb1daf","updated":"2016-12-08 14:41:23.000000000","message":"Good point, I think it should and I would also guess we trust the caller enough to know what they would like to update, so dropping volume_id makes sense here too.","commit_id":"b71e3e196766c1f213542f31b8f99e23ada2830d"},{"author":{"_account_id":7173,"name":"Scott DAngelo","email":"scott.dangelo@gmail.com","username":"scottda"},"change_message_id":"6befdb302635db165cfb1f974604e8aa6617d01c","unresolved":false,"context_lines":[{"line_number":107,"context_line":""},{"line_number":108,"context_line":"* attachment_remove"},{"line_number":109,"context_line":""},{"line_number":110,"context_line":"  Removes/Detaches an existing attachment, marks a volume as available."},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"  Args:"},{"line_number":113,"context_line":"      attachment_id"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a71b18c_801685a5","line":110,"updated":"2016-12-08 14:15:05.000000000","message":"Probably should be \"...marks a volume as available if this is the last attachment.\"","commit_id":"b71e3e196766c1f213542f31b8f99e23ada2830d"},{"author":{"_account_id":9562,"name":"Ildiko Vancsa","email":"ildiko.vancsa@gmail.com","username":"ildikov"},"change_message_id":"cc8532a1c91687d8d46557eb2d7419b1be17051b","unresolved":false,"context_lines":[{"line_number":107,"context_line":""},{"line_number":108,"context_line":"* attachment_remove"},{"line_number":109,"context_line":""},{"line_number":110,"context_line":"  Removes/Detaches an existing attachment, marks a volume as available."},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"  Args:"},{"line_number":113,"context_line":"      attachment_id"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a71b18c_f6361fe6","line":110,"in_reply_to":"3a71b18c_801685a5","updated":"2016-12-08 14:41:23.000000000","message":"Done","commit_id":"b71e3e196766c1f213542f31b8f99e23ada2830d"}]}
