)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"9cf44db3f11f3718540880b9f1b22b7355f3cc0c","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Dolph Mathews \u003cdolph.mathews@gmail.com\u003e"},{"line_number":5,"context_line":"CommitDate: 2015-10-30 13:35:07 -0500"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Shadow users: unified identity \u0026 multifactor auth"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Change-Id: I94b9345bc29173f90930d1f4a2a3d3244cf24d3b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"1a26ad4f_a21e1bb9","line":7,"updated":"2015-11-01 10:55:03.000000000","message":"the MFA mention here is not necessary","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Dolph Mathews \u003cdolph.mathews@gmail.com\u003e"},{"line_number":5,"context_line":"CommitDate: 2015-10-30 13:35:07 -0500"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Shadow users: unified identity \u0026 multifactor auth"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Change-Id: I94b9345bc29173f90930d1f4a2a3d3244cf24d3b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"da85f559_157e82ae","line":7,"in_reply_to":"1a26ad4f_a21e1bb9","updated":"2015-11-13 18:30:30.000000000","message":"++, shadow users have nothing to do with MFA","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1941,"name":"Lin Hua Cheng","email":"os.lcheng@gmail.com","username":"lin-hua-cheng"},"change_message_id":"e2a4504d219c1bd05f57eef4ac09be919a0a776d","unresolved":false,"context_lines":[{"line_number":5,"context_line":"CommitDate: 2015-10-30 13:35:07 -0500"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Shadow users: unified identity \u0026 multifactor auth"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Change-Id: I94b9345bc29173f90930d1f4a2a3d3244cf24d3b"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"ba8a016a_5dbb8ee5","line":8,"updated":"2015-11-18 06:38:44.000000000","message":"add link to launchpad bp","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"}],"specs/mitaka/shadow-users.rst":[{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_1990ef42","line":15,"updated":"2015-11-04 18:46:50.000000000","message":"Can you please expend on this? It would be nice to list them or provide examples.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"1ad3b58426c817a0669650ba18c2cf58fd6a535a","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_01bdc367","line":15,"in_reply_to":"ba8a016a_2f4553b2","updated":"2015-11-17 19:40:06.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_fc6584b4","line":15,"in_reply_to":"ba8a016a_2f4553b2","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_4e3339f7","line":15,"in_reply_to":"da85f559_198fbd24","updated":"2015-11-13 06:08:58.000000000","message":"it would be helpful to clarify what is meant by \"handled\".  I believe this is what guang-yee is asking to be expanded.  Perhaps you are trying to say the authentication process is handled differently...","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_2f4553b2","line":15,"in_reply_to":"da85f559_4e3339f7","updated":"2015-11-17 16:57:15.000000000","message":".\n\n  Available APIs, relevant APIs, and token validation responses all vary. For example, users receive different types of IDs, may or may not store passwords in keystone, and may or may not be able to receive direct role assignments in keystone. Adding additional authentication methods in the future poses a risk of complicating things further.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_5cc7f095","line":15,"in_reply_to":"da85f559_4e3339f7","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_198fbd24","line":15,"in_reply_to":"fa80f949_1990ef42","updated":"2015-11-08 15:02:14.000000000","message":"Do we really need to?  This is the summary;  let it be a summary.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Locally managed users are handled slightly differently than users backed by"},{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_1ce64844","line":15,"in_reply_to":"fa80f949_1990ef42","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_245fc6e6","line":17,"updated":"2015-11-04 18:46:50.000000000","message":"Not sure if I get this assertion. Authentication method convey information about the authentication payload and how it should be handled. So as long as the API spec is clearly specified, I don\u0027t see a problem.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_7776db3f","line":17,"in_reply_to":"da85f559_1055301c","updated":"2015-11-14 23:11:50.000000000","message":"It complicates it because we keep having to add edge cases for different methods to handle the same problems in multiple ways. New methods keep adding new edge cases.\n\nIt\u0027s a generalized statement, it\u0027s fine in the summary.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_1055301c","line":17,"in_reply_to":"da85f559_ae92e59d","updated":"2015-11-13 18:30:30.000000000","message":"Its only complicated if we don\u0027t have clear and concised documentation. This is true for any API. OpenStack use cases and deployment various to a greater degree. Our design favors flexibility. I don\u0027t see anything wrong with that.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_ae92e59d","line":17,"in_reply_to":"da85f559_d998556d","updated":"2015-11-13 06:08:58.000000000","message":"I think what Adam is trying to say in his comment is that if we added a new type of authentication source we havent thought of yet that one will probably be different as well.  I would just remove the statement or rephrase.  It doesn\u0027t add much and the motivation for the issue is well covered by the previous staements","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":14,"context_line":"LDAP, which are handled significantly differently than users backed by"},{"line_number":15,"context_line":"federation. Available APIs, relevant APIs, and token validation responses all"},{"line_number":16,"context_line":"vary. Adding additional authentication methods in the future poses a risk of"},{"line_number":17,"context_line":"complicating things further."},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_d998556d","line":17,"in_reply_to":"fa80f949_245fc6e6","updated":"2015-11-08 15:02:14.000000000","message":"Statement is true and trivially obvious.  PLease let it stand.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":21,"context_line":"will be a unified experience for both end users and operators."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Problem Description"},{"line_number":24,"context_line":"\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":2,"id":"fa80f949_e40f7edf","line":21,"updated":"2015-11-04 18:46:50.000000000","message":"Amen brother. API consistency matters. Unfortunately, can\u0027t say the same for pagination though. :)","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":21,"context_line":"will be a unified experience for both end users and operators."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Problem Description"},{"line_number":24,"context_line":"\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":2,"id":"da85f559_0ecf71b4","line":21,"updated":"2015-11-13 06:08:58.000000000","message":"I love how this is phrased!  Very nice","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1941,"name":"Lin Hua Cheng","email":"os.lcheng@gmail.com","username":"lin-hua-cheng"},"change_message_id":"e2a4504d219c1bd05f57eef4ac09be919a0a776d","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":20,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":21,"context_line":"will be a unified experience for both end users and operators."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Problem Description"},{"line_number":24,"context_line":"\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":2,"id":"ba8a016a_fd912253","line":21,"in_reply_to":"da85f559_0ecf71b4","updated":"2015-11-18 06:38:44.000000000","message":"I\u0027m sold! :)","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":25,"context_line":""},{"line_number":26,"context_line":"Use cases:"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"- **A consistent user experience.** A federated user should have the same user"},{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_90fa74bc","line":28,"updated":"2015-11-04 16:38:47.000000000","message":"It\u0027ll also make auditing much easier for the federated case","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":25,"context_line":""},{"line_number":26,"context_line":"Use cases:"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"- **A consistent user experience.** A federated user should have the same user"},{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_dc1440a1","line":28,"in_reply_to":"fa80f949_90fa74bc","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":25,"context_line":""},{"line_number":26,"context_line":"Use cases:"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"- **A consistent user experience.** A federated user should have the same user"},{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_5792b74c","line":28,"in_reply_to":"fa80f949_90fa74bc","updated":"2015-11-14 23:11:50.000000000","message":"hopefully it\u0027ll get rid of the concept of a federated user","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":32,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":35,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_64dcee12","line":32,"updated":"2015-11-04 18:46:50.000000000","message":"By definition, they are no longer \"ephemeral users\" if they are automatically provisioned in the shadow table locally upon successful authentication.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":32,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":35,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_8487a1d6","line":32,"in_reply_to":"da85f559_0ed21135","updated":"2015-11-13 18:30:30.000000000","message":"++, point is federation is about letting service providers to have the maximum control. That including not indiscriminately creating shadow users. That should be the choice of the service providers. I know for some deployers, user creation is a workflow. User creation carries a cost, especially for multi-site deployments.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":32,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":35,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_0ed21135","line":32,"in_reply_to":"da85f559_99a3ad9e","updated":"2015-11-13 06:08:58.000000000","message":"I remember very early on before we had real LDAP/federation support Henry used a similar approach. Let\u0027s call it a \"Henry Saturday Night Special\"  I vaguely recall management of the shadow users being nasty and unpleasant. Would be good to get his input to see if there were any real gotchas here.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":29,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":30,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":31,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":32,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":35,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_99a3ad9e","line":32,"in_reply_to":"fa80f949_64dcee12","updated":"2015-11-08 15:02:14.000000000","message":"Right.  Their \"epehmeral nature\" is the current state of things.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":39,"context_line":"  corporate LDAP), a public identity provider (such as a social media profile),"},{"line_number":40,"context_line":"  an X.509 certificate, and an existing account on a remote cloud. If all of"},{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_11ec438e","line":42,"updated":"2015-11-04 19:22:03.000000000","message":"What if all these authentication sources are not equally valid?\n\ne.g --  Passport and Driving license can be used to authenticate, but passport can\u0027t always be used for driving.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":39,"context_line":"  corporate LDAP), a public identity provider (such as a social media profile),"},{"line_number":40,"context_line":"  an X.509 certificate, and an existing account on a remote cloud. If all of"},{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_afd4e3bc","line":42,"in_reply_to":"da85f559_79aa8972","updated":"2015-11-17 16:57:15.000000000","message":"Agree, this isn\u0027t in the scope of this spec. How clients should \"recognize\" the \"validity\" of specific authentication methods is up to them. Keystone doesn\u0027t try to distinguish \"what you know\" from \"what you have\" for example (it\u0027s not communication authentication *factors*, only our nicknames for various methods).\n\nIt\u0027s the difference between how you authenticated:\n\n  methods: [password, sms, mother\u0027s maiden name]\n\nWhat what those authentication factors actually amount to:\n\n  factors: 2","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":39,"context_line":"  corporate LDAP), a public identity provider (such as a social media profile),"},{"line_number":40,"context_line":"  an X.509 certificate, and an existing account on a remote cloud. If all of"},{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_79aa8972","line":42,"in_reply_to":"fa80f949_11ec438e","updated":"2015-11-08 15:02:14.000000000","message":"Then it is specified in the token and we use the \"methods\" field to distinguish.  This is already available, although I don\u0027;t know of anyone taking advantage of it.  Not in the scope of this spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_5f6df796","line":44,"updated":"2015-11-04 18:46:50.000000000","message":"So this is essentially in the realm of linkability versus unlinkability, in internet privacy terminology. A user can also have multiple *accounts*. Our design should be flexibility enough to support both cases.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_4451b961","line":44,"in_reply_to":"da85f559_19b6dddc","updated":"2015-11-13 18:30:30.000000000","message":"need more details on how linking is accomplished","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_d79ea753","line":44,"in_reply_to":"da85f559_4451b961","updated":"2015-11-14 23:11:50.000000000","message":"How you link/unlink should be out of scope. I think we should keep the linked idea here though because i have no doubt that people will achieve this by directly editting the database.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_2fa6b353","line":44,"in_reply_to":"da85f559_d79ea753","updated":"2015-11-17 16:57:15.000000000","message":"Agree, account linking is a complex issue that should wait for the N release. We can move it into a new spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":41,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":42,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_19b6dddc","line":44,"in_reply_to":"fa80f949_5f6df796","updated":"2015-11-08 15:02:14.000000000","message":"It is.  You do not explicitly link accounts you want to keep separate.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"9cf44db3f11f3718540880b9f1b22b7355f3cc0c","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a26ad4f_c21b17c9","line":46,"updated":"2015-11-01 10:55:03.000000000","message":"you mention it under \u0027proposed changes\u0027 but i\u0027d suggest we repeat it here - MFA is out of scope for this spec","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_5f77373a","line":46,"in_reply_to":"1a26ad4f_c21b17c9","updated":"2015-11-04 18:46:50.000000000","message":"I don\u0027t understand this use case either. Keystone does not and should not assert anything. Keystone should only convey the information about the manner (method) in which the user is authentication. If IdP is willing to share that information via additional assertions, Keystone will simply convey it in token data.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_af7063dd","line":46,"in_reply_to":"da85f559_57b957ca","updated":"2015-11-17 16:57:15.000000000","message":"++ Our next steps for MFA should be another spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_57b957ca","line":46,"in_reply_to":"da85f559_84a3214f","updated":"2015-11-14 23:11:50.000000000","message":"++ ignore MFA","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_84a3214f","line":46,"in_reply_to":"da85f559_99cacd4e","updated":"2015-11-13 18:30:30.000000000","message":"huh? REMOTE_USER doesn\u0027t say anything about MFA. MFA is more than just linkable. In some cases, we need to preserve the authentication state. I think we had this discussion when the V3 spec was first introduced. Shadow users and MFA are two different things, lets not lump them together.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":44,"context_line":"  than resulting in distinct identities."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_99cacd4e","line":46,"in_reply_to":"fa80f949_5f77373a","updated":"2015-11-08 15:02:14.000000000","message":"Just that the REMOTE_USER value might be different from different forms of auth, and this spec makes it possible to link them.  This enables a key piece of MFA.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":46,"context_line":"- **Multi-factor authentication.** With multiple types of credentials tied to"},{"line_number":47,"context_line":"  the same user identity, users can authenticate themselves using more than one"},{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"Proposed Change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_cf37e7ae","line":49,"updated":"2015-11-17 16:57:15.000000000","message":"Let\u0027s delete this section from this spec, and move it to a new spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"Proposed Change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_9027d411","line":51,"updated":"2015-11-04 16:38:47.000000000","message":"will /users change to point to the new identity table?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"Proposed Change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_0f596f5a","line":51,"in_reply_to":"da85f559_f7af2b7b","updated":"2015-11-17 16:57:15.000000000","message":"++","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":48,"context_line":"  at a time. Keystone can then make stronger assertions about the identity of"},{"line_number":49,"context_line":"  the user and the path to viable multi-factor authentication is shortened."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"Proposed Change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_f7af2b7b","line":51,"in_reply_to":"fa80f949_9027d411","updated":"2015-11-14 23:11:50.000000000","message":"I think that /users will now solely link to the identity table as this is our source of users. We don\u0027t want to restrict this based on the type of user.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":55,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":56,"context_line":"   table."},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"2. **Shadow federated users.** Create a shadow table for federated users."},{"line_number":59,"context_line":"   Federated users have an `idp_id`, `protocol_id`, a display name, and a"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_3fe113d1","line":56,"updated":"2015-11-04 18:46:50.000000000","message":"This should not preclude us from mapping a federated ephemeral user to a local user right? That is supported today.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":55,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":56,"context_line":"   table."},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"2. **Shadow federated users.** Create a shadow table for federated users."},{"line_number":59,"context_line":"   Federated users have an `idp_id`, `protocol_id`, a display name, and a"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_8e94c144","line":56,"in_reply_to":"da85f559_79c1292e","updated":"2015-11-13 06:08:58.000000000","message":"That\u0027s good to hear!","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":55,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":56,"context_line":"   table."},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"2. **Shadow federated users.** Create a shadow table for federated users."},{"line_number":59,"context_line":"   Federated users have an `idp_id`, `protocol_id`, a display name, and a"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_ef1bcb7e","line":56,"in_reply_to":"da85f559_8e94c144","updated":"2015-11-17 16:57:15.000000000","message":"Agree, the \u0027id_mapping\u0027 table and the \u0027id\u0027 column of our \u0027user\u0027 table must essentially be merged, and that will provide the foundation for this spec. Additionally, you have to move password-related fields into a new table.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":55,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":56,"context_line":"   table."},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"2. **Shadow federated users.** Create a shadow table for federated users."},{"line_number":59,"context_line":"   Federated users have an `idp_id`, `protocol_id`, a display name, and a"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_79c1292e","line":56,"in_reply_to":"fa80f949_3fe113d1","updated":"2015-11-08 15:02:14.000000000","message":"Yeah, should still work as is.  It will be like the id_mapping table today;  id_mapping becomes this implementation,  and then Fedrated gets mapped via that to a real user entry.  If the real entry exists already, map to an existing entry.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_dffec76a","line":64,"updated":"2015-11-04 18:46:50.000000000","message":"What\u0027s the \u0027idp_id\u0027 and \u0027protocol_id\u0027 for LDAP and local users?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"049d3ca524fb2abd573f503264aad947d97de44b","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_caa58bb5","line":64,"in_reply_to":"ba8a016a_41350bd8","updated":"2015-11-17 19:48:19.000000000","message":"Would be good to have a sentence or two in the spec explaining why lifecycle management aspects are of minimal impact/a good problem to have... I missed that conversation at the summit I guess.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_d7c5473d","line":64,"in_reply_to":"da85f559_445bf954","updated":"2015-11-14 23:11:50.000000000","message":"I would clean up this paragraph, i think i know what you mean (based more off discussion than above) but it\u0027s obviously leading to confusion. \n\nThere should not be an idp_id for local users. We have the standard user/identity table and then another table linked from that for \"federated credentials\" or something that is only relevant to federated logins. The local users should not exist on this table.\n\nI think we ignore lifecycle management. If the SAML provider stops authenticating the user then they can still not login to keystone. Having what number of inactive users lying around is a problem? ~10, ~1k, 1M? Do we care? There\u0027s an argument there for audit trail.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6460,"name":"Brad Topol","email":"btopol@us.ibm.com","username":"btopol"},"change_message_id":"cc3fa1692a4dbdb90529a3242c5a7e445eeb74db","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_eeb30de8","line":64,"in_reply_to":"da85f559_99d16dfe","updated":"2015-11-13 06:08:58.000000000","message":"How do we handle lifecycle management of shadow federated users.  How do we clean these up?   Will we end up with orphaned shadow users if they are removed from their true LDAP/IDP source?  Are we gonna end up with chron jobs to now purge the shadow user table???  Need to address this","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_41350bd8","line":64,"in_reply_to":"da85f559_d7c5473d","updated":"2015-11-17 16:57:15.000000000","message":"Tracking for federated users needs to happen in a new table that we don\u0027t have today (perhaps a \"federated user\" table or something, which maps idp_id + protocol_id + unique ID into a local identity ID -- a user_id).\n\nAs for lifecycle management, we generally agreed to ignore it at the summit because it\u0027s a \"good problem to have.\" Worst case, we can provide creation / last used timestamps if it really poses a problem, and allow to deployers to write their own cleanup tooling. If there\u0027s consensus after a couple releases about how to go about cleaning up these records, then it\u0027d be time to discuss adding one of those approaches to keystone -- but not in mitaka.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_445bf954","line":64,"in_reply_to":"da85f559_eeb30de8","updated":"2015-11-13 18:30:30.000000000","message":"Lets not let deployers to decision what the idp_id for the local users. :)","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   data required to identify returning users and provide them with a consistent"},{"line_number":62,"context_line":"   `identity`. Thus, these pieces of data should be tracked in keystone"},{"line_number":63,"context_line":"   regardless of the type of federation involved, and refer to a user"},{"line_number":64,"context_line":"   `identity`, creating one if necessary."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_99d16dfe","line":64,"in_reply_to":"fa80f949_dffec76a","updated":"2015-11-08 15:02:14.000000000","message":"I think we can let the deployer specify those if necessary, ot keep things working as they do today if they do not want to include LDAP into the mix.  In some cases, it might make sense to have LDAP treated as the core user table, as it has lready issued the userid values used in Other serivces (both Nova and Barbican use UserId).  So we might want to allow maing to LDAP backed domains.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_11ba036f","line":68,"updated":"2015-11-04 19:22:03.000000000","message":"example please?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_8197738f","line":68,"in_reply_to":"fa80f949_11ba036f","updated":"2015-11-17 16:57:15.000000000","message":"Going to move account linking into it\u0027s own spec, anyway. Examples will be there!","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_ff386b3f","line":69,"updated":"2015-11-04 18:46:50.000000000","message":"What does linking have anything to do with MFA? Are we going to store the state information for authentication? Also, linkability should be optional, not mandatory.\n\nAlso, can you please specify addition details on how do we link them?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_0f564fd2","line":69,"in_reply_to":"da85f559_c41cc979","updated":"2015-11-17 16:57:15.000000000","message":"Let\u0027s move this section to another spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"509a0c4b8958819077784720b9b1640592280084","unresolved":false,"context_lines":[{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_c41cc979","line":69,"in_reply_to":"da85f559_d9fdb56f","updated":"2015-11-13 18:30:30.000000000","message":"huh? what does that have anything to do with MFA?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":66,"context_line":"3. **Linking authentication sources.** When keystone is presented a valid token"},{"line_number":67,"context_line":"   and a new authentication method it already understands, it should"},{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_d9fdb56f","line":69,"in_reply_to":"fa80f949_ff386b3f","updated":"2015-11-08 15:02:14.000000000","message":"If you have a rule that would map both X509 and Kerberos to the same user, they should be automatically linked.  Auth used to get the token s reflected in the token.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":68,"context_line":"   automatically associate the new credential with the existing `identity`, and"},{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"},{"line_number":73,"context_line":"   an identity provider might require both a password and an SMS verification"},{"line_number":74,"context_line":"   code. When combined with another type of trusted credential on keystone\u0027s"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_cf5e07e9","line":71,"updated":"2015-11-17 16:57:15.000000000","message":"Let\u0027s move this section to another spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":69,"context_line":"   return a token reflecting the multi-factor authentication."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"4. **Multi-factor authentication.** Federated identity providers should be able"},{"line_number":72,"context_line":"   to assert the type of authentication which they have performed. For example,"},{"line_number":73,"context_line":"   an identity provider might require both a password and an SMS verification"},{"line_number":74,"context_line":"   code. When combined with another type of trusted credential on keystone\u0027s"},{"line_number":75,"context_line":"   side such as X.509, keystone should include all three in any resulting"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_ff0f0b86","line":72,"updated":"2015-11-04 18:46:50.000000000","message":"yes, that can be conveyed via additional (SAML) assertions","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":73,"context_line":"   an identity provider might require both a password and an SMS verification"},{"line_number":74,"context_line":"   code. When combined with another type of trusted credential on keystone\u0027s"},{"line_number":75,"context_line":"   side such as X.509, keystone should include all three in any resulting"},{"line_number":76,"context_line":"   tokens, thus establishing true multi-factor tokens. A next step, out of"},{"line_number":77,"context_line":"   scope here, would be to expose multi-factor authentication requirements in"},{"line_number":78,"context_line":"   `policy.json` and enforce them in the policy engine."},{"line_number":79,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_d15a3b50","line":76,"updated":"2015-11-04 19:22:03.000000000","message":"Is multi factor tokens scope of this spec?  I thought this spec is about unified users? ( shadow users)","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":75,"context_line":"   side such as X.509, keystone should include all three in any resulting"},{"line_number":76,"context_line":"   tokens, thus establishing true multi-factor tokens. A next step, out of"},{"line_number":77,"context_line":"   scope here, would be to expose multi-factor authentication requirements in"},{"line_number":78,"context_line":"   `policy.json` and enforce them in the policy engine."},{"line_number":79,"context_line":""},{"line_number":80,"context_line":"Alternatives"},{"line_number":81,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_d40f892c","line":78,"updated":"2015-11-04 19:22:03.000000000","message":"How can you specify multi factor tokens in fernet?  How does the payload look?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":84,"context_line":"either going to result in additional metadata being included in token payloads,"},{"line_number":85,"context_line":"ultimately bloating Fernet tokens beyond their intended capacity, or"},{"line_number":86,"context_line":"increasingly disparate user experiences depending on the user\u0027s authentication"},{"line_number":87,"context_line":"source."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"Security Impact"},{"line_number":90,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_ffa2ab4a","line":87,"updated":"2015-11-04 18:46:50.000000000","message":"It depends, if the additional (SAML) assertions (i.e. auth mechanism to indicate MFA) are unbound, then Fernet will have to adjust to that as well.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":84,"context_line":"either going to result in additional metadata being included in token payloads,"},{"line_number":85,"context_line":"ultimately bloating Fernet tokens beyond their intended capacity, or"},{"line_number":86,"context_line":"increasingly disparate user experiences depending on the user\u0027s authentication"},{"line_number":87,"context_line":"source."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"Security Impact"},{"line_number":90,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_a1d95785","line":87,"in_reply_to":"fa80f949_ffa2ab4a","updated":"2015-11-17 16:57:15.000000000","message":"It already does. Deployers will eventually violate the 255 byte convention.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":89,"context_line":"Security Impact"},{"line_number":90,"context_line":"---------------"},{"line_number":91,"context_line":""},{"line_number":92,"context_line":"* User IDs may be used as a source of authorization in OpenStack, and this spec"},{"line_number":93,"context_line":"  will impact how they are controlled. We need to ensure that two unique users"},{"line_number":94,"context_line":"  cannot be accidentally mapped to the same identity. The simplest solution to"},{"line_number":95,"context_line":"  this is to always assign random UUIDs to serve as the OpenStack-facing user"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_ff706bde","line":92,"updated":"2015-11-04 18:46:50.000000000","message":"Who\u0027s authorizing on user_id today besides Keystone and Barbican?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":89,"context_line":"Security Impact"},{"line_number":90,"context_line":"---------------"},{"line_number":91,"context_line":""},{"line_number":92,"context_line":"* User IDs may be used as a source of authorization in OpenStack, and this spec"},{"line_number":93,"context_line":"  will impact how they are controlled. We need to ensure that two unique users"},{"line_number":94,"context_line":"  cannot be accidentally mapped to the same identity. The simplest solution to"},{"line_number":95,"context_line":"  this is to always assign random UUIDs to serve as the OpenStack-facing user"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_61ba8f0c","line":92,"in_reply_to":"da85f559_b9f55184","updated":"2015-11-17 16:57:15.000000000","message":"Boo. And swift ACLs.\n\nWe won\u0027t break any of that until we get into account linking / unlinking, though.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":89,"context_line":"Security Impact"},{"line_number":90,"context_line":"---------------"},{"line_number":91,"context_line":""},{"line_number":92,"context_line":"* User IDs may be used as a source of authorization in OpenStack, and this spec"},{"line_number":93,"context_line":"  will impact how they are controlled. We need to ensure that two unique users"},{"line_number":94,"context_line":"  cannot be accidentally mapped to the same identity. The simplest solution to"},{"line_number":95,"context_line":"  this is to always assign random UUIDs to serve as the OpenStack-facing user"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_b9f55184","line":92,"in_reply_to":"fa80f949_ff706bde","updated":"2015-11-08 15:02:14.000000000","message":"Nova just started doing it, too, unfortunately.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":97,"context_line":"  other source."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":100,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_f0850007","line":100,"updated":"2015-11-04 16:38:47.000000000","message":"\\o/","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":97,"context_line":"  other source."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":100,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_a16e3794","line":100,"in_reply_to":"da85f559_b93c7130","updated":"2015-11-17 16:57:15.000000000","message":"Agree, we will not be breaking the spec. Instead, we\u0027ll simply stop using a part of the spec in favor of another existing, valid, and more common part of the API :)\n\n... You know, unify them.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":97,"context_line":"  other source."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":100,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_b93c7130","line":100,"in_reply_to":"fa80f949_71358fd8","updated":"2015-11-08 15:02:14.000000000","message":"It will not break the spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"6c5b355bf2b1aa249e2521cbb524f0050e1b385d","unresolved":false,"context_lines":[{"line_number":97,"context_line":"  other source."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":100,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_71358fd8","line":100,"in_reply_to":"fa80f949_f0850007","updated":"2015-11-04 18:46:50.000000000","message":"This design should be backward compatible. Breaking the spec is bad.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":104,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":105,"context_line":"  token which reflects those roles. During the lifetime of that token, an"},{"line_number":106,"context_line":"  operator might assign or revoke additional roles which be immediately"},{"line_number":107,"context_line":"  reflected in the user\u0027s existing token. This behavior is already the case for"},{"line_number":108,"context_line":"  locally-managed users, but will be new to federated users."},{"line_number":109,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_5478798f","line":106,"updated":"2015-11-04 19:22:03.000000000","message":"Unfortunately this behavior is not the case of UUID. You add a role, existing token won\u0027t get the role. Only new will get it.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":104,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":105,"context_line":"  token which reflects those roles. During the lifetime of that token, an"},{"line_number":106,"context_line":"  operator might assign or revoke additional roles which be immediately"},{"line_number":107,"context_line":"  reflected in the user\u0027s existing token. This behavior is already the case for"},{"line_number":108,"context_line":"  locally-managed users, but will be new to federated users."},{"line_number":109,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_01d4e31d","line":106,"in_reply_to":"da85f559_f9399942","updated":"2015-11-17 16:57:15.000000000","message":"Agree with Adam, depending on how we see UUID\u0027s future over the next few releases. At the summit, we noted that the number of advantages of UUID over Fernet is rapidly dwindling, and eventually, will have no advantages over Fernet. It\u0027s just a list of bugs at this point, and all of them are solvable.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":103,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":104,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":105,"context_line":"  token which reflects those roles. During the lifetime of that token, an"},{"line_number":106,"context_line":"  operator might assign or revoke additional roles which be immediately"},{"line_number":107,"context_line":"  reflected in the user\u0027s existing token. This behavior is already the case for"},{"line_number":108,"context_line":"  locally-managed users, but will be new to federated users."},{"line_number":109,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_f9399942","line":106,"in_reply_to":"fa80f949_5478798f","updated":"2015-11-08 15:02:14.000000000","message":"We can rewrite UUID in terms of Fernet and standardize the behavior.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1941,"name":"Lin Hua Cheng","email":"os.lcheng@gmail.com","username":"lin-hua-cheng"},"change_message_id":"e2a4504d219c1bd05f57eef4ac09be919a0a776d","unresolved":false,"context_lines":[{"line_number":113,"context_line":"Auditing notifications which made use of federated identity information in"},{"line_number":114,"context_line":"tokens will no longer have access to that information, beyond multiple"},{"line_number":115,"context_line":"authentication methods presented by v3 tokens. All auditing notifications"},{"line_number":116,"context_line":"will effectively refer to a local user identity."},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"Other End User Impact"},{"line_number":119,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_7d2f72f6","line":116,"updated":"2015-11-18 06:38:44.000000000","message":"++ better UX for the pycadf events!","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Token formats requiring persistence will require less space on disk or in"},{"line_number":136,"context_line":"  memory."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Federated users will be able to consume local role assignments, slightly"},{"line_number":139,"context_line":"  increasing the response time of token validation."},{"line_number":140,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_94b58132","line":137,"updated":"2015-11-04 19:22:03.000000000","message":"How?  How is size of token in DB affected by shadow users?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Token formats requiring persistence will require less space on disk or in"},{"line_number":136,"context_line":"  memory."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Federated users will be able to consume local role assignments, slightly"},{"line_number":139,"context_line":"  increasing the response time of token validation."},{"line_number":140,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_01a2a3ae","line":137,"in_reply_to":"da85f559_5949c5cf","updated":"2015-11-17 16:57:15.000000000","message":"Token response bodies will be smaller, as we don\u0027t have to store federated user information or anything else in the token that we don\u0027t have to store for local users. If token response bodies are stored in the DB (UUID, PKI, PKIZ), then the storage requirements for the DB will be lighter.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Token formats requiring persistence will require less space on disk or in"},{"line_number":136,"context_line":"  memory."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Federated users will be able to consume local role assignments, slightly"},{"line_number":139,"context_line":"  increasing the response time of token validation."},{"line_number":140,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_5949c5cf","line":137,"in_reply_to":"fa80f949_94b58132","updated":"2015-11-08 15:02:14.000000000","message":"No effect.  Only the final userid is reflected in the token.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":143,"context_line":""},{"line_number":144,"context_line":"* No new configuration options."},{"line_number":145,"context_line":""},{"line_number":146,"context_line":"* Local identity persistence will need to be provided, even for deployments"},{"line_number":147,"context_line":"  exclusively using identity federation."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":"* During the upgrade, a database migration will be required to split"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_b0cd385f","line":146,"updated":"2015-11-04 16:38:47.000000000","message":"this is a safe assumption, we already require a database for plenty of other things in keystone, and we\u0027re acting as an identity-persistence/store, not modifying the identities","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":148,"context_line":""},{"line_number":149,"context_line":"* During the upgrade, a database migration will be required to split"},{"line_number":150,"context_line":"  locally-managed users. At runtime, shadow records will be created"},{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_30086891","line":151,"updated":"2015-11-04 16:38:47.000000000","message":"federated \u0026 ldap users*?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":148,"context_line":""},{"line_number":149,"context_line":"* During the upgrade, a database migration will be required to split"},{"line_number":150,"context_line":"  locally-managed users. At runtime, shadow records will be created"},{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_a19f174f","line":151,"in_reply_to":"da85f559_b7d1a3fd","updated":"2015-11-17 16:57:15.000000000","message":"++ Yes! We should cover LDAP users in this spec as well. We had them in mind at the summit, but I didn\u0027t capture that here because I was more interested in federation :) and I think local usernames+passwords and LDAP identities will be much simpler (and less interesting) than the impact on federation.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"118a434f92f70b4d1f030cfc608aadf861805d58","unresolved":false,"context_lines":[{"line_number":148,"context_line":""},{"line_number":149,"context_line":"* During the upgrade, a database migration will be required to split"},{"line_number":150,"context_line":"  locally-managed users. At runtime, shadow records will be created"},{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_b7d1a3fd","line":151,"in_reply_to":"fa80f949_30086891","updated":"2015-11-14 23:11:50.000000000","message":"++ is there any advantage to trying to do this proactively for ldap where there might be user_id specific roles etc already established?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":148,"context_line":""},{"line_number":149,"context_line":"* During the upgrade, a database migration will be required to split"},{"line_number":150,"context_line":"  locally-managed users. At runtime, shadow records will be created"},{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_8243aa41","line":151,"in_reply_to":"fa80f949_30086891","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":220,"name":"Haneef Ali","email":"haneef.ali@hp.com","username":"haneef"},"change_message_id":"6115156d9888c66f6512f058645e2260901f170f","unresolved":false,"context_lines":[{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"Developers supporting alternative authentication methods will be able to"},{"line_number":157,"context_line":"reference the common user identity, and create shadow records themselves."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_94e34133","line":154,"updated":"2015-11-04 19:22:03.000000000","message":"Need more information on listing_users  and user operations based on shadow users?  Specifically w.r.to user experience?\n\nCan I list them?  Can I enable/disable them?  How about update?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":151,"context_line":"  automatically as federated users successfully authenticate with keystone."},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"Developer Impact"},{"line_number":154,"context_line":"----------------"},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"Developers supporting alternative authentication methods will be able to"},{"line_number":157,"context_line":"reference the common user identity, and create shadow records themselves."}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_a1f4f786","line":154,"in_reply_to":"fa80f949_94e34133","updated":"2015-11-17 16:57:15.000000000","message":"Yes, we can\u0027t break any of those APIs. On the backend, we\u0027ll just be querying a new set of tables to answer those API calls, but things like \"id\", \"enabled\", \"description\", and \"email\" (those last two from the extras column) will still be supported and reflected in the API.\n\nI didn\u0027t cover any of that here because I don\u0027t expect to break or change the API at all. This is a refactor on the backend, primarily.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"4a8989d7f203b2d06a534913de406c85244ce913","unresolved":false,"context_lines":[{"line_number":172,"context_line":"Other contributors:"},{"line_number":173,"context_line":"  dstanek"},{"line_number":174,"context_line":""},{"line_number":175,"context_line":"Work Items"},{"line_number":176,"context_line":"----------"},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"1. Separate user identities from their local-managed credentials."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa80f949_904034a7","line":175,"updated":"2015-11-04 16:38:47.000000000","message":"6) notifications emitted when a federated user is authenticated will need to be updated","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":172,"context_line":"Other contributors:"},{"line_number":173,"context_line":"  dstanek"},{"line_number":174,"context_line":""},{"line_number":175,"context_line":"Work Items"},{"line_number":176,"context_line":"----------"},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"1. Separate user identities from their local-managed credentials."}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_212fa7f7","line":175,"in_reply_to":"fa80f949_904034a7","updated":"2015-11-17 16:57:15.000000000","message":"++","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":172,"context_line":"Other contributors:"},{"line_number":173,"context_line":"  dstanek"},{"line_number":174,"context_line":""},{"line_number":175,"context_line":"Work Items"},{"line_number":176,"context_line":"----------"},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"1. Separate user identities from their local-managed credentials."}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_e2439611","line":175,"in_reply_to":"fa80f949_904034a7","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":181,"context_line":""},{"line_number":182,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_0f040fbe","line":184,"updated":"2015-11-17 16:57:15.000000000","message":"Move this item to a second spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":181,"context_line":""},{"line_number":182,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_220c3eec","line":184,"in_reply_to":"ba8a016a_0f040fbe","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"8de184e7a901a4fdf8f772949f4468d5451aefc2","unresolved":false,"context_lines":[{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"Dependencies"},{"line_number":189,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_2f68f386","line":186,"updated":"2015-11-17 16:57:15.000000000","message":"Move this item into a second spec.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":6482,"name":"Steve Martinelli","email":"s.martinelli@gmail.com","username":"stevemar"},"change_message_id":"9cf44db3f11f3718540880b9f1b22b7355f3cc0c","unresolved":false,"context_lines":[{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"Dependencies"},{"line_number":189,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a26ad4f_8280bfc1","line":186,"updated":"2015-11-01 10:55:03.000000000","message":"thought this was out of scope?","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"Dependencies"},{"line_number":189,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_e2115605","line":186,"in_reply_to":"1a26ad4f_8280bfc1","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":2218,"name":"Adam Young","email":"adam@younglogic.com","username":"ayoung"},"change_message_id":"2a99ff5984e71016c2f742288747007c5bd1bc7a","unresolved":false,"context_lines":[{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"Dependencies"},{"line_number":189,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_d90b757c","line":186,"in_reply_to":"1a26ad4f_8280bfc1","updated":"2015-11-08 15:02:14.000000000","message":"I think it needs to be part of the discussion;  it might be implmented implicitly by the earlier steps.","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"8e649b31a5cf4d9bbfd141bd7d4a0469573a3a21","unresolved":false,"context_lines":[{"line_number":183,"context_line":""},{"line_number":184,"context_line":"4. Link multiple authentication sources to a single identity."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"5. Multi-factor authentication."},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"Dependencies"},{"line_number":189,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"ba8a016a_c20c52ee","line":186,"in_reply_to":"ba8a016a_2f68f386","updated":"2015-11-18 20:19:46.000000000","message":"Done","commit_id":"1096a23ef6204633f4523a2bd0de5c74ff6b24f0"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"48b14b4ce0cf579bfb7dc5978240d00a81627ce8","unresolved":false,"context_lines":[{"line_number":31,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":32,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":33,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":34,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":37,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_af420eac","line":34,"updated":"2015-11-20 00:26:41.000000000","message":"I think you can directly map a federate user to a local user.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"b4465507a4104a76ee0d73cf98e3d2ba34e5b3ac","unresolved":false,"context_lines":[{"line_number":31,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":32,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":33,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":34,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":37,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7a740942_b1e1fa00","line":34,"in_reply_to":"9a8ffd7b_1d68b786","updated":"2015-12-03 15:21:59.000000000","message":"Yes,  part of the reason for mapping a federated user to a local user, is to give federated users the same experience as local users.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"9b5fb3928c2d1c78209886da83019920247bb158","unresolved":false,"context_lines":[{"line_number":31,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":32,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":33,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":34,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":37,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_1d68b786","line":34,"in_reply_to":"9a8ffd7b_c722b401","updated":"2015-12-01 23:13:23.000000000","message":"When I did the K2K POC for public cloud, we have to map the user to a local user account because we need to meter on the usage, apply quota, and so on. I can imagine we have more use cases where mapping to an existing local account is preferred because give us more flexibility in tracking and controlling access.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":6534,"name":"David Chadwick","email":"d.w.chadwick@truetrust.co.uk","username":"d-w-chadwick"},"change_message_id":"ca5a6b5533969c27f81b518d5f8d098ada72a08f","unresolved":false,"context_lines":[{"line_number":31,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":32,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":33,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":34,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":37,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_c722b401","line":34,"in_reply_to":"ba8a016a_2151310f","updated":"2015-11-30 17:45:49.000000000","message":"Perhaps one of the reasons for mapping a federated user into a local user, was that the ephemeral user (mapped into a group) had a different experience to a local one. By mapping the federated user into a local user, the federated user gets the same experience.\n\nIf this spec can ensure that ephemeral users and local users get the same experience, then mapping to a local user will no longer be needed.\n\nBut currently work done in Hungary on onboarding new federated users (as part of the GN4 project), had to create local users and map them into it, since ephemeral users did not work correctly.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":8978,"name":"Marek Denis","email":"marek.denis+openstack@gmail.com","username":"marek-denis"},"change_message_id":"156358cbeb523506dd2423eb713ca3df6d86a939","unresolved":false,"context_lines":[{"line_number":31,"context_line":"  experience as a locally-managed user. For example, federated users should be"},{"line_number":32,"context_line":"  able to consume local role assignments just like locally-managed users can."},{"line_number":33,"context_line":"  Today, federated users can only be mapped into user groups, and receive"},{"line_number":34,"context_line":"  tokens with larger payloads to accommodate their ephemeral nature."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"- **One user, multiple credentials.** In the real world, a single person has"},{"line_number":37,"context_line":"  multiple means of authenticating themselves. For example, a person might"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_2151310f","line":34,"in_reply_to":"ba8a016a_af420eac","updated":"2015-11-20 15:15:03.000000000","message":"You can, but then I\u0027d say it\u0027s no longer a \u0027federation\u0027. we are just using SAML and all that stuff as a authN method.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":17860,"name":"Samuel de Medeiros Queiroz","email":"samueldmq@gmail.com","username":"samueldmq"},"change_message_id":"5389933faa89c2f7b1be27e95758eb64036d4958","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  these authentication sources are equally valid, then the resulting user and"},{"line_number":44,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":45,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":46,"context_line":"  than resulting in distinct identities."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"- **Additionally: Facilitates multi-factor authentication \u0026 account linking.**"},{"line_number":49,"context_line":"  With multiple types of credentials tied to the same user identity, users can"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_91857dae","line":46,"updated":"2015-11-25 18:12:19.000000000","message":"Great description","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":6534,"name":"David Chadwick","email":"d.w.chadwick@truetrust.co.uk","username":"d-w-chadwick"},"change_message_id":"ca5a6b5533969c27f81b518d5f8d098ada72a08f","unresolved":false,"context_lines":[{"line_number":44,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":45,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":46,"context_line":"  than resulting in distinct identities."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"- **Additionally: Facilitates multi-factor authentication \u0026 account linking.**"},{"line_number":49,"context_line":"  With multiple types of credentials tied to the same user identity, users can"},{"line_number":50,"context_line":"  authenticate themselves using more than one credential at a time. Keystone"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_a72450c6","line":47,"updated":"2015-11-30 17:45:49.000000000","message":"I would add a caveat here. Quite often, for security reasons, people authenticating by different mechanisms and from different locations are purposefully given different privileges. So you might want to allow a local user to kill VMs, but not if logging in from a remote location via federation.\n\nConsequently the mapping rules should still allow federated users to assume different roles in projects from local users if desired, even if they have the same identity.\n\nOne way to do this would be to indicate the Level of Assurance (LoA) tied into the authentication mechanism. This could also ensure that someone who performed 2FA had more privileges than when he authenticated with a password only.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":44,"context_line":"  operator experience should not vary based on which is chosen. All of these"},{"line_number":45,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":46,"context_line":"  than resulting in distinct identities."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"- **Additionally: Facilitates multi-factor authentication \u0026 account linking.**"},{"line_number":49,"context_line":"  With multiple types of credentials tied to the same user identity, users can"},{"line_number":50,"context_line":"  authenticate themselves using more than one credential at a time. Keystone"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_d1f545ed","line":47,"in_reply_to":"9a8ffd7b_a72450c6","updated":"2015-12-01 20:37:29.000000000","message":"I don\u0027t think the spec conflicts with what David is describing.  We actually would have everything in place except for a policy check.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"0f67d7d9c4db65575ad32e261b1c8ae4a745d76b","unresolved":false,"context_lines":[{"line_number":50,"context_line":"  authenticate themselves using more than one credential at a time. Keystone"},{"line_number":51,"context_line":"  can then make stronger assertions about the identity of the user and the path"},{"line_number":52,"context_line":"  to viable multi-factor authentication is shortened. Likewise, the proposed"},{"line_number":53,"context_line":"  changes supports linking multiple accounts for the same user to a single"},{"line_number":54,"context_line":"  account, as well as simplifies auditing around federated users."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_0e25635a","line":53,"updated":"2015-11-19 22:28:03.000000000","message":"support*","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"0f67d7d9c4db65575ad32e261b1c8ae4a745d76b","unresolved":false,"context_lines":[{"line_number":51,"context_line":"  can then make stronger assertions about the identity of the user and the path"},{"line_number":52,"context_line":"  to viable multi-factor authentication is shortened. Likewise, the proposed"},{"line_number":53,"context_line":"  changes supports linking multiple accounts for the same user to a single"},{"line_number":54,"context_line":"  account, as well as simplifies auditing around federated users."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Proposed Change"},{"line_number":57,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_cefc3bba","line":54,"updated":"2015-11-19 22:28:03.000000000","message":"This doesn\u0027t necessarily tie directly to this specification, but it does make it easier to work towards in the future (thinking about mfa, for example). Calling this out explicitly might be a nice addition.\n\nA follow on sentence might be: \n\n\"All though this specification lays the ground work for MFA and account linking, it is not in scope for this spec.\"","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":17860,"name":"Samuel de Medeiros Queiroz","email":"samueldmq@gmail.com","username":"samueldmq"},"change_message_id":"5389933faa89c2f7b1be27e95758eb64036d4958","unresolved":false,"context_lines":[{"line_number":51,"context_line":"  can then make stronger assertions about the identity of the user and the path"},{"line_number":52,"context_line":"  to viable multi-factor authentication is shortened. Likewise, the proposed"},{"line_number":53,"context_line":"  changes supports linking multiple accounts for the same user to a single"},{"line_number":54,"context_line":"  account, as well as simplifies auditing around federated users."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Proposed Change"},{"line_number":57,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_d1154549","line":54,"in_reply_to":"ba8a016a_284cb175","updated":"2015-11-25 18:12:19.000000000","message":"++","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"cd0494c9422202f9e71605fef8435ace82f17720","unresolved":false,"context_lines":[{"line_number":51,"context_line":"  can then make stronger assertions about the identity of the user and the path"},{"line_number":52,"context_line":"  to viable multi-factor authentication is shortened. Likewise, the proposed"},{"line_number":53,"context_line":"  changes supports linking multiple accounts for the same user to a single"},{"line_number":54,"context_line":"  account, as well as simplifies auditing around federated users."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Proposed Change"},{"line_number":57,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_284cb175","line":54,"in_reply_to":"ba8a016a_2a660c2a","updated":"2015-11-20 15:34:17.000000000","message":"Correct, it is taken out.  This paragraph just says that the refactoring done for this spec will make MFA and account linking easier.  We\u0027ll update the specs to make this more clear.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"48b14b4ce0cf579bfb7dc5978240d00a81627ce8","unresolved":false,"context_lines":[{"line_number":51,"context_line":"  can then make stronger assertions about the identity of the user and the path"},{"line_number":52,"context_line":"  to viable multi-factor authentication is shortened. Likewise, the proposed"},{"line_number":53,"context_line":"  changes supports linking multiple accounts for the same user to a single"},{"line_number":54,"context_line":"  account, as well as simplifies auditing around federated users."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Proposed Change"},{"line_number":57,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_2a660c2a","line":54,"in_reply_to":"ba8a016a_cefc3bba","updated":"2015-11-20 00:26:41.000000000","message":"I thought we agree to take the MFA and linking stuff out, no?","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"0f67d7d9c4db65575ad32e261b1c8ae4a745d76b","unresolved":false,"context_lines":[{"line_number":57,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":60,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":61,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":62,"context_line":"   remove the user table. Modify backend code to utilize the new tables."},{"line_number":63,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_ee7a5f23","line":60,"updated":"2015-11-19 22:28:03.000000000","message":"We will have to refactor the `id_mapping` table as well. Which is already sort-of shadowing ldap users.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"48b14b4ce0cf579bfb7dc5978240d00a81627ce8","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":62,"context_line":"   remove the user table. Modify backend code to utilize the new tables."},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"2. **Shadow LDAP and federated users.** Create new shadow tables for mapping"},{"line_number":65,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":66,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":67,"context_line":"   identity provider. These are the minimal pieces of data required to identify"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_2a34ccf9","line":64,"updated":"2015-11-20 00:26:41.000000000","message":"s/shadow LDAP and federated users/shadow externally-managed users/\n\nThis should not be just LDAP and federated users. Technically, LDAP users are also federated users. (i.e. SSSD configuration)","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"3aaa5ae0553b2b75a5de9741cc6498651bd3e8a1","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":62,"context_line":"   remove the user table. Modify backend code to utilize the new tables."},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"2. **Shadow LDAP and federated users.** Create new shadow tables for mapping"},{"line_number":65,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":66,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":67,"context_line":"   identity provider. These are the minimal pieces of data required to identify"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_f00f6c86","line":64,"in_reply_to":"9a8ffd7b_62626840","updated":"2015-11-30 18:38:20.000000000","message":"You talking about the soon-to-be-deprecated Keystone LDAP write driver? Because even with the per-domain backend, the LDAP users are not managed by Keystone. They are read-only to Keystone. We are already implicitly shadowing them right now. My point is we are basically shadowing all the externally-managed users. Whether they are come from mapping, LDAP, or an auth plugin, The framework needs to be consistent.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":13478,"name":"Boris Bobrov","email":"b.bobrov@sap.com","username":"bbobrov"},"change_message_id":"9258507a0c76e6d7e0fd3f8bc17e3b2427ca7aff","unresolved":false,"context_lines":[{"line_number":61,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":62,"context_line":"   remove the user table. Modify backend code to utilize the new tables."},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"2. **Shadow LDAP and federated users.** Create new shadow tables for mapping"},{"line_number":65,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":66,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":67,"context_line":"   identity provider. These are the minimal pieces of data required to identify"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_62626840","line":64,"in_reply_to":"ba8a016a_2a34ccf9","updated":"2015-11-26 10:10:47.000000000","message":"guang-yee, not really. LDAP users are locally-managed, because, well, we can manage them via keystone. They have domains, after all.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":13478,"name":"Boris Bobrov","email":"b.bobrov@sap.com","username":"bbobrov"},"change_message_id":"9258507a0c76e6d7e0fd3f8bc17e3b2427ca7aff","unresolved":false,"context_lines":[{"line_number":65,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":66,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":67,"context_line":"   identity provider. These are the minimal pieces of data required to identify"},{"line_number":68,"context_line":"   returning users and provide them with a consistent `identity`."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"Alternatives"},{"line_number":71,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_82f50c58","line":68,"updated":"2015-11-26 10:10:47.000000000","message":"What about a table for LDAP users?","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":65,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":66,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":67,"context_line":"   identity provider. These are the minimal pieces of data required to identify"},{"line_number":68,"context_line":"   returning users and provide them with a consistent `identity`."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"Alternatives"},{"line_number":71,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_310371ba","line":68,"in_reply_to":"9a8ffd7b_82f50c58","updated":"2015-12-01 20:37:29.000000000","message":"Currently, yes, we would have a table for LDAP users.  However, as we get into the details, we may be able to generalize this into a single table for Federated and LDAP users.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":6534,"name":"David Chadwick","email":"d.w.chadwick@truetrust.co.uk","username":"d-w-chadwick"},"change_message_id":"ca5a6b5533969c27f81b518d5f8d098ada72a08f","unresolved":false,"context_lines":[{"line_number":75,"context_line":"ultimately bloating Fernet tokens beyond their intended capacity, or"},{"line_number":76,"context_line":"increasingly disparate user experiences depending on the user\u0027s authentication"},{"line_number":77,"context_line":"source."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"Security Impact"},{"line_number":80,"context_line":"---------------"},{"line_number":81,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_8aed91f1","line":78,"updated":"2015-11-30 17:45:49.000000000","message":"Whatever you do, you need to make sure that privileges can be tied to the way the user was authenticated otherwise you are storing up problems for the future.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":75,"context_line":"ultimately bloating Fernet tokens beyond their intended capacity, or"},{"line_number":76,"context_line":"increasingly disparate user experiences depending on the user\u0027s authentication"},{"line_number":77,"context_line":"source."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"Security Impact"},{"line_number":80,"context_line":"---------------"},{"line_number":81,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_b10be17c","line":78,"in_reply_to":"9a8ffd7b_8aed91f1","updated":"2015-12-01 20:37:29.000000000","message":"We\u0027re not adding this feature now, but again, I don\u0027t believe we are preventing or conflicting with this being added in the future.  We would have everything in place, except the policy check.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":6534,"name":"David Chadwick","email":"d.w.chadwick@truetrust.co.uk","username":"d-w-chadwick"},"change_message_id":"ca5a6b5533969c27f81b518d5f8d098ada72a08f","unresolved":false,"context_lines":[{"line_number":85,"context_line":"  this is to always assign random UUIDs to serve as the OpenStack-facing user"},{"line_number":86,"context_line":"  identity. We\u0027ll have to carefully consider the implications of using any"},{"line_number":87,"context_line":"  other source."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":90,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":91,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_2a8ee527","line":88,"updated":"2015-11-30 17:45:49.000000000","message":"You also have to make sure that the same federated user, logging in several times to different sessions over a period of time, has the same user ID each time so that he can continue where he left off.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":85,"context_line":"  this is to always assign random UUIDs to serve as the OpenStack-facing user"},{"line_number":86,"context_line":"  identity. We\u0027ll have to carefully consider the implications of using any"},{"line_number":87,"context_line":"  other source."},{"line_number":88,"context_line":""},{"line_number":89,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":90,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":91,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_6c72807e","line":88,"in_reply_to":"9a8ffd7b_2a8ee527","updated":"2015-12-01 20:37:29.000000000","message":"Yes, the user should have the same experience if logging in several times of a period of time.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"3479d2960e36db8c828e4fdb5e6eb6b8853edf6a","unresolved":false,"context_lines":[{"line_number":93,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":94,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":95,"context_line":"  token which reflects those roles. During the lifetime of that token, an"},{"line_number":96,"context_line":"  operator might assign or revoke additional roles which be immediately"},{"line_number":97,"context_line":"  reflected in the user\u0027s existing token. This behavior is already the case for"},{"line_number":98,"context_line":"  locally-managed users, but will be new to federated users."},{"line_number":99,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_b9b675f7","line":96,"updated":"2015-12-01 16:38:13.000000000","message":"will be*","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"3479d2960e36db8c828e4fdb5e6eb6b8853edf6a","unresolved":false,"context_lines":[{"line_number":126,"context_line":"  memory."},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"* Federated users will be able to consume local role assignments, slightly"},{"line_number":129,"context_line":"  increasing the response time of token validation."},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"Other Deployer Impact"},{"line_number":132,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_59b0a98b","line":129,"updated":"2015-12-01 16:38:13.000000000","message":"Because... we don\u0027t have to do the mapping lookup?","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":6534,"name":"David Chadwick","email":"d.w.chadwick@truetrust.co.uk","username":"d-w-chadwick"},"change_message_id":"0e1c33b8ca96c0663aa0e0524bc186b9d5d90183","unresolved":false,"context_lines":[{"line_number":132,"context_line":"---------------------"},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"* No new configuration options."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":"* Local identity persistence will need to be provided, even for deployments"},{"line_number":137,"context_line":"  exclusively using identity federation."},{"line_number":138,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_bbc980ae","line":135,"updated":"2015-12-01 19:20:53.000000000","message":"You will need to have some housekeeping mechanism to delete federated users periodically otherwise the shadow table will grow uncontrollably. E.g. consider a university that enrolls 20,000 new users each year, and needs to purge them after they graduate 3 years later.\n\nA new config option might be the length of time to wait before an unused shadow entry is automatically purged from the database.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":132,"context_line":"---------------------"},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"* No new configuration options."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":"* Local identity persistence will need to be provided, even for deployments"},{"line_number":137,"context_line":"  exclusively using identity federation."},{"line_number":138,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_2cfc783d","line":135,"in_reply_to":"9a8ffd7b_bbc980ae","updated":"2015-12-01 20:37:29.000000000","message":"There a number of ways to solve this.  We\u0027ll be mindful of it during the detailed design.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":17860,"name":"Samuel de Medeiros Queiroz","email":"samueldmq@gmail.com","username":"samueldmq"},"change_message_id":"5389933faa89c2f7b1be27e95758eb64036d4958","unresolved":false,"context_lines":[{"line_number":138,"context_line":""},{"line_number":139,"context_line":"* During the upgrade, a database migration will be required to split"},{"line_number":140,"context_line":"  locally-managed users. At runtime, shadow records will be created"},{"line_number":141,"context_line":"  automatically as LDAP and federated users successfully authenticate with"},{"line_number":142,"context_line":"  keystone."},{"line_number":143,"context_line":""},{"line_number":144,"context_line":"Developer Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_e74d2b73","line":141,"updated":"2015-11-25 18:12:19.000000000","message":"Same as Guang\u0027s comment at L64, for consistency:\n\ns/LDAP and federated users/externally-managed users","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":8978,"name":"Marek Denis","email":"marek.denis+openstack@gmail.com","username":"marek-denis"},"change_message_id":"156358cbeb523506dd2423eb713ca3df6d86a939","unresolved":false,"context_lines":[{"line_number":176,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"4. Notifications emitted when a federated users is authenticated will need to"},{"line_number":179,"context_line":"   be updated."},{"line_number":180,"context_line":""},{"line_number":181,"context_line":"Dependencies"},{"line_number":182,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba8a016a_42d79423","line":179,"updated":"2015-11-20 15:15:03.000000000","message":"as per \"[...] A federated user might automatically receive a concrete role assignment as the result of the federation mapping and receive a Fernet token which reflects those roles.\" I think a mapping engine will also require a proper refactorization.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":8978,"name":"Marek Denis","email":"marek.denis+openstack@gmail.com","username":"marek-denis"},"change_message_id":"c36ceed8f4196ccd184ef32e10712e2bcfae5af9","unresolved":false,"context_lines":[{"line_number":176,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"4. Notifications emitted when a federated users is authenticated will need to"},{"line_number":179,"context_line":"   be updated."},{"line_number":180,"context_line":""},{"line_number":181,"context_line":"Dependencies"},{"line_number":182,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_e7bbeb73","line":179,"in_reply_to":"9a8ffd7b_879e9f52","updated":"2015-12-01 20:51:34.000000000","message":"I think mapping engine will be responsible for computing unique and constant user_id  as well as it\u0027s roles. which looks like a core feature of shadow users. The rest needs to stay for backwards compatibility .","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"68ebe6ed527bd35d200326d75890bbbeb5cb5189","unresolved":false,"context_lines":[{"line_number":176,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"4. Notifications emitted when a federated users is authenticated will need to"},{"line_number":179,"context_line":"   be updated."},{"line_number":180,"context_line":""},{"line_number":181,"context_line":"Dependencies"},{"line_number":182,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_879e9f52","line":179,"in_reply_to":"ba8a016a_42d79423","updated":"2015-12-01 20:48:36.000000000","message":"This means that the mapping engine will have to be refactored to handle: \n\n- mapping to group ids\n- mapping to local users (implemented in the federated direct user mapping spec)\n- mapping to shadow users (possibly the same as above?)\n\nIs that what you mean?","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9527cdb8c69fa4413ca26f6a85d05b28c9d5f449","unresolved":false,"context_lines":[{"line_number":176,"context_line":"3. Make concrete role assignments after federated mapping is evaluated."},{"line_number":177,"context_line":""},{"line_number":178,"context_line":"4. Notifications emitted when a federated users is authenticated will need to"},{"line_number":179,"context_line":"   be updated."},{"line_number":180,"context_line":""},{"line_number":181,"context_line":"Dependencies"},{"line_number":182,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_0702af09","line":179,"in_reply_to":"ba8a016a_42d79423","updated":"2015-12-01 20:37:29.000000000","message":"Yes, most likely it will require some refactoring.","commit_id":"eacc832531ac6f260a484fef06f5abc41c0c465c"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"7aa8f2c8d469356d253e557afe2ae67514023f6f","unresolved":false,"context_lines":[{"line_number":18,"context_line":"able to receive direct role assignments. Future additional authentication"},{"line_number":19,"context_line":"methods pose a risk of complicating things further."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":22,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":23,"context_line":"will be a unified experience for both end users and operators."},{"line_number":24,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_4869038b","line":21,"updated":"2015-12-03 16:25:03.000000000","message":"(nit) using refactor for describing a rewrite is a pet peeve of mine","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"be07304fa235fdcc291a1abe27aa32683567d2b4","unresolved":false,"context_lines":[{"line_number":18,"context_line":"able to receive direct role assignments. Future additional authentication"},{"line_number":19,"context_line":"methods pose a risk of complicating things further."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":22,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":23,"context_line":"will be a unified experience for both end users and operators."},{"line_number":24,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_18f34a75","line":21,"in_reply_to":"7a740942_15e233d6","updated":"2015-12-03 17:00:47.000000000","message":"Refactoring is changing the implementation without changing the external interfaces or behavior. The whole point of this is to change the behavior, to make all the types of users the same.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"2b1315687b4e2977b387b48ad20c6f4192b49d22","unresolved":false,"context_lines":[{"line_number":18,"context_line":"able to receive direct role assignments. Future additional authentication"},{"line_number":19,"context_line":"methods pose a risk of complicating things further."},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Instead of continuing down this path, we can refactor our user persistence to"},{"line_number":22,"context_line":"separate identities from their locally-managed credentials, if any. The result"},{"line_number":23,"context_line":"will be a unified experience for both end users and operators."},{"line_number":24,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_15e233d6","line":21,"in_reply_to":"7a740942_4869038b","updated":"2015-12-03 16:49:07.000000000","message":"I really don\u0027t think of this as a rewrite -- it should ultimately be migrations to shuffle our backend models around and code changes to support the new data model.\n\nCan you explain your reasoning for thinking of this as a rewrite?","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"3b9bd442fafd61871e4109e313dd198612ba78ca","unresolved":false,"context_lines":[{"line_number":45,"context_line":"  means of authentications should be tied to the same user identity, rather"},{"line_number":46,"context_line":"  than resulting in distinct identities."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"  Note: We discussed the potential future need to tie privileges to"},{"line_number":49,"context_line":"  authentication methods. This is out-of-scope for this spec, however the spec"},{"line_number":50,"context_line":"  does not prevent or conflict from this feature being added in the future."},{"line_number":51,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_78f84e7d","line":48,"updated":"2015-12-03 18:57:18.000000000","message":"(nit) s/potential future/potential/","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"7aa8f2c8d469356d253e557afe2ae67514023f6f","unresolved":false,"context_lines":[{"line_number":56,"context_line":"  to viable multi-factor authentication (MFA) is shortened. Likewise, the"},{"line_number":57,"context_line":"  proposed changes support linking multiple accounts for the same user to a"},{"line_number":58,"context_line":"  single account, as well as simplifies auditing around federated users. While"},{"line_number":59,"context_line":"  this specification does not solve MFA or account linking, the refactoring"},{"line_number":60,"context_line":"  done for this spec will make the development for those features easier."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_88a04b92","line":59,"updated":"2015-12-03 16:25:03.000000000","message":"rewriting","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"a9d531a85478d74dd664617e206145dbf9bf0e16","unresolved":false,"context_lines":[{"line_number":56,"context_line":"  to viable multi-factor authentication (MFA) is shortened. Likewise, the"},{"line_number":57,"context_line":"  proposed changes support linking multiple accounts for the same user to a"},{"line_number":58,"context_line":"  single account, as well as simplifies auditing around federated users. While"},{"line_number":59,"context_line":"  this specification does not solve MFA or account linking, the refactoring"},{"line_number":60,"context_line":"  done for this spec will make the development for those features easier."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_aeb3cc09","line":59,"in_reply_to":"7a740942_88a04b92","updated":"2015-12-03 17:30:03.000000000","message":"Refactoring, rewriting, what\u0027s the difference? Pardon my ESL English. :)","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"7aa8f2c8d469356d253e557afe2ae67514023f6f","unresolved":false,"context_lines":[{"line_number":62,"context_line":"Proposed Change"},{"line_number":63,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":66,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":67,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":68,"context_line":"   remove the user table. Modify backend code to utilize the new tables."}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_883ccb30","line":65,"updated":"2015-12-03 16:25:03.000000000","message":"s/Refactor/Split/","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"3b9bd442fafd61871e4109e313dd198612ba78ca","unresolved":false,"context_lines":[{"line_number":62,"context_line":"Proposed Change"},{"line_number":63,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":66,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":67,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":68,"context_line":"   remove the user table. Modify backend code to utilize the new tables."}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_f8943ee9","line":65,"in_reply_to":"7a740942_756db731","updated":"2015-12-03 18:57:18.000000000","message":"(^nit)","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"2b1315687b4e2977b387b48ad20c6f4192b49d22","unresolved":false,"context_lines":[{"line_number":62,"context_line":"Proposed Change"},{"line_number":63,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"1. **Separate user identities from their local-managed credentials.** Refactor"},{"line_number":66,"context_line":"   `user` table into an `identity` table and a locally-managed `password`"},{"line_number":67,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":68,"context_line":"   remove the user table. Modify backend code to utilize the new tables."}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_756db731","line":65,"in_reply_to":"7a740942_883ccb30","updated":"2015-12-03 16:49:07.000000000","message":"++ on this instance","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"3b9bd442fafd61871e4109e313dd198612ba78ca","unresolved":false,"context_lines":[{"line_number":67,"context_line":"   table. Migrate data from the user table to these new tables and ultimately"},{"line_number":68,"context_line":"   remove the user table. Modify backend code to utilize the new tables."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"2. **Shadow LDAP and federated users.** Create new shadow tables for mapping"},{"line_number":71,"context_line":"   LDAP and federated users to local identities. Federated users have a"},{"line_number":72,"context_line":"   `idp_id`, `protocol_id`, display name, and a unique ID asserted by the"},{"line_number":73,"context_line":"   identity provider. These are the minimal pieces of data required to identify"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_985bba54","line":70,"updated":"2015-12-03 18:57:18.000000000","message":"(nit) In patchset 4, guang suggested s/shadow LDAP and federated users/shadow externally-managed users/ which I also prefer.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":18338,"name":"Ron De Rose","email":"ronald.de.rose@intel.com","username":"derosenet"},"change_message_id":"9207a4ee3fd421d1ecec0e52848420dc4e041028","unresolved":false,"context_lines":[{"line_number":73,"context_line":"   identity provider. These are the minimal pieces of data required to identify"},{"line_number":74,"context_line":"   returning users and provide them with a consistent `identity`. Likewise,"},{"line_number":75,"context_line":"   LDAP users have a \u0027domain_id\u0027 and \u0027dn\u0027 used to identify the user. That being"},{"line_number":76,"context_line":"   said, there may be an opportunity to generalize LDAP and federated users"},{"line_number":77,"context_line":"   a single table. This will be solved during development."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_a81f475a","line":76,"updated":"2015-12-03 16:13:47.000000000","message":"into*","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"2b1315687b4e2977b387b48ad20c6f4192b49d22","unresolved":false,"context_lines":[{"line_number":73,"context_line":"   identity provider. These are the minimal pieces of data required to identify"},{"line_number":74,"context_line":"   returning users and provide them with a consistent `identity`. Likewise,"},{"line_number":75,"context_line":"   LDAP users have a \u0027domain_id\u0027 and \u0027dn\u0027 used to identify the user. That being"},{"line_number":76,"context_line":"   said, there may be an opportunity to generalize LDAP and federated users"},{"line_number":77,"context_line":"   a single table. This will be solved during development."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_f89afec4","line":76,"in_reply_to":"7a740942_a81f475a","updated":"2015-12-03 16:49:07.000000000","message":"(nit) ++","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"7aa8f2c8d469356d253e557afe2ae67514023f6f","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_489023ed","line":101,"updated":"2015-12-03 16:25:03.000000000","message":"Does this mean that for every authentication we\u0027ll have to make sure the concrete group associations and roles match the assertion?","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":6804,"name":"bruce-benjamin","email":"bruce.benjamin@jhuapl.edu","username":"bruce-benjamin"},"change_message_id":"f6cd3e3982f089a3bad5ebf840779f8454eef404","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_c33f14d5","line":101,"in_reply_to":"7a740942_489023ed","updated":"2015-12-03 16:40:59.000000000","message":"+1","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"2b1315687b4e2977b387b48ad20c6f4192b49d22","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_5848f25b","line":101,"in_reply_to":"7a740942_489023ed","updated":"2015-12-03 16:49:07.000000000","message":"That\u0027s a good question I was going to leave for implementation, but I\u0027ll state my preference here. The two approaches:\n\nA) The simplest behavior: apply the mapping (group memberships) once, when the \"shadow\" is initially created. This is like bootstrapping the shadow user using the data from the mapping.\n\nB) The complex, bug-prone behavior (IMO): apply the mapping (group memberships) on each authentication. This is like using the mapping as a declarative model of the user\u0027s authorization state. But this leads to further questions; such as, should the mapping output be able to revoke group memberships that were formerly applied by a mapping? Obviously, I\u0027m not in favor of this approach :)","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7191,"name":"Jamie Lennox","email":"jamielennox@gmail.com","username":"jamielennox"},"change_message_id":"2c036a835cf7f5d3ddf64cfb5b1587deb1e09efe","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_9b3b4d9c","line":101,"in_reply_to":"7a740942_4e46e8c0","updated":"2015-12-07 05:44:22.000000000","message":"can we not make it configurable unless someone really needs it. There are already too many switches for subtly different behaviour that we need to test. \n\nI\u0027d be happy with A) until there is an explicit need for something more. \n\nI haven\u0027t put a whole lot of thought into this but if you need something more complex here now that federated users are not ephemeral then should we be looking at this control coming from a new/additional assignment backend instead of the mapping?","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":7725,"name":"David Stanek","email":"dstanek@dstanek.com","username":"dstanek"},"change_message_id":"be07304fa235fdcc291a1abe27aa32683567d2b4","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_780cce42","line":101,"in_reply_to":"7a740942_5848f25b","updated":"2015-12-03 17:00:47.000000000","message":"Would it make sense to put you comment in the doc? I wonder is here are other systems that do this sort of thing. That way we know if the expectation for an operator is that changing group roles in AD will change the group roles in OS or if they will have to script it.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":1916,"name":"Guang Yee","email":"gyee@suse.com","username":"guang-yee"},"change_message_id":"a9d531a85478d74dd664617e206145dbf9bf0e16","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_4e46e8c0","line":101,"in_reply_to":"7a740942_780cce42","updated":"2015-12-03 17:30:03.000000000","message":"How about make it configurable? This should give deployers options to mitigate risk. If we have multiple DB transactions on auth, making it atomic will be an adventure.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":4,"name":"Dolph Mathews","email":"dolph.mathews@gmail.com","username":"dolph"},"change_message_id":"fa541db97b24f457337d9bbbb483cad0566e19dc","unresolved":false,"context_lines":[{"line_number":98,"context_line":"* We\u0027ll gain the ability to remove information about federated users from"},{"line_number":99,"context_line":"  tokens, thus eliminating the concept of \"federated tokens\" altogether."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Federated user group membership will become persistent rather than ephemeral"},{"line_number":102,"context_line":"  per authentication. A federated user might automatically receive a concrete"},{"line_number":103,"context_line":"  role assignment as the result of the federation mapping and receive a Fernet"},{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_48a88720","line":101,"in_reply_to":"7a740942_9b3b4d9c","updated":"2015-12-07 22:40:55.000000000","message":"Unless we have conflicting use cases and decide it\u0027s worth the cost to properly support them both, I agree with Jamie.\n\nJamie: As for where this code lives, I don\u0027t really care. Mapping or assignment: they\u0027re both handling authorization management / CRUD already.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"},{"author":{"_account_id":8978,"name":"Marek Denis","email":"marek.denis+openstack@gmail.com","username":"marek-denis"},"change_message_id":"6679d6fd79182a13732bc1f4ee9537ca8da69fbe","unresolved":false,"context_lines":[{"line_number":104,"context_line":"  token which reflects those roles. During the lifetime of that token, an"},{"line_number":105,"context_line":"  operator might assign or revoke additional roles which will be immediately"},{"line_number":106,"context_line":"  reflected in the user\u0027s existing token. This behavior is already the case for"},{"line_number":107,"context_line":"  locally-managed users, but will be new to federated users."},{"line_number":108,"context_line":""},{"line_number":109,"context_line":"Notifications Impact"},{"line_number":110,"context_line":"--------------------"}],"source_content_type":"text/x-rst","patch_set":5,"id":"7a740942_a0d654c3","line":107,"updated":"2015-12-04 10:17:00.000000000","message":"++, however I reckon change of role assignments will not be able only during token lifetime as multiple tokens will be issued for a single shadow user.","commit_id":"16398bc5abf2756f071b759b60c5c50e7349ecaa"}]}
