)]}'
{"nova/conductor/manager.py":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"80393fa8a0363ff2508b5d567b7cd95cc8c9900a","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_76e04ccd","line":602,"updated":"2014-09-05 07:56:59.000000000","message":"While I understand the need you mention, I can\u0027t see an example of where it would be helpful to access that info (I guess in a filter)\n\nWe\u0027re also trying to stop adding blobs to blobs, ie. putting unversioned things to JSON dicts. As this change doesn\u0027t provide why it would be needed, I can\u0027t agree this patch, sorry.","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"f5ed81f17fbae6f093f1690d045a336114bff22d","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_4d5d1848","line":602,"in_reply_to":"fa98f980_18e60617","updated":"2014-09-09 07:47:33.000000000","message":"I discussed with Joe yesterday and we agreed on abandoning this patch.","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":5511,"name":"Nikola Dipanov","email":"ndipanov@redhat.com","username":"ndipanov"},"change_message_id":"af0acaeee173cd483ed3a17e5fffef842b30591f","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_8a62aa14","line":602,"in_reply_to":"fa98f980_3b7dabe8","updated":"2014-09-08 10:06:04.000000000","message":"As discussed on IRC:\n\n* We want to fix the request_spec blob - redesign how we keep that data,\n* It is ok to add stuff to it in case it fixes a bug in tree, we don\u0027t want to wait for the refactoring to happen\n* but this is not used by any of the intree filters so -2","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"50c6c52c645aa0a4014bd9ca43e91dc0e1633723","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_f628bce3","line":602,"in_reply_to":"fa98f980_76e04ccd","updated":"2014-09-05 08:01:28.000000000","message":"Nothing in tree needs it (as of now).  Within our project, we have custom filters that use the scheduling framework and the conductor refactoring now makes it impossible to know which networks are being requested.  Without this, less-than-ideal monkey patching is needed.  :-(","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":5441,"name":"Andrew Laski","email":"andrew@lascii.com","username":"alaski"},"change_message_id":"2d3a83364deea2c23ffcb8830272b5d28be4131b","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_18e60617","line":602,"in_reply_to":"fa98f980_8a62aa14","updated":"2014-09-08 17:10:56.000000000","message":"+1 for what Nikola said.\n\nIf there was code in Nova that used the requested_networks information it would likely not have been removed when this code was refactored.  But as it stands now there\u0027s no reason to add this data here as it is unused.  Adding a simple filter to Nova that uses that information would justify it\u0027s existence.","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"e80fa86aa11981f66a3a9f4f535c91c13fcf6d29","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_3b7dabe8","line":602,"in_reply_to":"fa98f980_8b37304c","updated":"2014-09-06 22:16:27.000000000","message":"BTW, the primary use case for this is...\n\nImagine you\u0027ve got networks A, B and C.\nHost-1: has access to use networks A and B.\nHost-2: has access to use network C.\n\nThis patch allows the scheduler to continue having ability to eliminate the proper hosts given the network being requested.  :)","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"b8167218bb01249d3bc152e7cdccac07b4a8205f","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_f9fa5941","line":602,"in_reply_to":"fa98f980_d65c987f","updated":"2014-09-05 20:45:18.000000000","message":"It might be helpful if you can point to the commit that removed / changed this for context.","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"0c295b6f370196e81a7437b261337f04f1cb41be","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_d65c987f","line":602,"in_reply_to":"fa98f980_f628bce3","updated":"2014-09-05 08:03:59.000000000","message":"So, I was hoping that we might recognize that someone using OpenStack pre-Juno might have written their own scheduler/filter to rely on networks, and without this patch, we\u0027ve taken away capability folks may have needed for their extensions.","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"02dbf58e915965f167a59e2767dec32171b12386","unresolved":false,"context_lines":[{"line_number":599,"context_line":"            requested_networks \u003d objects.NetworkRequestList("},{"line_number":600,"context_line":"                objects\u003d[objects.NetworkRequest.from_tuple(t)"},{"line_number":601,"context_line":"                         for t in requested_networks])"},{"line_number":602,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":603,"context_line":"        try:"},{"line_number":604,"context_line":"            # check retry policy. Rather ugly use of instances[0]..."},{"line_number":605,"context_line":"            # but if we\u0027ve exceeded max retries... then we really only"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa98f980_8b37304c","line":602,"in_reply_to":"fa98f980_f9fa5941","updated":"2014-09-05 21:11:17.000000000","message":"Matt, thanks for adding Andrew to the review as well.\n\nThe issue was introduced when this code was no longer called:\n\nhttps://github.com/openstack/nova/blob/master/nova/scheduler/manager.py#L75\n\nThe scheduler, at this time, had access to \u0027requested_networks\u0027, which could then in turn be used to make scheduling decisions.\n\nNow that only selected_destinations is called via conductor, the scheduler has no visibility to the networks anymore and we\u0027ve (probably unintentionally) taken away any ability for the scheduler to know what networks are involved for potential host filtering.\n\nHoping we can get this in so as to continue giving folks access to this information for more \u0027advanced\u0027 use cases in which network data is critical to scheduling decisions.  :)","commit_id":"7d326bad984db9700f3bf753a0ef18355194579a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c9adeeb6a0b10eac03e2261d8043349badd254d6","unresolved":false,"context_lines":[{"line_number":673,"context_line":"            flavor \u003d objects.Flavor.get_by_id(context, flavor[\u0027id\u0027])"},{"line_number":674,"context_line":"            filter_properties \u003d dict(filter_properties, instance_type\u003dflavor)"},{"line_number":675,"context_line":""},{"line_number":676,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":677,"context_line":"        try:"},{"line_number":678,"context_line":"            scheduler_utils.setup_instance_group(context, request_spec,"},{"line_number":679,"context_line":"                                                 filter_properties)"}],"source_content_type":"text/x-python","patch_set":4,"id":"7a016987_529f335f","line":676,"updated":"2015-05-27 21:06:21.000000000","message":"Since we\u0027re in the middle of implementing blueprint request-spec-object, I would prefer to defer that specific request to once the above is implemented honestly.\n\nAlso, as said mriedem in the dependent patch, I honestly think we need a spec (even if we usually say that a filter doesn\u0027t require it) since we need to agree on a good way to provide that to the Scheduler.","commit_id":"0845dc5101ca03603480d61125facf7175ddf80d"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"f7500a40799845f93d6216803699233a67ddbcfb","unresolved":false,"context_lines":[{"line_number":673,"context_line":"            flavor \u003d objects.Flavor.get_by_id(context, flavor[\u0027id\u0027])"},{"line_number":674,"context_line":"            filter_properties \u003d dict(filter_properties, instance_type\u003dflavor)"},{"line_number":675,"context_line":""},{"line_number":676,"context_line":"        filter_properties[\u0027requested_networks\u0027] \u003d requested_networks"},{"line_number":677,"context_line":"        try:"},{"line_number":678,"context_line":"            scheduler_utils.setup_instance_group(context, request_spec,"},{"line_number":679,"context_line":"                                                 filter_properties)"}],"source_content_type":"text/x-python","patch_set":4,"id":"7a016987_f2be1f63","line":676,"in_reply_to":"7a016987_529f335f","updated":"2015-05-27 21:16:48.000000000","message":"Isn\u0027t this pretty trivial though?  In Icehouse, the requested network was available and it was removed.  And now Stephen is just creating a simple filter for this?\n\nIt doesn\u0027t seem like we\u0027re making any drastic changes for this that should need a blueprint?  I just want to make sure we\u0027re not using a sledgehammer to drive a nail.  :-)","commit_id":"0845dc5101ca03603480d61125facf7175ddf80d"}]}