)]}'
{"doc/source/cells.rst":[{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"66cae08277e67b0c1f5da295a0bb19002ab6f9d6","unresolved":false,"context_lines":[{"line_number":185,"context_line":"instance_groups"},{"line_number":186,"context_line":"instance_group_member"},{"line_number":187,"context_line":"instance_group_policy"},{"line_number":188,"context_line":"compute_nodes"},{"line_number":189,"context_line":"quotas"},{"line_number":190,"context_line":"project_user_quotas"},{"line_number":191,"context_line":"quota_classes"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a5ae5dd_d87dca6d","line":188,"updated":"2016-02-08 19:53:24.000000000","message":"I had been thinking this would be a cell-level table","commit_id":"a5665e549290a9617a4c563868b08e6b04ace269"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"898af8ef2718e5a12a596973c168b46873a6384f","unresolved":false,"context_lines":[{"line_number":185,"context_line":"instance_groups"},{"line_number":186,"context_line":"instance_group_member"},{"line_number":187,"context_line":"instance_group_policy"},{"line_number":188,"context_line":"compute_nodes"},{"line_number":189,"context_line":"quotas"},{"line_number":190,"context_line":"project_user_quotas"},{"line_number":191,"context_line":"quota_classes"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a5ae5dd_187fe2e0","line":188,"in_reply_to":"5a5ae5dd_d87dca6d","updated":"2016-02-08 20:06:31.000000000","message":"I think that was the decision made in IRC also. The resource provider usage tables are going in the cell, makes sense for this to go there also.","commit_id":"a5665e549290a9617a4c563868b08e6b04ace269"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"73dc2b06a04586a0fb8afbaca3ea5562573d1243","unresolved":false,"context_lines":[{"line_number":220,"context_line":"instance_actions"},{"line_number":221,"context_line":"instance_actions_events"},{"line_number":222,"context_line":"instance_id_mappings"},{"line_number":223,"context_line":"migrations"},{"line_number":224,"context_line":"pci_devices"},{"line_number":225,"context_line":"block_device_mapping"},{"line_number":226,"context_line":"snapshots"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a5ae5dd_b8424e66","line":223,"updated":"2016-02-08 20:13:53.000000000","message":"The biggest difference to the previous list (apart from adding missing tables) is that I have moved migrations to the cell-level table. For now, where we arn\u0027t allowing cross-cell migrations this is preferable.\n\nI think it should go in the cell level table, as this removes a foreign key problem. None of the infrastructure for cross-cell migrations are there anyway. We should cross this bridge when we come to it?\n\nOn the other hand leaving it in the API table and doing the work to fix the foreign key issue might be more future friendly.","commit_id":"a5665e549290a9617a4c563868b08e6b04ace269"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"0a2c1da0748ff709dbdf4cc76104e569e14075f1","unresolved":false,"context_lines":[{"line_number":175,"context_line":""},{"line_number":176,"context_line":"agent_builds"},{"line_number":177,"context_line":"s3_images"},{"line_number":178,"context_line":"services"},{"line_number":179,"context_line":"aggregate_hosts"},{"line_number":180,"context_line":"aggregate_metadata"},{"line_number":181,"context_line":"aggregates"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a57f1b5_393cb7c7","line":178,"updated":"2016-02-11 08:41:55.000000000","message":"Given the denormalized relationship we have with compute_nodes (see for example objects.ServiceList.get_by_host() and given also that a service is only a oslo_service/RPC component, it should be really in the cells DB IMHO.","commit_id":"a2d02cf7123790e4860bc70ddc72ab79457e7166"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"61f299af1b179a5ae31e81d427d3be6e93ef0a1d","unresolved":false,"context_lines":[{"line_number":175,"context_line":""},{"line_number":176,"context_line":"agent_builds"},{"line_number":177,"context_line":"s3_images"},{"line_number":178,"context_line":"services"},{"line_number":179,"context_line":"aggregate_hosts"},{"line_number":180,"context_line":"aggregate_metadata"},{"line_number":181,"context_line":"aggregates"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a57f1b5_508804c5","line":178,"in_reply_to":"3a57f1b5_393cb7c7","updated":"2016-02-12 19:13:09.000000000","message":"Thanks Sylvain good to know. Will move back to cell tables. I created this list only looking at foreign key relationships and my best guess. Denormalized data in multiple tables won\u0027t show up so its really good to have you look over it.","commit_id":"a2d02cf7123790e4860bc70ddc72ab79457e7166"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"1466f2a74301362d6894f6e9e99161f9e1049974","unresolved":false,"context_lines":[{"line_number":175,"context_line":""},{"line_number":176,"context_line":"agent_builds"},{"line_number":177,"context_line":"s3_images"},{"line_number":178,"context_line":"aggregate_hosts"},{"line_number":179,"context_line":"aggregate_metadata"},{"line_number":180,"context_line":"aggregates"},{"line_number":181,"context_line":"instance_types"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_edae2397","line":178,"updated":"2016-03-02 11:51:39.000000000","message":"It feels like the hosts should live in the child cell to me, as when we add extra cells this table will grow.\n\nWe could use uuids in the aggregate to map these between the two tables, I think?\n\nI agreed the aggregate and metadata live at the top.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"b671c625f7b54e66b8b1a88c68a96514e63d82a1","unresolved":false,"context_lines":[{"line_number":175,"context_line":""},{"line_number":176,"context_line":"agent_builds"},{"line_number":177,"context_line":"s3_images"},{"line_number":178,"context_line":"aggregate_hosts"},{"line_number":179,"context_line":"aggregate_metadata"},{"line_number":180,"context_line":"aggregates"},{"line_number":181,"context_line":"instance_types"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_9798d31d","line":178,"in_reply_to":"7af24918_edae2397","updated":"2016-03-03 21:52:52.000000000","message":"There might be some performance issues with keeping aggregate_hosts in the cell database, but right now they are minimal. Basically when calling AggregateList.get_all we need to make a call for every cell database. This function isn\u0027t called that often. Will keep this in cells  table for now.\n\nI\u0027m less worried about this being in the cell db than the others you flagged though. It only grows with the number of hosts as opposed to instances.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"1466f2a74301362d6894f6e9e99161f9e1049974","unresolved":false,"context_lines":[{"line_number":182,"context_line":"instance_type_projects"},{"line_number":183,"context_line":"instance_type_extra_specs"},{"line_number":184,"context_line":"instance_groups"},{"line_number":185,"context_line":"instance_group_member"},{"line_number":186,"context_line":"instance_group_policy"},{"line_number":187,"context_line":"quotas"},{"line_number":188,"context_line":"project_user_quotas"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_0db1cfe1","line":185,"updated":"2016-03-02 11:51:39.000000000","message":"it feels like the members could be in the instance DBs, because of how this table will grow as you add extra cells?\n\nA bit like aggregates?","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"b671c625f7b54e66b8b1a88c68a96514e63d82a1","unresolved":false,"context_lines":[{"line_number":182,"context_line":"instance_type_projects"},{"line_number":183,"context_line":"instance_type_extra_specs"},{"line_number":184,"context_line":"instance_groups"},{"line_number":185,"context_line":"instance_group_member"},{"line_number":186,"context_line":"instance_group_policy"},{"line_number":187,"context_line":"quotas"},{"line_number":188,"context_line":"project_user_quotas"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_d711bb5e","line":185,"in_reply_to":"7af24918_0db1cfe1","updated":"2016-03-03 21:52:52.000000000","message":"Agree, this will grow with number of instances across cells....so should go in the cell db. Its strange that the instance_id doesn\u0027t have a FK relationship to the instances table otherwise I might have looked at this more closely.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"1466f2a74301362d6894f6e9e99161f9e1049974","unresolved":false,"context_lines":[{"line_number":192,"context_line":"security_groups"},{"line_number":193,"context_line":"security_group_rules"},{"line_number":194,"context_line":"security_group_default_rules"},{"line_number":195,"context_line":"security_group_instance_association"},{"line_number":196,"context_line":"provider_fw_rules"},{"line_number":197,"context_line":"key_pairs"},{"line_number":198,"context_line":"certificates"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_4dd2b701","line":195,"updated":"2016-03-02 11:51:39.000000000","message":"Hmm, I wonder if this should live in the cell DB, as above.\n\nAlthough this is basically only for nova-network, that we are trying to delete, so I worry less about that.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"b671c625f7b54e66b8b1a88c68a96514e63d82a1","unresolved":false,"context_lines":[{"line_number":192,"context_line":"security_groups"},{"line_number":193,"context_line":"security_group_rules"},{"line_number":194,"context_line":"security_group_default_rules"},{"line_number":195,"context_line":"security_group_instance_association"},{"line_number":196,"context_line":"provider_fw_rules"},{"line_number":197,"context_line":"key_pairs"},{"line_number":198,"context_line":"certificates"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_f294c58b","line":195,"in_reply_to":"7af24918_4dd2b701","updated":"2016-03-03 21:52:52.000000000","message":"Will move. This was already flagged as an issue by a FK violation. We have had discussions about keeping it in cell db during previous cells meetings.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"1466f2a74301362d6894f6e9e99161f9e1049974","unresolved":false,"context_lines":[{"line_number":200,"context_line":"tags"},{"line_number":201,"context_line":"fixed_ips"},{"line_number":202,"context_line":"floating_ips"},{"line_number":203,"context_line":"dns_domains"},{"line_number":204,"context_line":""},{"line_number":205,"context_line":"Cell-level Tables"},{"line_number":206,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_ad153b9c","line":203,"updated":"2016-03-02 11:51:39.000000000","message":"so currently, nova-network wants all IPs restricted to each cell, but I think we might want to just drop nova-network support for that kind of setup, and make it a simple global-ish service.\n\nSo I like this move.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"b671c625f7b54e66b8b1a88c68a96514e63d82a1","unresolved":false,"context_lines":[{"line_number":200,"context_line":"tags"},{"line_number":201,"context_line":"fixed_ips"},{"line_number":202,"context_line":"floating_ips"},{"line_number":203,"context_line":"dns_domains"},{"line_number":204,"context_line":""},{"line_number":205,"context_line":"Cell-level Tables"},{"line_number":206,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7af24918_521ad917","line":203,"in_reply_to":"7af24918_ad153b9c","updated":"2016-03-03 21:52:52.000000000","message":"Fixed ip has a FK link to the instance table. We will need to either just kill the FK altogether or make it link to the instance_mapping table in the API db.","commit_id":"346dfdcdfac9cd51601e99ea165c4a60e6f9b6eb"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"fef9cef99700b031f07a075fbec3e101f0a43074","unresolved":false,"context_lines":[{"line_number":201,"context_line":"Cell-level Tables"},{"line_number":202,"context_line":"-----------------"},{"line_number":203,"context_line":""},{"line_number":204,"context_line":"aggregate_hosts"},{"line_number":205,"context_line":"bw_usage_cache"},{"line_number":206,"context_line":"compute_nodes"},{"line_number":207,"context_line":"consoles"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a122d0e_efb91817","line":204,"range":{"start_line":204,"start_character":0,"end_line":204,"end_character":15},"updated":"2016-05-03 01:59:07.000000000","message":"Is this going to the api db? In the compute rpc api it isn\u0027t related to an instance","commit_id":"d0614061cefcb7e0d1dd4e38630233e2d1280038"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"e0310288357f4c76d250a08e56ff7edafddf0ba0","unresolved":false,"context_lines":[{"line_number":201,"context_line":"Cell-level Tables"},{"line_number":202,"context_line":"-----------------"},{"line_number":203,"context_line":""},{"line_number":204,"context_line":"aggregate_hosts"},{"line_number":205,"context_line":"bw_usage_cache"},{"line_number":206,"context_line":"compute_nodes"},{"line_number":207,"context_line":"consoles"}],"source_content_type":"text/x-rst","patch_set":4,"id":"dada55a8_058b61c4","line":204,"range":{"start_line":204,"start_character":0,"end_line":204,"end_character":15},"in_reply_to":"1a122d0e_efb91817","updated":"2016-07-19 20:07:36.000000000","message":"Yes, this is going to the API db.","commit_id":"d0614061cefcb7e0d1dd4e38630233e2d1280038"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"fef9cef99700b031f07a075fbec3e101f0a43074","unresolved":false,"context_lines":[{"line_number":205,"context_line":"bw_usage_cache"},{"line_number":206,"context_line":"compute_nodes"},{"line_number":207,"context_line":"consoles"},{"line_number":208,"context_line":"console_pools"},{"line_number":209,"context_line":"instances"},{"line_number":210,"context_line":"instance_extra"},{"line_number":211,"context_line":"instance_info_caches"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a122d0e_4f35aca6","line":208,"range":{"start_line":208,"start_character":0,"end_line":208,"end_character":13},"updated":"2016-05-03 01:59:07.000000000","message":"Is this going to the api db? In the compute rpc api it isn\u0027t related to an instance","commit_id":"d0614061cefcb7e0d1dd4e38630233e2d1280038"},{"author":{"_account_id":18247,"name":"Mark Doffman","email":"mdoffman@linux.vnet.ibm.com","username":"doffm"},"change_message_id":"e0310288357f4c76d250a08e56ff7edafddf0ba0","unresolved":false,"context_lines":[{"line_number":205,"context_line":"bw_usage_cache"},{"line_number":206,"context_line":"compute_nodes"},{"line_number":207,"context_line":"consoles"},{"line_number":208,"context_line":"console_pools"},{"line_number":209,"context_line":"instances"},{"line_number":210,"context_line":"instance_extra"},{"line_number":211,"context_line":"instance_info_caches"}],"source_content_type":"text/x-rst","patch_set":4,"id":"dada55a8_0117d0ed","line":208,"range":{"start_line":208,"start_character":0,"end_line":208,"end_character":13},"in_reply_to":"1a122d0e_4f35aca6","updated":"2016-07-19 20:07:36.000000000","message":"I think these are staying in the cell. These specific tables are related to an instance or a compute node, so we should be able to find them via a lookup in the instance mapping.","commit_id":"d0614061cefcb7e0d1dd4e38630233e2d1280038"}]}