)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Sean Dague \u003csean@dague.net\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-05-18 15:48:51 -0400"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Declare plainly the current state of Posgresql in OpenStack"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This declares the current state of the world, including problems that"},{"line_number":10,"context_line":"are made more complicated by a perceived support of PostgreSQL. It has"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":7,"id":"ff0f0b1f_5789e6c7","line":7,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This declares the current state of the world, including problems that"},{"line_number":10,"context_line":"are made more complicated by a perceived support of PostgreSQL. It has"},{"line_number":11,"context_line":"concrete steps to move forward and prevent us from continuing to"},{"line_number":12,"context_line":"mislead our users."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"Change-Id: I332cef8ec4539520adcf37c6d2ea11488289fcfd"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":7,"id":"ff0f0b1f_d7be1664","line":12,"updated":"2017-05-19 08:23:50.000000000","message":"I understand the emotions behind it, but do we need to word it that way in a git commit message?","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"dbf6d136ed78eca3d17ce9c311fc92d6b9eeb60a","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Dirk Mueller \u003cdirk@dmllr.de\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-06-11 22:28:23 +0200"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Declare plainly the current state of MySQL in OpenStack"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This declares the current state of OpenStack\u0027s database support"},{"line_number":10,"context_line":"including the impact of a perceived support of PostgreSQL."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":10,"id":"7f231b9d_1cddeb41","line":7,"updated":"2017-06-12 15:37:02.000000000","message":"I thought the goal was to talk about PostgreSQL and _not_ talk about MySQL?","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Dirk Mueller \u003cdirk@dmllr.de\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-06-11 22:28:23 +0200"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Declare plainly the current state of MySQL in OpenStack"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This declares the current state of OpenStack\u0027s database support"},{"line_number":10,"context_line":"including the impact of a perceived support of PostgreSQL."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":10,"id":"7f231b9d_dec4fd61","line":7,"in_reply_to":"7f231b9d_1cddeb41","updated":"2017-06-12 21:47:12.000000000","message":"Good point.","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":11904,"name":"Sean McGinnis","email":"sean.mcginnis@gmail.com","username":"SeanM"},"change_message_id":"9eb854f000a1037c78bcb97e6f6c41516c80b5ba","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     Dirk Mueller \u003cdirk@dmllr.de\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-06-13 20:05:37 +0200"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Declare plainly the current state of MySQL in OpenStack"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This declares the current state of OpenStack\u0027s database support"},{"line_number":10,"context_line":"including the impact of a perceived support of PostgreSQL."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":11,"id":"7f231b9d_28b393ec","line":7,"range":{"start_line":7,"start_character":37,"end_line":7,"end_character":42},"updated":"2017-06-13 18:07:59.000000000","message":"Aren\u0027t we trying to plainly state of PostreSQL?","commit_id":"2cef1c3c6d113392a1e5dff775f13ee8e33e03da"},{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"5995659dc95f07165457c7ec95fb10e73e8e4a4c","unresolved":false,"context_lines":[{"line_number":12,"context_line":"It advises concrete steps to take in order to avoid a false"},{"line_number":13,"context_line":"impression of having a sustaining PostgreSQL support available."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"Re-worded by: Dirk Mueller \u003cdirk@dmllr.de\u003e"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"Change-Id: I332cef8ec4539520adcf37c6d2ea11488289fcfd"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":12,"id":"7f231b9d_4d4928ce","line":15,"updated":"2017-06-14 10:34:08.000000000","message":"I think it would be fair for you to do a \u0027Co-Authored-By\u0027 for this.","commit_id":"38d3df2567043b7fc26c5b87a3a7bfaa902075e2"}],"resolutions/20170201-deprecate-pg.rst":[{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"98685be30ee7d682694fa93a3b38c92eabea4fe9","unresolved":false,"context_lines":[{"line_number":2,"context_line":" 2017-02-01 Deprecation of Postgresql support in the community"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""},{"line_number":5,"context_line":"The If you look at the myriad of deployment options in OpenStack,"},{"line_number":6,"context_line":"there are many many ways that you can configure fundamental services"},{"line_number":7,"context_line":"for OpenStack. If you look at the `OpenStack User Survey`_, there are"},{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_870f8e2d","line":5,"updated":"2017-02-01 20:09:26.000000000","message":"The If","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"31bf021fd1d06097858e44431b8ebd6947f90f53","unresolved":false,"context_lines":[{"line_number":2,"context_line":" 2017-02-01 Deprecation of Postgresql support in the community"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""},{"line_number":5,"context_line":"The If you look at the myriad of deployment options in OpenStack,"},{"line_number":6,"context_line":"there are many many ways that you can configure fundamental services"},{"line_number":7,"context_line":"for OpenStack. If you look at the `OpenStack User Survey`_, there are"},{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_2625e3ad","line":5,"in_reply_to":"3a461143_870f8e2d","updated":"2017-02-02 13:00:05.000000000","message":"Done","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"72e5da452b6780adaee500ff9c1f1ada21a631af","unresolved":false,"context_lines":[{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem has largely settled on MySQL as the backend."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Common patterns help deployers because there are pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"},{"line_number":13,"context_line":"knowledge needed to actually run a cloud. The moment operators head in"},{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_0df2df06","line":11,"range":{"start_line":11,"start_character":49,"end_line":11,"end_character":53},"updated":"2017-02-01 21:16:13.000000000","message":"pools? or, \u0027there is a pool\u0027.","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"31bf021fd1d06097858e44431b8ebd6947f90f53","unresolved":false,"context_lines":[{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem has largely settled on MySQL as the backend."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Common patterns help deployers because there are pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"},{"line_number":13,"context_line":"knowledge needed to actually run a cloud. The moment operators head in"},{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_661feb5b","line":11,"range":{"start_line":11,"start_character":49,"end_line":11,"end_character":53},"in_reply_to":"3a461143_0df2df06","updated":"2017-02-02 13:00:05.000000000","message":"Done","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"72e5da452b6780adaee500ff9c1f1ada21a631af","unresolved":false,"context_lines":[{"line_number":11,"context_line":"Common patterns help deployers because there are pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"},{"line_number":13,"context_line":"knowledge needed to actually run a cloud. The moment operators head in"},{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"},{"line_number":15,"context_line":"tribal knowledge."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"There are also time where the backend database impacts the format of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_6dd9a383","line":14,"range":{"start_line":14,"start_character":37,"end_line":14,"end_character":42},"updated":"2017-02-01 21:16:13.000000000","message":"lose","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"31bf021fd1d06097858e44431b8ebd6947f90f53","unresolved":false,"context_lines":[{"line_number":11,"context_line":"Common patterns help deployers because there are pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"},{"line_number":13,"context_line":"knowledge needed to actually run a cloud. The moment operators head in"},{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"},{"line_number":15,"context_line":"tribal knowledge."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"There are also time where the backend database impacts the format of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_8624d7af","line":14,"range":{"start_line":14,"start_character":37,"end_line":14,"end_character":42},"in_reply_to":"3a461143_6dd9a383","updated":"2017-02-02 13:00:05.000000000","message":"Done","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"72e5da452b6780adaee500ff9c1f1ada21a631af","unresolved":false,"context_lines":[{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"},{"line_number":15,"context_line":"tribal knowledge."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"There are also time where the backend database impacts the format of"},{"line_number":18,"context_line":"data accepted from the user on APIs. While some projects like Nova now"},{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_cdce17b7","line":17,"range":{"start_line":17,"start_character":15,"end_line":17,"end_character":19},"updated":"2017-02-01 21:16:13.000000000","message":"times","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"31bf021fd1d06097858e44431b8ebd6947f90f53","unresolved":false,"context_lines":[{"line_number":14,"context_line":"a radically different direction they loose access to those pools of"},{"line_number":15,"context_line":"tribal knowledge."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"There are also time where the backend database impacts the format of"},{"line_number":18,"context_line":"data accepted from the user on APIs. While some projects like Nova now"},{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_a639d3da","line":17,"range":{"start_line":17,"start_character":15,"end_line":17,"end_character":19},"in_reply_to":"3a461143_cdce17b7","updated":"2017-02-02 13:00:05.000000000","message":"Done","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"98685be30ee7d682694fa93a3b38c92eabea4fe9","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_2756da06","line":22,"updated":"2017-02-01 20:09:26.000000000","message":"Is there any way to work the word \"bonghits\" in here?","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"2bbe1732f63875c714c0b67a981d289cddd68d33","unresolved":false,"context_lines":[{"line_number":30,"context_line":"backend. And that the recommendation is they remove it from their"},{"line_number":31,"context_line":"code, documentation, and testing at the earliest convenience."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":34,"context_line":".. _[1]: https://review.openstack.org/#/c/357444/"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_8d616fd6","line":33,"updated":"2017-02-01 21:12:24.000000000","message":"I was going to say to use the latest:\n\nhttps://www.openstack.org/assets/survey/October2016SurveyReport.pdf\n\nBut that just sends me here:\n\nhttps://www.openstack.org/analytics\n\nWhich doesn\u0027t have the backing database info anymore, at least from what I can tell.","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"31bf021fd1d06097858e44431b8ebd6947f90f53","unresolved":false,"context_lines":[{"line_number":30,"context_line":"backend. And that the recommendation is they remove it from their"},{"line_number":31,"context_line":"code, documentation, and testing at the earliest convenience."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":34,"context_line":".. _[1]: https://review.openstack.org/#/c/357444/"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_264c0376","line":33,"in_reply_to":"3a461143_8d616fd6","updated":"2017-02-02 13:00:05.000000000","message":"Yeh, this should be at least some documented data there.","commit_id":"df7abd4538e28cb9d42296e381231703ae08c88e"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"c6397f3a8b659b311bb2093161d83d6493874007","unresolved":false,"context_lines":[{"line_number":6,"context_line":"there are many many ways that you can configure fundamental services"},{"line_number":7,"context_line":"for OpenStack. If you look at the `OpenStack User Survey`_, there are"},{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem has largely settled on MySQL as the backend."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Common patterns help deployers because there is pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_22f63cf9","line":9,"range":{"start_line":9,"start_character":33,"end_line":9,"end_character":38},"updated":"2017-02-07 17:50:13.000000000","message":"Its tempting to say MySQL, Maria (both +/- galera) are the most popular choices. Although maybe that confuses the message?\n\nWhat makes me say that, is we seem to be re-working such that we remove the SELECT...FOR...UPDATE pattern so we are more multi-master cluster friendly.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"3a0b7f5833e4f5cfecdab2e9273e92f57d833d5b","unresolved":false,"context_lines":[{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem has largely settled on MySQL as the backend."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Common patterns help deployers because there is pool of knowledge"},{"line_number":12,"context_line":"about use at scale, about tuning, about HA practices. The tribal"},{"line_number":13,"context_line":"knowledge needed to actually run a cloud. The moment operators head in"},{"line_number":14,"context_line":"a radically different direction they lose access to those pools of"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_37fc243e","line":11,"range":{"start_line":11,"start_character":48,"end_line":11,"end_character":52},"updated":"2017-02-02 16:35:05.000000000","message":"nit: a pool","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":2472,"name":"Doug Hellmann","email":"dhellmann@redhat.com","username":"doug-hellmann"},"change_message_id":"2983557f773606ad8173adc7d1191cf7c20614b2","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_904e9694","line":22,"updated":"2017-02-06 15:14:07.000000000","message":"Is this a hypothetical problem or can you cite a real example?","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"25e597d1eaadc4e826840e74142f6d8bca77b79d","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_e727f0c7","line":22,"in_reply_to":"1a430d35_904e9694","updated":"2017-02-07 15:09:36.000000000","message":"A few things that come to mind:\n\n1. There are differences in how strict they are and how they handle types, e.g. you can see some pg specific conditions in the base nova db schema migration:\n\nhttps://github.com/openstack/nova/blob/46c778efafa3ff3c7bb99d74656bb2d2869ceb16/nova/db/sqlalchemy/migrate_repo/versions/216_havana.py\n\nNow, things like postgres not allowing duplicate indexes is a thing that mysql can enforce too, and actually has already started emitting warnings if you are creating duplicate indexes (you can see these warnings in the devstack logs of any gate run today).\n\nWhen I added DB2 support it was much of the same since DB2 is super strict about types, indexes, constraints and naming, so when nova supported DB2 we had conditions in the code for stuff like that (not a ton, but it\u0027s annoying enough to maintain and not something 99% of people working on nova know anything about since they don\u0027t work with DB2).\n\n2. I *think* this nova spec is related to how different DBs store things differently, in this case it\u0027s MySQL that\u0027s called out:\n\nhttps://specs.openstack.org/openstack/nova-specs/specs/newton/approved/lowercase-metadata-keys.html\n\nThat\u0027s essentially a problem where by default mysql stores strings as case insensitive so nova has bugs leaking out of the API for certain key/value tables like for aggregates and metadata. On postgres I don\u0027t think it\u0027s a problem because it doesn\u0027t have the case insensitivity issue, but I could be wrong.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"486a4e04619fb1fc674cc5bd3215d91e61685b1c","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_47d7e4ec","line":22,"in_reply_to":"1a430d35_904e9694","updated":"2017-02-07 14:59:33.000000000","message":"The class of issues that I\u0027ve typically found are things where mysql does a type conversion automatically from \"2\" to 2, where as postgresql will not. There are also issues about UTF8. Postgresql supports 4byte by default when enabled, mysql only 3byte. So characters that work with pg clouds will fail with mysql ones.\n\nThe thing is, it\u0027s a lot of rough edges. You can fix every one of those in python code pushing the validation layer closer to the user, but having to write that additional layer is something that takes time and resources and is part of why there is a tax by supporting pg.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"c6397f3a8b659b311bb2093161d83d6493874007","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_e23d1490","line":22,"in_reply_to":"1a430d35_e2fa7e74","updated":"2017-02-07 17:50:13.000000000","message":"I am torn also. I quite like the simplicity of the current resolution.\n\nMaybe just link to one example problem?","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":2472,"name":"Doug Hellmann","email":"dhellmann@redhat.com","username":"doug-hellmann"},"change_message_id":"d46a24fe3dc9cb5065771aa82b28b0fc7d811872","unresolved":false,"context_lines":[{"line_number":19,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":20,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":21,"context_line":"clouds actually have different forward facing APIs. Code that works on"},{"line_number":22,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_e2fa7e74","line":22,"in_reply_to":"1a430d35_e727f0c7","updated":"2017-02-07 15:17:31.000000000","message":"Thanks Sean and Matt, those are good details. It seems like it would be useful to have that all written down somewhere so we can point to it when someone asks why. I don\u0027t know if the resolution body is the best place for that. Maybe?","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"25e597d1eaadc4e826840e74142f6d8bca77b79d","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_a2a316ff","line":26,"range":{"start_line":26,"start_character":37,"end_line":26,"end_character":45},"updated":"2017-02-07 15:09:36.000000000","message":"FWIW, this was the mailing list thread that started the ball rolling on dropping the postgres job from the integrated gate:\n\nhttp://lists.openstack.org/pipermail/openstack-dev/2016-August/101892.html\n\nThat\u0027s probably worth linking in here as footnote [2].","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"125503b6c36f6da2ea5a33f20eb99c6db4599b79","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_b37d4a1e","line":26,"updated":"2017-02-06 22:08:48.000000000","message":"In neutron we tried to keep postgres support going for a while, flipping the job between voting and non-voting for quite sometime. Then we gave in entirely and had it experimental only. That said, we kept on getting bug reports from users using Postgres. Recently we chose to go periodic [1], so that we can keep an eye on what happens but without wasting too many resources and avoid RC firefight (caused by last minute bugs reported against this backend). I think it\u0027s wrong to make an executive decision to force postgres support out of our gate infrastructure. One thing is promoting the use of well establishing practices, another is banning niche ones without serious justification as to why it\u0027s detrimental to the progress of the wider community.\n\n[1] https://review.openstack.org/#/c/408847/","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":3153,"name":"Emilien Macchi","email":"emilien@redhat.com","username":"emilienm"},"change_message_id":"362c34f814d0a018e71d66c322e4ae88020fdffe","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_a2c876a9","line":26,"in_reply_to":"1a430d35_47f8c449","updated":"2017-02-07 15:10:12.000000000","message":"fyi, packstack doesn\u0027t support postgresql and we only support mysql, same thing for tripleo.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"c6397f3a8b659b311bb2093161d83d6493874007","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_e20b54d4","line":26,"in_reply_to":"1a430d35_a2c876a9","updated":"2017-02-07 17:50:13.000000000","message":"+1 for shouting loud so user expectations are clear.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"3cfd17127cee194c2fc4cb27b126e176dbd76150","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_47f8c449","line":26,"in_reply_to":"1a430d35_b37d4a1e","updated":"2017-02-07 15:03:37.000000000","message":"This is also about user expectations. Saying that projects will handle bug fixes when they have time, is vastly different than a statement of support.\n\nMany of the early complaints about this were about the fact that \"well packstack tells me how to do it with postgresql, so I assume it\u0027s as good as mysql.\" We in the community all know that\u0027s not the case. However, that\u0027s not messaged anywhere. People make product decisions based on that assumption, which means through omission we are giving people bad advice.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"ebdfb0ab919493cdf25c3b07a6a34f983cd621c0","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"There is very little value to our user or operator community to keep"},{"line_number":25,"context_line":"class A support for a second database backend. We stopped gating"},{"line_number":26,"context_line":"changes based on it over 5 months ago `[1]`_."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_05422a68","line":26,"in_reply_to":"1a430d35_e20b54d4","updated":"2017-02-07 18:35:25.000000000","message":"I appreciate the desire of making a more explicit statement about the state of DB support in OpenStack, and I agree that MySql has the lion share of what developers and deployers use. Personally I have always recommended adoption of gate-exercised technologies. \n\nHowever, one thing is to make a recommendation (and allowing best effort initiatives to proceed), the other is using a wording that effectively prevents those initiatives from proceeding. To some degree the same considerations could be made for distros where we wildly gate on Canonical/Ubuntu images. What about Red Hat Fedora/CentOS/ and SUSE OpenSuse/SLES? Are we going to deprecate those also in the future because the package fragmentation is too much to deal with? It\u0027s a slippery slope, IMO. My negative vote is mainly about the wording. Diversity is what makes this community strong and this resolution seems to send a conflicting signal against our principle of open collaboration.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"ebdfb0ab919493cdf25c3b07a6a34f983cd621c0","unresolved":false,"context_lines":[{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"},{"line_number":30,"context_line":"backend. And that the recommendation is they remove it from their"},{"line_number":31,"context_line":"code, documentation, and testing at the earliest convenience."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":34,"context_line":".. _[1]: https://review.openstack.org/#/c/357444/"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_e8889373","line":31,"range":{"start_line":30,"start_character":9,"end_line":31,"end_character":60},"updated":"2017-02-07 18:35:25.000000000","message":"I guess this is the part of the recommendation that bugs me the most (and the commit header \u0027deprecate postgresql in OpenStack\u0027). Recommending to focus on MySql is one thing, removing support for PostGres is another.","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"2600f2e5341e77fca8db92ac7744453cb40735a1","unresolved":false,"context_lines":[{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that Postgresql support is not required"},{"line_number":29,"context_line":"of any projects in OpenStack that have a relational database"},{"line_number":30,"context_line":"backend. And that the recommendation is they remove it from their"},{"line_number":31,"context_line":"code, documentation, and testing at the earliest convenience."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":34,"context_line":".. _[1]: https://review.openstack.org/#/c/357444/"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1a430d35_393e6bda","line":31,"range":{"start_line":30,"start_character":9,"end_line":31,"end_character":60},"in_reply_to":"1a430d35_e8889373","updated":"2017-02-07 20:30:30.000000000","message":"What about remove it from documentation immediately, and consider it deprecated in code and testing?","commit_id":"69ed94c00582e864b822b031b27f635d6555a94c"},{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"3b8c0666dbe8ad952aff414d1d79e96387274504","unresolved":false,"context_lines":[{"line_number":27,"context_line":""},{"line_number":28,"context_line":"It is time to explicitly state that:"},{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"}],"source_content_type":"text/x-rst","patch_set":3,"id":"1a430d35_e2d410ed","line":30,"updated":"2017-02-08 12:39:55.000000000","message":"one \u0027l\u0027 in Galera","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"16c1b7ac0b3e84eab92fae944b7a1e4e3745d06f","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_2ab377ea","line":32,"updated":"2017-02-23 16:16:21.000000000","message":"I agree we need to be extra careful as a community when using the word \"support\" anywhere in the future. I still feel that putting any instructions in official documentation without a strongly-worded warning may give the impression a solution is \"supported\" simply by virtue of it being documented at all, but in this case use of the word muddies the waters even more.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":7872,"name":"Marc Koderer","email":"marc@koderer.com","username":"mkoderer"},"change_message_id":"bd7e949137242c06f0000a1798fdbefff5321399","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_decad76d","line":32,"range":{"start_line":32,"start_character":35,"end_line":32,"end_character":56},"updated":"2017-02-22 12:59:15.000000000","message":"There is no automated way of migration from PostgreSQL to MySQL/Maria - this resolution will break the migration contract.. OpenStack users that are on PostgreSQL can\u0027t upgrade to a new release.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"9e4e8c9ec8d2c5a05dfd9dc570e61936fb978b08","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_9a7275e4","line":32,"updated":"2017-02-20 15:25:50.000000000","message":"What exactly does that mean? does that mean that projects can go with PyMySQL directly instead of using some middle layer like oslo.db or SQLAlchemy? or does that mean that they still are enouraged/required to use those layers but only test one particular backend?","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"aae9fa9b6c4b6e216b3e904b890fbb8306408242","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_0f588589","line":32,"range":{"start_line":32,"start_character":35,"end_line":32,"end_character":56},"in_reply_to":"da36d5c6_0a26c914","updated":"2017-02-23 16:05:17.000000000","message":"The Ocata installation guides (https://docs.openstack.org/ocata/install-guide-obs/environment-sql-database.html, https://docs.openstack.org/ocata/install-guide-rdo/environment-sql-database.html, https://docs.openstack.org/ocata/install-guide-ubuntu/environment-sql-database.html) all say \"OpenStack services also support other SQL databases including PostgreSQL.\"  So I think operators could be forgiven for thinking that PostgreSQL is supported.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"57ba6493f6d5be5791dcc5a566038d4f690bfa55","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_0a26c914","line":32,"range":{"start_line":32,"start_character":35,"end_line":32,"end_character":56},"in_reply_to":"da36d5c6_decad76d","updated":"2017-02-23 02:59:13.000000000","message":"This assumes there has been an upstream contract here, which I don\u0027t think is a valid assumption. That\u0027s part of the current issue, there is an assumed support statement for a thing which really doesn\u0027t have first class upstream support.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"9e4e8c9ec8d2c5a05dfd9dc570e61936fb978b08","unresolved":false,"context_lines":[{"line_number":30,"context_line":" - The MySQL database ecosystem (MySQL, Maria, Gallera) is the assumed"},{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_da5b0d53","line":33,"updated":"2017-02-20 15:25:50.000000000","message":"It would be nice to get at least the spelling right here - it is PostgreSQL","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"3beaacdc7d6875634926bf35a4a72d375ec56dbe","unresolved":false,"context_lines":[{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa31d9ce_2b39acf2","line":34,"updated":"2017-02-14 21:03:15.000000000","message":"Should official documentation also remove sections/references describing how to implement alternatives, or at least be expected to annotate them with warnings that this is not tested upstream and may perform poorly or not work at all so only the primary model is strongly recommended for production deployments?","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"e34f7bb865d9c4eda297cb3c1bd402c07574adbb","unresolved":false,"context_lines":[{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa31d9ce_d5834954","line":34,"updated":"2017-02-14 23:36:40.000000000","message":"The counter-concern raised is that by documenting how to do something we are making implicit assertions that it\u0027s supported and expected to work well (or even at all).","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f84b7c26658fdb91603a8d7957178b076dcd39b0","unresolved":false,"context_lines":[{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa31d9ce_5ab14846","line":34,"in_reply_to":"fa31d9ce_2b39acf2","updated":"2017-02-14 23:11:18.000000000","message":"i would prefer that we keep this to what we know as fact:\n  1. postgresql isn\u0027t tested (at least not across all projects).\n  2. postgresql isn\u0027t being considered in development (at least not across all projects).\n\ni don\u0027t think it\u0027s safe to make comments on performance or whether it works.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"4fa77d2c790894974e0fc62282e45ab58667d1e8","unresolved":false,"context_lines":[{"line_number":31,"context_line":"   default backend for projects needing a relational database."},{"line_number":32,"context_line":" - Projects do not need to support alternative databases (like"},{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa31d9ce_baa14848","line":34,"in_reply_to":"fa31d9ce_d5834954","updated":"2017-02-15 13:54:21.000000000","message":"that\u0027s cool. i\u0027m not suggesting this either. :)\n\nall i said was that we should provide the facts: not all projects test postgres and not all projects develop for postgres. unless we have benchmarks, i don\u0027t think we can objectively say one is better than the other. we\u0027ve mentioned how mysql stores strings as an issue, which is arguably something we should help zzzeek with, regardless i don\u0027t think is a performance swing either way.\n\ni just want to re-emphasis, i\u0027m not suggesting all projects develop for postgres and we gate on postgres. each project should do what they feel fit (as long as mysql is supported). there is obviously a strong opinion from select projects that postgres is too much of a burden. that\u0027s cool. in telemetry, we\u0027ve handled both, mysql sometimes needing tweaks. that\u0027s also cool.\n\nas mriedem mentioned, this arose from telemetry team asking if nova supported postgres, if not, telemetry team would drop gate which required nova+postgres. we just wanted to clarify that postgres is not considered by nova and i suggested this be raised since it conflicts with existing docs. i _like_ this patch as it. i just didn\u0027t like the prior revisions which said \"let\u0027s rip out existing postgres for some unknown reason.\"","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"9e4e8c9ec8d2c5a05dfd9dc570e61936fb978b08","unresolved":false,"context_lines":[{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""},{"line_number":38,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":39,"context_line":".. [1] MySQL is used as a stand in term in this document for MySQL"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_fae531a8","line":36,"updated":"2017-02-20 15:25:50.000000000","message":"Can we add that projects removing support for non-MySQL that there is a way to convert user data to MySQL then? I\u0027m fine with removing a non-default choice if there is some way to get the users converted without loosing all the data. \n\nAt this point I don\u0027t think its a good idea to let the individual downstream openstack distributors (not sure if there is more than SUSE that chose to go with PostgreSQL) develop the conversion path themselves.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"1e2c62f5c361dd47aa884e79268dd7d046dce3d4","unresolved":false,"context_lines":[{"line_number":33,"context_line":"   Postgresql)"},{"line_number":34,"context_line":" - OpenStack documentation should reference MySQL usage."},{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""},{"line_number":38,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":39,"context_line":".. [1] MySQL is used as a stand in term in this document for MySQL"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_bf171850","line":36,"in_reply_to":"da36d5c6_fae531a8","updated":"2017-02-21 02:55:56.000000000","message":"Gordon said that Huawei was supporting PG in their distro.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"9e4e8c9ec8d2c5a05dfd9dc570e61936fb978b08","unresolved":false,"context_lines":[{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""},{"line_number":38,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":39,"context_line":".. [1] MySQL is used as a stand in term in this document for MySQL"},{"line_number":40,"context_line":"       and its close variants, MariaDB, Gallera."},{"line_number":41,"context_line":".. [2] Common issues include the fact that MySQL char strings are case"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_5af39d73","line":38,"updated":"2017-02-20 15:25:50.000000000","message":"There is no more recent survey that has relevant numbers?","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"1e2c62f5c361dd47aa884e79268dd7d046dce3d4","unresolved":false,"context_lines":[{"line_number":35,"context_line":" - Projects are free to remove support for non MySQL usage when"},{"line_number":36,"context_line":"   appropriate for their project."},{"line_number":37,"context_line":""},{"line_number":38,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":39,"context_line":".. [1] MySQL is used as a stand in term in this document for MySQL"},{"line_number":40,"context_line":"       and its close variants, MariaDB, Gallera."},{"line_number":41,"context_line":".. [2] Common issues include the fact that MySQL char strings are case"}],"source_content_type":"text/x-rst","patch_set":3,"id":"da36d5c6_1f07441d","line":38,"in_reply_to":"da36d5c6_5af39d73","updated":"2017-02-21 02:55:56.000000000","message":"I couldn\u0027t find anything in the latest published survey from October. I don\u0027t know if the one being generated this week has anything in it about the DB backend (I failed to take the survey).","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"20bb1f887e999e4582a5ce17c66e69ba63dc9f89","unresolved":false,"context_lines":[{"line_number":37,"context_line":""},{"line_number":38,"context_line":".. _OpenStack User Survey: https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":39,"context_line":".. [1] MySQL is used as a stand in term in this document for MySQL"},{"line_number":40,"context_line":"       and its close variants, MariaDB, Gallera."},{"line_number":41,"context_line":".. [2] Common issues include the fact that MySQL char strings are case"},{"line_number":42,"context_line":"       insensitive by default. That storage defaults for utf8 are"},{"line_number":43,"context_line":"       different between the two."}],"source_content_type":"text/x-rst","patch_set":3,"id":"1a430d35_3a9b8b19","line":40,"updated":"2017-02-08 18:04:03.000000000","message":"Typo Galera","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"2b92703a491bb9e201f231aaaa202dd38522cc45","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"fa31d9ce_38dbc750","line":45,"updated":"2017-02-14 18:54:14.000000000","message":"Maybe we should include how people can step up to help reverse the deprecation and reinstate support before the support is fully dropped?\n\nThat should include educating the community on why its worth using a different database backend, and what the benefits are.","commit_id":"ee9337188ff658f524a35ee72ff92b3810f673df"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"13a7e924c1fa0aeedcb375d0d666d472b57895b8","unresolved":false,"context_lines":[{"line_number":15,"context_line":"a radically different direction they lose access to those pools of"},{"line_number":16,"context_line":"tribal knowledge."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"There are also times where the backend database impacts the format of"},{"line_number":19,"context_line":"data accepted from the user on APIs. While some projects like Nova now"},{"line_number":20,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":21,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":22,"context_line":"clouds actually have different forward facing APIs. [#f4]_ Code that works on"},{"line_number":23,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":24,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_90354fdd","line":21,"range":{"start_line":18,"start_character":0,"end_line":21,"end_character":4},"updated":"2017-05-16 09:21:40.000000000","message":"Is this a nice way of saying that using PostgreSQL as database backend, through its sane-by-default config, discovers bugs that developers using unsane-by-default MySQL lets slip through? I.e., that by deprecating PostgreSQL support, developers can continue to ignore these latent bugs, lowering data correctness level in OpenStack as a whole?","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"b032b4be5bcb8ad2ecdf7a99f62a298f0c784135","unresolved":false,"context_lines":[{"line_number":15,"context_line":"a radically different direction they lose access to those pools of"},{"line_number":16,"context_line":"tribal knowledge."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"There are also times where the backend database impacts the format of"},{"line_number":19,"context_line":"data accepted from the user on APIs. While some projects like Nova now"},{"line_number":20,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":21,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":22,"context_line":"clouds actually have different forward facing APIs. [#f4]_ Code that works on"},{"line_number":23,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":24,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_1f3388ad","line":21,"range":{"start_line":18,"start_character":0,"end_line":21,"end_character":4},"in_reply_to":"1f013ff3_90354fdd","updated":"2017-05-16 12:18:15.000000000","message":"One of the common issues I had to fix when first adding pg to the gate was that MySQL will do type coersion from int -\u003e string fine. PG will error under such conditions. So\n\n \"resource\": {\n    \"id\": 1\n }\n\nvs.\n\n \"resource\": {\n    \"id\": \"1\"\n }\n\nWould fail on PG backends (potentially very late in the process), but work fine with MySQL. Strict json and oslo.versioned object validation shifted that up so that we\u0027re doing this much earlier in the process in a project like Nova, but if you didn\u0027t do that to your internals you are subject to a lot of late fails that turn into server 500 issues.","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":5638,"name":"Davanum Srinivas","email":"davanum@gmail.com","username":"dims-v"},"change_message_id":"15fae61b248fb548206a3f8d69bb8670c3e1e26d","unresolved":false,"context_lines":[{"line_number":27,"context_line":"community. That statement of support makes people think the upstream"},{"line_number":28,"context_line":"development teams are equally thinking about Postgresql and MySQL when"},{"line_number":29,"context_line":"reviewing impacts to the database. The reality is, they are"},{"line_number":30,"context_line":"not. Upstream is largely MySQL focused."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"The current known users of Postgresql are"},{"line_number":33,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_842e31a3","line":30,"updated":"2017-05-16 12:01:16.000000000","message":"\"As a result, most of the continuous integration testing in upstream is based on MySQL and NOT Postgresql\"","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"b032b4be5bcb8ad2ecdf7a99f62a298f0c784135","unresolved":false,"context_lines":[{"line_number":27,"context_line":"community. That statement of support makes people think the upstream"},{"line_number":28,"context_line":"development teams are equally thinking about Postgresql and MySQL when"},{"line_number":29,"context_line":"reviewing impacts to the database. The reality is, they are"},{"line_number":30,"context_line":"not. Upstream is largely MySQL focused."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"The current known users of Postgresql are"},{"line_number":33,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_ff27bc63","line":30,"in_reply_to":"1f013ff3_842e31a3","updated":"2017-05-16 12:18:15.000000000","message":"Done","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":6547,"name":"Andreas Jaeger","email":"jaegerandi@gmail.com","username":"jaegerandi"},"change_message_id":"3aef4486aa9d42c451e106b79f0c619307edca5c","unresolved":false,"context_lines":[{"line_number":29,"context_line":"reviewing impacts to the database. The reality is, they are"},{"line_number":30,"context_line":"not. Upstream is largely MySQL focused."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"The current known users of Postgresql are"},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"We should be clear to future operators about this situation through"},{"line_number":35,"context_line":"the following actions:"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_0affecaa","line":32,"range":{"start_line":32,"start_character":0,"end_line":32,"end_character":41},"updated":"2017-05-15 11:06:51.000000000","message":"That sentence is broken.","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"b032b4be5bcb8ad2ecdf7a99f62a298f0c784135","unresolved":false,"context_lines":[{"line_number":29,"context_line":"reviewing impacts to the database. The reality is, they are"},{"line_number":30,"context_line":"not. Upstream is largely MySQL focused."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"The current known users of Postgresql are"},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"We should be clear to future operators about this situation through"},{"line_number":35,"context_line":"the following actions:"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_bf213450","line":32,"range":{"start_line":32,"start_character":0,"end_line":32,"end_character":41},"in_reply_to":"1f013ff3_0446a17d","updated":"2017-05-16 12:18:15.000000000","message":"Done","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":5638,"name":"Davanum Srinivas","email":"davanum@gmail.com","username":"dims-v"},"change_message_id":"15fae61b248fb548206a3f8d69bb8670c3e1e26d","unresolved":false,"context_lines":[{"line_number":29,"context_line":"reviewing impacts to the database. The reality is, they are"},{"line_number":30,"context_line":"not. Upstream is largely MySQL focused."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"The current known users of Postgresql are"},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"We should be clear to future operators about this situation through"},{"line_number":35,"context_line":"the following actions:"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_0446a17d","line":32,"range":{"start_line":32,"start_character":0,"end_line":32,"end_character":41},"in_reply_to":"1f013ff3_0affecaa","updated":"2017-05-16 12:01:16.000000000","message":"Let\u0027s drop this listing","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"13a7e924c1fa0aeedcb375d0d666d472b57895b8","unresolved":false,"context_lines":[{"line_number":38,"context_line":""},{"line_number":39,"context_line":"    .. warning::"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":42,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":43,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":44,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- Investigate the viability of the migration path from Postgresql to"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_d0407738","line":43,"range":{"start_line":41,"start_character":7,"end_line":43,"end_character":26},"updated":"2017-05-16 09:21:40.000000000","message":"There\u0027s at least one counter-example to this in Gnocchi which for its indexer prefers PostgreSQL: https://docs.openstack.org/developer/gnocchi/architecture.html#available-index-back-ends","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"b032b4be5bcb8ad2ecdf7a99f62a298f0c784135","unresolved":false,"context_lines":[{"line_number":38,"context_line":""},{"line_number":39,"context_line":"    .. warning::"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":42,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":43,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":44,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"- Investigate the viability of the migration path from Postgresql to"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_bfc5b490","line":43,"range":{"start_line":41,"start_character":7,"end_line":43,"end_character":26},"in_reply_to":"1f013ff3_d0407738","updated":"2017-05-16 12:18:15.000000000","message":"gnocchi actually delisted itself as part of OpenStack 7 weeks ago - https://review.openstack.org/#/c/447438 - it\u0027s been largely developed independently (does not use the OpenStack release cadence) and is going to be more so in the future.","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"13a7e924c1fa0aeedcb375d0d666d472b57895b8","unresolved":false,"context_lines":[{"line_number":55,"context_line":".. [#f1] April 2016 User Survey"},{"line_number":56,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":57,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":58,"context_line":"         and its close variants, MariaDB, Gallera."},{"line_number":59,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of Postgresql"},{"line_number":60,"context_line":"         are:"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_5076a7a6","line":58,"range":{"start_line":58,"start_character":42,"end_line":58,"end_character":49},"updated":"2017-05-16 09:21:40.000000000","message":"(Nit): Galera","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"13a7e924c1fa0aeedcb375d0d666d472b57895b8","unresolved":false,"context_lines":[{"line_number":56,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":57,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":58,"context_line":"         and its close variants, MariaDB, Gallera."},{"line_number":59,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of Postgresql"},{"line_number":60,"context_line":"         are:"},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"         - Suse - (investigating migrating to Gallera)"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_f074939b","line":59,"range":{"start_line":59,"start_character":58,"end_line":59,"end_character":68},"updated":"2017-05-16 09:21:40.000000000","message":"(Nit): PostgreSQL","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":5293,"name":"Nikhil","email":"SlickNik@gmail.com","username":"slicknik"},"change_message_id":"d5a7488f1a988410f05d53e3bfb6b5c523efd9f0","unresolved":false,"context_lines":[{"line_number":59,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of Postgresql"},{"line_number":60,"context_line":"         are:"},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"         - Suse - (investigating migrating to Gallera)"},{"line_number":63,"context_line":"         - Huawei - not actually postgresql, but an internal db based"},{"line_number":64,"context_line":"           on pg."},{"line_number":65,"context_line":"         - Windriver"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1f013ff3_4b9e0c64","line":62,"updated":"2017-05-15 19:04:45.000000000","message":"(nitpick) It\u0027s \u0027Galera\u0027, and not \u0027Gallera\u0027","commit_id":"fd8cfc1456eec646239037945aa4e87f7767cbdd"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"ab4688bdb385b4c13c8aca1ab70c1e8fcd499db4","unresolved":false,"context_lines":[{"line_number":19,"context_line":"data accepted from the user on APIs. While some projects like Nova now"},{"line_number":20,"context_line":"have strict input validation on inbound requests, most projects do"},{"line_number":21,"context_line":"not. The choice of backends means that MySQL and Postgresql based"},{"line_number":22,"context_line":"clouds actually have different forward facing APIs. [#f4]_ Code that works on"},{"line_number":23,"context_line":"a MySQL backend db, may not actually work on a Postgresql one."},{"line_number":24,"context_line":""},{"line_number":25,"context_line":"We have however caused confusion in our operator community by"},{"line_number":26,"context_line":"sometimes stating that Postgresql was supported by the upstream"}],"source_content_type":"text/x-rst","patch_set":5,"id":"1f013ff3_195a7129","line":23,"range":{"start_line":22,"start_character":59,"end_line":23,"end_character":62},"updated":"2017-05-17 01:08:31.000000000","message":"[citation needed]?","commit_id":"0eb73983795f5988a67311ae2ab704b162022826"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"9ea2d47a9641d733ac1c90e8e4d96d9f40956a45","unresolved":false,"context_lines":[{"line_number":38,"context_line":"    .. warning::"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":41,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":42,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":43,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":44,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"1f013ff3_7f872211","line":41,"range":{"start_line":41,"start_character":14,"end_line":41,"end_character":49},"updated":"2017-05-16 16:43:48.000000000","message":"Just FYI, it looks like we\u0027re running the pg-full job in the experimental queue via the integrated gate jobs (which runs on all core projects). I dug into one of the test failures and it looks like it has nothing to do with postgresql (schema restrictions or anything), and everything to do with timing differences which actually exposed a race bug in the neutron subnet/network delete flow:\n\nhttps://bugs.launchpad.net/neutron/+bug/1691176\n\nSo this is a data point that we\u0027re actually getting improved test coverage by having different backends which different timing characteristics.","commit_id":"0eb73983795f5988a67311ae2ab704b162022826"},{"author":{"_account_id":11526,"name":"Martin Millnert","email":"martin@millnert.se","username":"Anticimex"},"change_message_id":"972c39f0683f2de429809197730aeca48b4051d9","unresolved":false,"context_lines":[{"line_number":38,"context_line":"    .. warning::"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":41,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":42,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":43,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":44,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"1f013ff3_b9913d9c","line":41,"range":{"start_line":41,"start_character":14,"end_line":41,"end_character":49},"in_reply_to":"1f013ff3_7f872211","updated":"2017-05-17 01:20:04.000000000","message":"This has consistently been the pattern on every postgresql-related bug that I have ever seen from 2013 (i.e not just timing characteristics): pg has provided increased coverage, mostly due to poor MySQL configs. Happily seeing any counter example. Hence my worry about reducing test coverage unless MySQL configuration in CI is improved w.r.t sanity \u0026 correctness.","commit_id":"0eb73983795f5988a67311ae2ab704b162022826"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"cc3b6a5640d30a5d729450991cb7efcf6d5c707e","unresolved":false,"context_lines":[{"line_number":45,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":46,"context_line":"  MySQL (Suse team is currently doing this, will report back)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"- Communicate to the OpenStack Foundation that there needs to be a way"},{"line_number":49,"context_line":"  for distributors to report back aggregate data about their users,"},{"line_number":50,"context_line":"  who are not self reporting to the survey. Our ability to make"},{"line_number":51,"context_line":"  decisions off the survey is impacted by large bits of missing"},{"line_number":52,"context_line":"  information that distributors have access to."},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"Depending on the outcome of the migration investigation we will assess"},{"line_number":55,"context_line":"what other actions might be appropriate to signal this lack of"}],"source_content_type":"text/x-rst","patch_set":5,"id":"1f013ff3_c28f8421","line":52,"range":{"start_line":48,"start_character":0,"end_line":52,"end_character":47},"updated":"2017-05-16 14:59:11.000000000","message":"Another thing that came up in the Forum session was actually getting real cost estimates from postgresql users (SAP?) about how long and what the cost would be to migrate to mysql in production, i.e. does it take a year and cost a billion USD to do this?","commit_id":"0eb73983795f5988a67311ae2ab704b162022826"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"b423244a097300f18d2fbaf135d9ea97d02e4031","unresolved":false,"context_lines":[{"line_number":45,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":46,"context_line":"  MySQL (Suse team is currently doing this, will report back)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"- Communicate to the OpenStack Foundation that there needs to be a way"},{"line_number":49,"context_line":"  for distributors to report back aggregate data about their users,"},{"line_number":50,"context_line":"  who are not self reporting to the survey. Our ability to make"},{"line_number":51,"context_line":"  decisions off the survey is impacted by large bits of missing"},{"line_number":52,"context_line":"  information that distributors have access to."},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"Depending on the outcome of the migration investigation we will assess"},{"line_number":55,"context_line":"what other actions might be appropriate to signal this lack of"}],"source_content_type":"text/x-rst","patch_set":5,"id":"1f013ff3_fc722049","line":52,"range":{"start_line":48,"start_character":0,"end_line":52,"end_character":47},"in_reply_to":"1f013ff3_c28f8421","updated":"2017-05-16 16:14:00.000000000","message":"In my head thats part of the migration path investigation, but yes that cost needs considering.","commit_id":"0eb73983795f5988a67311ae2ab704b162022826"}],"resolutions/20170201-postgresql-mysql-status.rst":[{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":7,"context_line":"User Survey 2017 [#f1]_ there is a trend visible that most production"},{"line_number":8,"context_line":"deployments today settle on MySQL as the configured database backend."},{"line_number":9,"context_line":""},{"line_number":10,"context_line":"Although the OpenStack User Survey is non-authoritative, it\u0027s reported"},{"line_number":11,"context_line":"bias (\u003e 10 to 1 for MySQL [#f2]_ vs. PostgreSQL [#f3]_) is also"},{"line_number":12,"context_line":"mirrored in openstack-operator mailing list posts, ask.openstack.org"},{"line_number":13,"context_line":"posts, and content during operator meetups. In contrast, production"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_c28d6f7c","line":10,"updated":"2017-06-10 12:30:03.000000000","message":"its?","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":7,"context_line":"User Survey 2017 [#f1]_ there is a trend visible that most production"},{"line_number":8,"context_line":"deployments today settle on MySQL as the configured database backend."},{"line_number":9,"context_line":""},{"line_number":10,"context_line":"Although the OpenStack User Survey is non-authoritative, it\u0027s reported"},{"line_number":11,"context_line":"bias (\u003e 10 to 1 for MySQL [#f2]_ vs. PostgreSQL [#f3]_) is also"},{"line_number":12,"context_line":"mirrored in openstack-operator mailing list posts, ask.openstack.org"},{"line_number":13,"context_line":"posts, and content during operator meetups. In contrast, production"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_4e5ce24a","line":10,"in_reply_to":"7f231b9d_c28d6f7c","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":13,"context_line":"posts, and content during operator meetups. In contrast, production"},{"line_number":14,"context_line":"deployments using PostgreSQL seem to be less popular ([#f4]_)."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":"The OpenStack community is generally offering wide ranges of options"},{"line_number":17,"context_line":"to the users and advocates developers to use proper DBMS abstraction"},{"line_number":18,"context_line":"layers like for example oslo.db and SQLAlchemy to avoid a lock-in on"},{"line_number":19,"context_line":"a specific database implementation. Still there are projects and"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_42f71ff7","line":16,"updated":"2017-06-10 12:30:03.000000000","message":"this sentence is long and could benefit from being decomposed into smaller ones.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":13,"context_line":"posts, and content during operator meetups. In contrast, production"},{"line_number":14,"context_line":"deployments using PostgreSQL seem to be less popular ([#f4]_)."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":"The OpenStack community is generally offering wide ranges of options"},{"line_number":17,"context_line":"to the users and advocates developers to use proper DBMS abstraction"},{"line_number":18,"context_line":"layers like for example oslo.db and SQLAlchemy to avoid a lock-in on"},{"line_number":19,"context_line":"a specific database implementation. Still there are projects and"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_6e5f9e41","line":16,"in_reply_to":"7f231b9d_42f71ff7","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":20,"context_line":"individual features that are being designed and developed with"},{"line_number":21,"context_line":"specific backends in mind."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Even when portability issues are catched during code review and various"},{"line_number":24,"context_line":"levels of automated QA, there is still impact on the community like"},{"line_number":25,"context_line":"documenting and circumventing peculiarities of each DBMS, or"},{"line_number":26,"context_line":"a significant demand for knowledgeable operators and developers"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_02d92782","line":23,"updated":"2017-06-10 12:30:03.000000000","message":"caught","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":20,"context_line":"individual features that are being designed and developed with"},{"line_number":21,"context_line":"specific backends in mind."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Even when portability issues are catched during code review and various"},{"line_number":24,"context_line":"levels of automated QA, there is still impact on the community like"},{"line_number":25,"context_line":"documenting and circumventing peculiarities of each DBMS, or"},{"line_number":26,"context_line":"a significant demand for knowledgeable operators and developers"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_8e4afa7d","line":23,"in_reply_to":"7f231b9d_02d92782","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":23,"context_line":"Even when portability issues are catched during code review and various"},{"line_number":24,"context_line":"levels of automated QA, there is still impact on the community like"},{"line_number":25,"context_line":"documenting and circumventing peculiarities of each DBMS, or"},{"line_number":26,"context_line":"a significant demand for knowledgeable operators and developers"},{"line_number":27,"context_line":"fielding questions of performance tuning, reference architectures"},{"line_number":28,"context_line":"and general discovery and sharing of best practices for each"},{"line_number":29,"context_line":"implementation choice."}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_e2e50b3a","line":26,"updated":"2017-06-10 12:30:03.000000000","message":"again, this sentence is long and difficult to follow.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":23,"context_line":"Even when portability issues are catched during code review and various"},{"line_number":24,"context_line":"levels of automated QA, there is still impact on the community like"},{"line_number":25,"context_line":"documenting and circumventing peculiarities of each DBMS, or"},{"line_number":26,"context_line":"a significant demand for knowledgeable operators and developers"},{"line_number":27,"context_line":"fielding questions of performance tuning, reference architectures"},{"line_number":28,"context_line":"and general discovery and sharing of best practices for each"},{"line_number":29,"context_line":"implementation choice."}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_ae45b672","line":26,"in_reply_to":"7f231b9d_e2e50b3a","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":32,"context_line":"includes contributors willing to step up, and fluent in either the"},{"line_number":33,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":34,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":35,"context_line":"MySQL, we potentially reduce the chances of them happening. Examples"},{"line_number":36,"context_line":"of potential impact are (but are not limited to):"},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"- OpenStack services taking a more active role in managing the DBMS"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_42d03f51","line":35,"updated":"2017-06-10 12:30:03.000000000","message":"after reading this paragraph, it is unclear what the question about future work is. The second sentence starts to talk of cost but doesn\u0027t make it clear what that cost is.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":32,"context_line":"includes contributors willing to step up, and fluent in either the"},{"line_number":33,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":34,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":35,"context_line":"MySQL, we potentially reduce the chances of them happening. Examples"},{"line_number":36,"context_line":"of potential impact are (but are not limited to):"},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"- OpenStack services taking a more active role in managing the DBMS"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_ce50f235","line":35,"in_reply_to":"7f231b9d_42d03f51","updated":"2017-06-11 20:28:14.000000000","message":"Removed the introductionary question, it confused me as well.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":39,"context_line":"- The ability to have zero down time upgrade for services such as"},{"line_number":40,"context_line":"  Keystone and Glance."},{"line_number":41,"context_line":"- Consistent UTF8 4 byte support in our APIs."},{"line_number":42,"context_line":"- Consistency around case sensitivity collation defaults that lead to"},{"line_number":43,"context_line":"  strange bugs around searching/updating names in resources ([#f5]_)."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"Therefore, we should clearly state that currently in the OpenStack"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_22c5438f","line":42,"updated":"2017-06-10 12:30:03.000000000","message":"some commas would help.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":39,"context_line":"- The ability to have zero down time upgrade for services such as"},{"line_number":40,"context_line":"  Keystone and Glance."},{"line_number":41,"context_line":"- Consistent UTF8 4 byte support in our APIs."},{"line_number":42,"context_line":"- Consistency around case sensitivity collation defaults that lead to"},{"line_number":43,"context_line":"  strange bugs around searching/updating names in resources ([#f5]_)."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"Therefore, we should clearly state that currently in the OpenStack"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_ee4bae84","line":42,"in_reply_to":"7f231b9d_22c5438f","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  strange bugs around searching/updating names in resources ([#f5]_)."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"Therefore, we should clearly state that currently in the OpenStack"},{"line_number":46,"context_line":"community there is a bias on MySQL [#f2]_ over than for other"},{"line_number":47,"context_line":"databases like for example PostgreSQL."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"As a set of concrete steps we should do the following:"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_02c24786","line":46,"range":{"start_line":46,"start_character":26,"end_line":46,"end_character":28},"updated":"2017-06-10 12:30:03.000000000","message":"bias towards; not on. \u0027over than\u0027 is not valid. just over, maybe?","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":43,"context_line":"  strange bugs around searching/updating names in resources ([#f5]_)."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"Therefore, we should clearly state that currently in the OpenStack"},{"line_number":46,"context_line":"community there is a bias on MySQL [#f2]_ over than for other"},{"line_number":47,"context_line":"databases like for example PostgreSQL."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"As a set of concrete steps we should do the following:"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_0e7dcab0","line":46,"in_reply_to":"7f231b9d_02c24786","updated":"2017-06-11 20:28:14.000000000","message":"Done","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":9664,"name":"amrith","email":"amrith.kumar@gmail.com","username":"amrith"},"change_message_id":"360a6a612695290d9eb2a90ebab13b3fe90d5840","unresolved":false,"context_lines":[{"line_number":49,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- The OpenStack documentation removes any references to PostgreSQL"},{"line_number":52,"context_line":"  being first-class supported by the upstream developer community."},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"- The OpenStack documentation should be updated to document the"},{"line_number":55,"context_line":"  community preference::"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_c2bb2f0c","line":52,"updated":"2017-06-10 12:30:03.000000000","message":"first class what?\n\nthere is also a clash in tense between these bullets ..\n\nThis one is \"removes\", the next is \"should be\", the third is \"should\", and the fourth is \"State\".","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"5ba9891d41f57451899b6c9ed10c56bda79db42e","unresolved":false,"context_lines":[{"line_number":49,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- The OpenStack documentation removes any references to PostgreSQL"},{"line_number":52,"context_line":"  being first-class supported by the upstream developer community."},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"- The OpenStack documentation should be updated to document the"},{"line_number":55,"context_line":"  community preference::"}],"source_content_type":"text/x-rst","patch_set":9,"id":"7f231b9d_2e80868b","line":52,"in_reply_to":"7f231b9d_c2bb2f0c","updated":"2017-06-11 20:28:14.000000000","message":"I\u0027ll remove this again as it wasn\u0027t in Sean\u0027s original proposal. I am not sure what the intent of leaving that out was, feels a bit like a toothless tiger then, but anyway.","commit_id":"595645be31892554cda41437f484d4b5d63cf908"},{"author":{"_account_id":308,"name":"Thierry Carrez","email":"thierry@openstack.org","username":"ttx"},"change_message_id":"2aa871204145b6447ca8930d9f8ab070d68de7e1","unresolved":false,"context_lines":[{"line_number":1,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":2,"context_line":" 2017-02-01 Document current level of community database support"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""},{"line_number":5,"context_line":"Over the years, OpenStack\u0027s ecosystem and developer community grew"}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_eecba3a2","line":2,"range":{"start_line":2,"start_character":1,"end_line":2,"end_character":11},"updated":"2017-06-12 08:55:44.000000000","message":"If you do another patchset, update the date to ~today","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":29,"context_line":""},{"line_number":30,"context_line":"There is also an impact on future work. All abstractions have a cost,"},{"line_number":31,"context_line":"and includes contributors willing to step up, and fluent in either the"},{"line_number":32,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":33,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":34,"context_line":"MySQL, we potentially reduce the chances of them happening. Examples"},{"line_number":35,"context_line":"of potential impact are (but are not limited to):"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"- OpenStack services taking a more active role in managing the DBMS"}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_7e0e098e","line":34,"range":{"start_line":32,"start_character":50,"end_line":34,"end_character":59},"updated":"2017-06-12 21:47:12.000000000","message":"I\u0027m not sure why this is still in here. During the TC meeting I thought we said this isn\u0027t being decided here. For example, it could be possible for the community to decide that zero-downtime upgrades with Keystone is a thing you only get with MySQL unless/until someone steps up and makes it a reality for PostgreSQL too - but that doesn\u0027t mean we can\u0027t move forward with the advanced feature with MySQL and if you\u0027re using PG, then you don\u0027t get zero-downtime upgrades (as has been the case for years already). Essentially making PG second-class, and we make sure that\u0027s documented and understood by peopling weighing the costs of picking a DB backend. As I\u0027ve noted several times before, the exact same is true based on which compute drivers you\u0027re running in Nova; there are a lot of things that will work with libvirt, which is the reference implementation we test against, but if you\u0027re picking vCenter or Ironic you can still do that, but the API isn\u0027t as feature complete.","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":43,"context_line":""},{"line_number":44,"context_line":"Therefore, we should clearly state that currently in the OpenStack"},{"line_number":45,"context_line":"community there is a bias towards MySQL [#f2]_ over other"},{"line_number":46,"context_line":"databases like for example PostgreSQL."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_3e56515e","line":46,"range":{"start_line":46,"start_character":10,"end_line":46,"end_character":37},"updated":"2017-06-12 21:47:12.000000000","message":"nit: this seems redundant as this doc is already using pg in earlier examples and arguments for only supporting mysql, but you don\u0027t need to change it, just seemed odd to me.","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"b4f73669a433a4c764d292bd4f11163b04b45252","unresolved":false,"context_lines":[{"line_number":50,"context_line":"- The OpenStack documentation should be updated to document the"},{"line_number":51,"context_line":"  community preference::"},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"    .. warning::"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"       While OpenStack can be used with many SQL DBMS, please"},{"line_number":56,"context_line":"       be aware that the upstream OpenStack developer community"}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_6e2b05eb","line":53,"updated":"2017-06-12 14:48:39.000000000","message":"do we know where we\u0027re adding this disclaimer?","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"- We should embrace contributions that enable a community supported"},{"line_number":67,"context_line":"  seamless migration path for deployments from PostgreSQL to"},{"line_number":68,"context_line":"  MySQL to mitigate a potentially upcoming OpenStack upgrade trap."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"- We should state to the OpenStack Foundation that there needs"},{"line_number":71,"context_line":"  to be a way for distributors to report back aggregate data about"}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_fee459ae","line":68,"range":{"start_line":68,"start_character":22,"end_line":68,"end_character":65},"updated":"2017-06-12 21:47:12.000000000","message":"Is this regarding the keystone triggers thing? Will that make all upgrades fail if not using mysql? Or only online zero-downtime upgrades?","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":11904,"name":"Sean McGinnis","email":"sean.mcginnis@gmail.com","username":"SeanM"},"change_message_id":"9abd47dc49741f6e69022d9b4c1150e4edc3790f","unresolved":false,"context_lines":[{"line_number":81,"context_line":"class supported upstream, and has never been and there is so far a"},{"line_number":82,"context_line":"definitive lack of consolidated effort to change that."},{"line_number":83,"context_line":""},{"line_number":84,"context_line":"This should not exclude vendors claim support for other DBMS choices"},{"line_number":85,"context_line":"for their customers. However, for users  bypassing vendors and"},{"line_number":86,"context_line":"going straight to upstream, the current reality of OpenStack"},{"line_number":87,"context_line":"enablement needs to be clear."}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_21e65746","line":84,"updated":"2017-06-11 21:41:31.000000000","message":"claiming - if you need to do another update.","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":82,"context_line":"definitive lack of consolidated effort to change that."},{"line_number":83,"context_line":""},{"line_number":84,"context_line":"This should not exclude vendors claim support for other DBMS choices"},{"line_number":85,"context_line":"for their customers. However, for users  bypassing vendors and"},{"line_number":86,"context_line":"going straight to upstream, the current reality of OpenStack"},{"line_number":87,"context_line":"enablement needs to be clear."},{"line_number":88,"context_line":""}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_7eea8975","line":85,"range":{"start_line":85,"start_character":39,"end_line":85,"end_character":41},"updated":"2017-06-12 21:47:12.000000000","message":"nit: s/  / /","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":97,"context_line":"         database and close derivatives."},{"line_number":98,"context_line":".. [#f4] As of the Boston Forum (May 2017) known users of PostgreSQL"},{"line_number":99,"context_line":"         are:"},{"line_number":100,"context_line":"         - SUSE: investigating migrating to Galera"},{"line_number":101,"context_line":"         - Huawei: not actually PostgreSQL, but an internal db based on pg."},{"line_number":102,"context_line":"         - Windriver"},{"line_number":103,"context_line":"         - SAP: consuming OpenStack directly, not via vendor. They"},{"line_number":104,"context_line":"         did not respond to the user survey, so their usage is not"},{"line_number":105,"context_line":"         recorded there."},{"line_number":106,"context_line":".. [#f5] Common issues include the fact that MySQL char strings are case"},{"line_number":107,"context_line":"         insensitive by default. That storage defaults for utf8 are"},{"line_number":108,"context_line":"         different between the two."}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_9e92054b","line":105,"range":{"start_line":100,"start_character":9,"end_line":105,"end_character":24},"updated":"2017-06-12 21:47:12.000000000","message":"Maybe it doesn\u0027t matter, but these don\u0027t render as a list:\n\nhttp://docs-draft.openstack.org/80/427880/10/check/gate-governance-docs-ubuntu-xenial/cd4d07d//doc/build/html/resolutions/20170201-postgresql-mysql-status.html#f4","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":104,"context_line":"         did not respond to the user survey, so their usage is not"},{"line_number":105,"context_line":"         recorded there."},{"line_number":106,"context_line":".. [#f5] Common issues include the fact that MySQL char strings are case"},{"line_number":107,"context_line":"         insensitive by default. That storage defaults for utf8 are"},{"line_number":108,"context_line":"         different between the two."}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_3e6d3132","line":107,"range":{"start_line":107,"start_character":33,"end_line":107,"end_character":37},"updated":"2017-06-12 21:47:12.000000000","message":"The?","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"fe78b5714b4df956c181b30031246c9719d0ddb9","unresolved":false,"context_lines":[{"line_number":105,"context_line":"         recorded there."},{"line_number":106,"context_line":".. [#f5] Common issues include the fact that MySQL char strings are case"},{"line_number":107,"context_line":"         insensitive by default. That storage defaults for utf8 are"},{"line_number":108,"context_line":"         different between the two."}],"source_content_type":"text/x-rst","patch_set":10,"id":"7f231b9d_1e68351f","line":108,"range":{"start_line":108,"start_character":19,"end_line":108,"end_character":34},"updated":"2017-06-12 21:47:12.000000000","message":"two what? MySQL and PG?","commit_id":"21ad13ef515828054aa56c8a698db172da4e21ae"}],"resolutions/20170201-postgresql-status.rst":[{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"ea13cd7dd24066c77790c108faf170f9184876ca","unresolved":false,"context_lines":[{"line_number":88,"context_line":"         - SAP - consuming OpenStack directly, not via vendor. They"},{"line_number":89,"context_line":"           did not respond to the user survey, so their usage is not"},{"line_number":90,"context_line":"           recorded there."},{"line_number":91,"context_line":".. [#f4] Common issues include the fact that MySQL char strings are case"},{"line_number":92,"context_line":"       insensitive by default. That storage defaults for utf8 are"},{"line_number":93,"context_line":"       different between the two."}],"source_content_type":"text/x-rst","patch_set":6,"id":"ff0f0b1f_27d1b6ce","line":93,"range":{"start_line":91,"start_character":3,"end_line":93,"end_character":33},"updated":"2017-05-18 15:53:00.000000000","message":"It could be argued that this is a *good* thing as it forces us to be explicit about what we expect in the DB rather than relying on defaults.  (Since sqlite also differs from MySQL in some ways.)","commit_id":"c5bb6bed85a9c71eb5c74a49e5c193ddfa0699e2"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"9b437a0884a15d335fd227fb168a386ae974733c","unresolved":false,"context_lines":[{"line_number":88,"context_line":"         - SAP - consuming OpenStack directly, not via vendor. They"},{"line_number":89,"context_line":"           did not respond to the user survey, so their usage is not"},{"line_number":90,"context_line":"           recorded there."},{"line_number":91,"context_line":".. [#f4] Common issues include the fact that MySQL char strings are case"},{"line_number":92,"context_line":"       insensitive by default. That storage defaults for utf8 are"},{"line_number":93,"context_line":"       different between the two."}],"source_content_type":"text/x-rst","patch_set":6,"id":"ff0f0b1f_41b80fc7","line":93,"range":{"start_line":91,"start_character":3,"end_line":93,"end_character":33},"in_reply_to":"ff0f0b1f_27d1b6ce","updated":"2017-05-18 20:17:52.000000000","message":"I agree that we should be explicit. However, we should just be explicit. Supporting multiple database backends just so that one of them might accidentally point to a mis-configuration of another one, while a thing that will produce positive results some times, doesn\u0027t seem like a GREAT design for long-term success. I think focusing on the other reasons (users in the field) and both the costs and benefits of potential path will bring us all more of the happy.","commit_id":"c5bb6bed85a9c71eb5c74a49e5c193ddfa0699e2"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":1,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":2,"context_line":" 2017-02-01 Document current level of Postgresql support in community"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""},{"line_number":5,"context_line":"If you look at the myriad of deployment options in OpenStack,"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_97b89e76","line":2,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem grew up around MySQL [#f2]_, and has settled on MySQL as the"},{"line_number":10,"context_line":"backend for most deploys. [#f3]_ While the user survey is non"},{"line_number":11,"context_line":"authoritative, it\u0027s reported bias (\u003e 10 to 1 mysql vs. pg) is also"},{"line_number":12,"context_line":"mirrored in openstack-operator mailing list posts, ask.openstack.org"},{"line_number":13,"context_line":"posts, and content during operator meetups."},{"line_number":14,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_e98ffe20","line":11,"updated":"2017-05-19 08:23:50.000000000","message":"Actually the April 2017 survey says:\n\n\"MySQL with Galera also dropped significantly, [...] this follows a longer term trend where deployments are moving away from MySQL [...] to MariaDB *and* *others*. \"\n\n(OpenStack User Survey April 2017 Page 47)","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":8,"context_line":"a few common patterns that have emerged. One of which is that our"},{"line_number":9,"context_line":"ecosystem grew up around MySQL [#f2]_, and has settled on MySQL as the"},{"line_number":10,"context_line":"backend for most deploys. [#f3]_ While the user survey is non"},{"line_number":11,"context_line":"authoritative, it\u0027s reported bias (\u003e 10 to 1 mysql vs. pg) is also"},{"line_number":12,"context_line":"mirrored in openstack-operator mailing list posts, ask.openstack.org"},{"line_number":13,"context_line":"posts, and content during operator meetups."},{"line_number":14,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_3285365b","line":11,"in_reply_to":"ff0f0b1f_e98ffe20","updated":"2017-05-21 17:23:48.000000000","message":"For the purposes of this discussion MySQL is defined as noted in the footnote. Perhaps a parenthetical here like \"MySQL (as defined in [#f2]_) would make it clear for folks that there is a definition.\n\nSimilarly, Huawei do not actually run PG, they run a PG-based DB. We mention this later, but maybe for the purposes of this we should also make a definition and reference it like \"PostgreSQL (as defined in [#fX]_)\"","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":15,"context_line":"Upstream support statements in Open Source are complicated. They are a"},{"line_number":16,"context_line":"combination of developers designing with backends in mind, developers"},{"line_number":17,"context_line":"catching issues during code review, various levels of automated QA,"},{"line_number":18,"context_line":"documentation of multiple forms, knowledgable operators fielding"},{"line_number":19,"context_line":"questions of performance tuning, reference architectures, and general"},{"line_number":20,"context_line":"best practices. On all of these metrics, the OpenStack community very"},{"line_number":21,"context_line":"strong leans in the MySQL direction."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_8c0c103c","line":18,"updated":"2017-05-19 08:23:50.000000000","message":"knowledgeable","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"da100cfdecde239ab64bd4f9d6a7015a7bf9132a","unresolved":false,"context_lines":[{"line_number":18,"context_line":"documentation of multiple forms, knowledgable operators fielding"},{"line_number":19,"context_line":"questions of performance tuning, reference architectures, and general"},{"line_number":20,"context_line":"best practices. On all of these metrics, the OpenStack community very"},{"line_number":21,"context_line":"strong leans in the MySQL direction."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"There is also a question about future work. All work has a cost, and"},{"line_number":24,"context_line":"includes find people willing to do it, and fluent in either the"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_41148fa0","line":21,"updated":"2017-05-18 20:27:20.000000000","message":"strongly #notaminusone","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"d065db410039cecd09d171c068cdb07407625ebe","unresolved":false,"context_lines":[{"line_number":21,"context_line":"strong leans in the MySQL direction."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"There is also a question about future work. All work has a cost, and"},{"line_number":24,"context_line":"includes find people willing to do it, and fluent in either the"},{"line_number":25,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":26,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":27,"context_line":"MySQL, we make the chances of them happening smaller. These include"}],"source_content_type":"text/x-rst","patch_set":7,"id":"df140735_1eecc8f1","line":24,"updated":"2017-06-07 15:35:59.000000000","message":"finding*","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":25,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":26,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":27,"context_line":"MySQL, we make the chances of them happening smaller. These include"},{"line_number":28,"context_line":"(but are not limited to):"},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"- OpenStack services taking a more active role in managing the DBMS"},{"line_number":31,"context_line":"- The ability to have zero down time upgrade for services such as"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_8c81b04d","line":28,"updated":"2017-05-19 08:23:50.000000000","message":"So I think this is getting to the gut of the proposal. I think what you\u0027re saying is : \n\n\"Feature work for OpenStack might enhance services for MySQL (and its derivatives) more than for other databases like for example PostgreSQL and others\". \n\nSo with other words, MySQL is the first class citizen in the OpenStack upstream deployment, and any other database comes 2nd class. \n\nIf so, then it would be good to explicitely mention that this is not about PostgreSQL specifically but about any database not derived from the MySQL/MariaDB flavor (depending on your base OS).","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":25,"context_line":"abstraction layers, or the underlying datastores. If we prevent any"},{"line_number":26,"context_line":"forward progress on these unless PostgreSQL is also enhanced equal to"},{"line_number":27,"context_line":"MySQL, we make the chances of them happening smaller. These include"},{"line_number":28,"context_line":"(but are not limited to):"},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"- OpenStack services taking a more active role in managing the DBMS"},{"line_number":31,"context_line":"- The ability to have zero down time upgrade for services such as"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_72928ea0","line":28,"in_reply_to":"ff0f0b1f_8c81b04d","updated":"2017-05-21 17:23:48.000000000","message":"Yes, I agree with that characterization. The developer community designs around the MySQL family. Other DB backends that work do so by a combination of happenstance and the sqlalchemy abstraction layer. Things accidentally working isn\u0027t a good basis for people running a production service.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":34,"context_line":"- Consistency around case sensitivity collation defaults that lead to"},{"line_number":35,"context_line":"  strange bugs around searching/updating names in resources."},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We should clearly state that Postgresql isn\u0027t first class supported in"},{"line_number":38,"context_line":"Upstream OpenStack, which has been the reality for many years."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"As a set of concrete steps we should do the following:"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_57b20696","line":37,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"da100cfdecde239ab64bd4f9d6a7015a7bf9132a","unresolved":false,"context_lines":[{"line_number":35,"context_line":"  strange bugs around searching/updating names in resources."},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We should clearly state that Postgresql isn\u0027t first class supported in"},{"line_number":38,"context_line":"Upstream OpenStack, which has been the reality for many years."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_a1248b0e","line":38,"updated":"2017-05-18 20:27:20.000000000","message":"I\u0027d even go further to \"and actually never has been\"","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":35,"context_line":"  strange bugs around searching/updating names in resources."},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We should clearly state that Postgresql isn\u0027t first class supported in"},{"line_number":38,"context_line":"Upstream OpenStack, which has been the reality for many years."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_12ed9224","line":38,"in_reply_to":"ff0f0b1f_24850209","updated":"2017-05-21 17:23:48.000000000","message":"I think the key in my mind is \"a normal consideration in design choices\". We may not gate on ceph, but a substantial portion of our deployments use it, it\u0027s well integrated, has champions and nobody would dream of designing a new feature that would somehow architecturally break ceph. So I consider ceph support to be \"first class\" for us. I do not believe we can say the same thing about PG - a design of a feature could TOTALLY be done in such a way as to either break PG or at least be a really bad idea for anyone doing PG in production and I\u0027m pretty sure we\u0027d never know until we released it.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"ba55a7cab04751590d07fb258965949d9f316438","unresolved":false,"context_lines":[{"line_number":35,"context_line":"  strange bugs around searching/updating names in resources."},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We should clearly state that Postgresql isn\u0027t first class supported in"},{"line_number":38,"context_line":"Upstream OpenStack, which has been the reality for many years."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_24850209","line":38,"in_reply_to":"ff0f0b1f_841bd519","updated":"2017-05-19 22:36:32.000000000","message":"What does \"first class\" mean then, or \"never has been\". We used to gate on it. There are lots of things we don\u0027t gate on (ceph) but still support.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"443a2c2f1ba6a3075e8eea1847b64bdeb4adfe97","unresolved":false,"context_lines":[{"line_number":35,"context_line":"  strange bugs around searching/updating names in resources."},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We should clearly state that Postgresql isn\u0027t first class supported in"},{"line_number":38,"context_line":"Upstream OpenStack, which has been the reality for many years."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"As a set of concrete steps we should do the following:"},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_841bd519","line":38,"in_reply_to":"ff0f0b1f_a1248b0e","updated":"2017-05-18 20:53:44.000000000","message":"fair, given that moving to \"supported\" was because I fixed 4 nova bugs and made it gating :)","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":43,"context_line":""},{"line_number":44,"context_line":"    .. warning::"},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":47,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_572566c1","line":46,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":44,"context_line":"    .. warning::"},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":47,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_094b52b0","line":47,"updated":"2017-05-19 08:23:50.000000000","message":"This (re-) introduces the \"supported\" wording which is only weakenend later that OpenStack vendors can overwrite the support statement in the amendmend later. I would suggest a less ambiguous wording that makes it imho equally clear what you\u0027re trying to say (I think): \n\n  \"While OpenStack is open to be used with many SQL database \n  implementations, please be aware that anything other than MariaDB or\n   MySQL (depending on your Operating System distribution) is much less\n   tested within OpenStack and may trigger compatibility issues or cause\n  feature  degradations. If you rely on the OpenStack upstream community for \n  support of your particular OpenStack installation, you are proceeding at\n   your own risk by using anything else other than the default for your\n   distribution documented and upstream tested SQL database backend.\"","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":44,"context_line":"    .. warning::"},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":47,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_72e72e42","line":47,"in_reply_to":"ff0f0b1f_094b52b0","updated":"2017-05-21 17:23:48.000000000","message":"I gave a suggestion in Chris Dent\u0027s proposal - but I\u0027ll somewhat repeat it here:\n\n\"be aware that the OpenStack developer community is primarily focused on the MySQL family of database as a backend.\" (I worry about focusing too much weight to the word \"tested\" as that might imply to someone that simply re-enabling the PG devstack jobs would \u0027fix\u0027 things.)\n\nThe problem we\u0027re trying to warn about isn\u0027t lack of a testing job, but lack of design attention and thought. The coverage we\u0027d get by enabling testing today would really only be about SQL-level correctness, not validation of intent or fitness for purpose. If we gained developer design attention and thought, we could certainly add tests that would attempt to validate those things further, but the fundamental issue at hand currently isn\u0027t the testing- the testing is but a symptom of a deeper issue.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":46,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":47,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_a99fc601","line":49,"updated":"2017-05-19 08:23:50.000000000","message":"Again, the current wording in the installation guide is: \n(quote from: openstack-manuals/doc/install-guide/source/environment-sql-database.rst )\n\n  Most OpenStack services use an SQL database to store information. The\n  database typically runs on the controller node. The procedures in this\n  guide use MariaDB or MySQL depending on the distribution. OpenStack\n  services also support other SQL databases including\n  `PostgreSQL \u003chttps://www.postgresql.org/\u003e`__.\n\nAre you suggesting that MariaDB is no longer an option or are you saying that anything other than  the default documented for the distribution (which means its the default tested one for that distribution ) is not a good idea?\n\nWhat if upstream is testing on one distro in the gate based on mysql, but the user chooses to deploy mariadb instead of mysql? is that \"supported\" or is it unsupported then?\n\nThe wording in the documentation is currently saying that operators should follow the OS default flavor of MySQL. and it would be good to make it explicit here on whether we consider it completely interchangeable (which I think isn\u0027t a good idea because the non-default mySQL variant of the OS mgiht be buggy or fairly outdated) or whether we say upstream only supports the documented default MySQL family flavor.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":46,"context_line":"       While you may find reference to Postgresql within the project"},{"line_number":47,"context_line":"       trees, be aware that it\u0027s much less tested and supported than"},{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_12d6b250","line":49,"in_reply_to":"ff0f0b1f_a99fc601","updated":"2017-05-21 17:23:48.000000000","message":"Same as above comment. MySQL colloqually means the set of MySQL derived databases.\n\nI also think we need to add to the list of things we need to do, in addition to adding the warning, is removing the statements Dirk pasted pointing people towards PostgreSQL as an option. We should not add such statements back until such a time that we\u0027re all comfortable that PG is actually getting appropriate design attention and developer support.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":48,"context_line":"       MySQL as a backend. Proceed at your own risk in using something"},{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_3720b2b0","line":51,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."},{"line_number":55,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_c99e7ab7","line":52,"updated":"2017-05-19 08:23:50.000000000","message":"SUSE","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."},{"line_number":55,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_f2cc5ebd","line":52,"in_reply_to":"ff0f0b1f_c99e7ab7","updated":"2017-05-21 17:23:48.000000000","message":"TIL ... I totally thought it was SuSE in my brain.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"9fc0575caae667c16102da05eb301347537eb28f","unresolved":false,"context_lines":[{"line_number":49,"context_line":"       other than MySQL as the database for OpenStack."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."},{"line_number":55,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_036182d3","line":52,"in_reply_to":"ff0f0b1f_f2cc5ebd","updated":"2017-05-23 16:30:00.000000000","message":"It was...name changed in 2004 apparently.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"d065db410039cecd09d171c068cdb07407625ebe","unresolved":false,"context_lines":[{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"- Communicate to the OpenStack Foundation that there needs to be a way"},{"line_number":57,"context_line":"  for distributors to report back aggregate data about their users,"}],"source_content_type":"text/x-rst","patch_set":7,"id":"df140735_fe8d942a","line":54,"updated":"2017-06-07 15:35:59.000000000","message":"on an*","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":51,"context_line":"- Investigate the viability of the migration path from Postgresql to"},{"line_number":52,"context_line":"  MySQL (Suse team is currently doing this, will report back). This"},{"line_number":53,"context_line":"  provides data, and potentially a path forward, for people that are"},{"line_number":54,"context_line":"  feeling trapped on unsupported upstream leaf."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"- Communicate to the OpenStack Foundation that there needs to be a way"},{"line_number":57,"context_line":"  for distributors to report back aggregate data about their users,"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_2993b6fc","line":54,"updated":"2017-05-19 08:23:50.000000000","message":"s,unsupported,untested,","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":62,"context_line":"  is via the User Survey and what vendors decide to share."},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"In the future we may want to make stronger statements here, but the"},{"line_number":65,"context_line":"important thing is being clear to users that Postgresql is not first"},{"line_number":66,"context_line":"class supported upstream, and has not been for quite some time. And"},{"line_number":67,"context_line":"that there is currently no will or effort being applied to change that"},{"line_number":68,"context_line":"fact in upstream development."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_172feea1","line":65,"updated":"2017-05-19 08:23:50.000000000","message":"PostgreSQL","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"ba55a7cab04751590d07fb258965949d9f316438","unresolved":false,"context_lines":[{"line_number":63,"context_line":""},{"line_number":64,"context_line":"In the future we may want to make stronger statements here, but the"},{"line_number":65,"context_line":"important thing is being clear to users that Postgresql is not first"},{"line_number":66,"context_line":"class supported upstream, and has not been for quite some time. And"},{"line_number":67,"context_line":"that there is currently no will or effort being applied to change that"},{"line_number":68,"context_line":"fact in upstream development."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"This doesn\u0027t mean that vendors can\u0027t claim support here for their"},{"line_number":71,"context_line":"customers. The OpenStack code is Apache2 licensed. But for"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_24e862b2","line":68,"range":{"start_line":66,"start_character":64,"end_line":68,"end_character":29},"updated":"2017-05-19 22:36:32.000000000","message":"Oh I don\u0027t know about this, I just fixed this bug on Tuesday:\n\nhttps://bugs.launchpad.net/neutron/+bug/1691176\n\nWhich made the postgresql job start passing in the experimental queue. I\u0027d be totally fine with running the nova-next job, or any other voting job that runs against nova, run with postgresql as the backend.\n\nI feel like I\u0027m constantly needing to remind people that we didn\u0027t remove the postgresql job because it was broken, we removed it because it wasn\u0027t as much of a snowflake was it was in years before when that job also ran with keystone under eventlet. Turning it back on in a voting job is trivial.\n\nWould this statement about will or effort change if I pushed a patch to run a voting job on nova under postgresql instead of mysql?","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":63,"context_line":""},{"line_number":64,"context_line":"In the future we may want to make stronger statements here, but the"},{"line_number":65,"context_line":"important thing is being clear to users that Postgresql is not first"},{"line_number":66,"context_line":"class supported upstream, and has not been for quite some time. And"},{"line_number":67,"context_line":"that there is currently no will or effort being applied to change that"},{"line_number":68,"context_line":"fact in upstream development."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"This doesn\u0027t mean that vendors can\u0027t claim support here for their"},{"line_number":71,"context_line":"customers. The OpenStack code is Apache2 licensed. But for"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_92c1a283","line":68,"range":{"start_line":66,"start_character":64,"end_line":68,"end_character":29},"in_reply_to":"ff0f0b1f_24e862b2","updated":"2017-05-21 17:23:48.000000000","message":"Not for me - but I think I\u0027ve covered that in my comments above about design intent.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":70,"context_line":"This doesn\u0027t mean that vendors can\u0027t claim support here for their"},{"line_number":71,"context_line":"customers. The OpenStack code is Apache2 licensed. But for"},{"line_number":72,"context_line":"organizations bypassing vendors and going straight to upstream, the"},{"line_number":73,"context_line":"reality of the world should be made clear."},{"line_number":74,"context_line":""},{"line_number":75,"context_line":""},{"line_number":76,"context_line":".. rubric:: Footnotes"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_c94f9a90","line":73,"updated":"2017-05-19 08:23:50.000000000","message":"I agree with all of the wording here. It would be nice to see it reflected in the actual concrete action step for the documentation (line 42++ of this proposal) rather than the amendmend of a governance resolution that is in no way linked by the concrete documentation.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":70,"context_line":"This doesn\u0027t mean that vendors can\u0027t claim support here for their"},{"line_number":71,"context_line":"customers. The OpenStack code is Apache2 licensed. But for"},{"line_number":72,"context_line":"organizations bypassing vendors and going straight to upstream, the"},{"line_number":73,"context_line":"reality of the world should be made clear."},{"line_number":74,"context_line":""},{"line_number":75,"context_line":""},{"line_number":76,"context_line":".. rubric:: Footnotes"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_3238969e","line":73,"in_reply_to":"ff0f0b1f_c94f9a90","updated":"2017-05-21 17:23:48.000000000","message":"Honestly, and this came up elsewhere this past week as well - it makes me feel we need better documentation broadly on what \"supported\" means, what we as an upstream feel we provide, and what things we feel consumers are either on their own for or should seek help from a vendor for.\n\nOS Distro Packages is an example of a thing where we have an understanding upstream that if you want them you should get them from a vendor. However, other than tribal knowledge, I do not believe that is documented, nor are the philosophical underpinnings of that tribal knowledge. The same could be said of \"supporting LTS releases\" - and is a discussion happening right now about publication of container images.\n\nIf we can produce such a document, I agree that we should be aggressive in including summaries of it with links directly in our documentation in places like this.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":76,"context_line":".. rubric:: Footnotes"},{"line_number":77,"context_line":""},{"line_number":78,"context_line":".. [#f1] April 2016 User Survey"},{"line_number":79,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":80,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":81,"context_line":"         and its close variants, MariaDB, Galera."},{"line_number":82,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of PostgreSQL"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_a9ab4687","line":79,"updated":"2017-05-19 08:23:50.000000000","message":"this should be updated to the April 2017 survey instead. There is nothing older than the newspaper from yesterday.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":78,"context_line":".. [#f1] April 2016 User Survey"},{"line_number":79,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":80,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":81,"context_line":"         and its close variants, MariaDB, Galera."},{"line_number":82,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of PostgreSQL"},{"line_number":83,"context_line":"         are:"},{"line_number":84,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_c91d5a74","line":81,"updated":"2017-05-19 08:23:50.000000000","message":"\"close variants MariaDB and Galera\"","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":78,"context_line":".. [#f1] April 2016 User Survey"},{"line_number":79,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":80,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":81,"context_line":"         and its close variants, MariaDB, Galera."},{"line_number":82,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of PostgreSQL"},{"line_number":83,"context_line":"         are:"},{"line_number":84,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_d25ddae9","line":81,"in_reply_to":"ff0f0b1f_c91d5a74","updated":"2017-05-21 17:23:48.000000000","message":"I ACTUALLY think what we should say here is:\n\nMySQL is used as a stand in term in this document for the family of MySQL-derived and compatible database servers running InnoDB-derived storage engines with or without Galera clustering. MySQL-derived database servers include but are not necessarily limited to Oracle MySQL, MariaDB and Percona while InnoDB-derived storage engines include but are not necessarily limited to Oracle InnoDB and Percona XtraDB. MySQL-derived but incompatible servers such as Drizzle are not included, nor are other MySQL storage engines such as MyRocks, TokuDB or NDB.\n\nI also think we should define that set clearly somewhere so that it\u0027s not just in this document. The above is what OpenStack developers typically mean when they use the word \"MySQL\" - but it can be confusing.\n\nWe had angst a little while ago when someone from Oracle attempted to add support for NDB Cluster. As a former NDB consultant I think it would be an amazing choice for OpenStack to run on - especially given our increasing use in Telcos. BUT - as an OpenStack developer I have to admit that we have even less community knowledge or operator experience with the NDB storage engine, and there is no way we\u0027d be able to adequately include NDB in our database design decisions.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":11816,"name":"mike bayer","display_name":"Mike Bayer","email":"mike_mp@zzzcomputing.com","username":"zzzeek","status":"Red Hat"},"change_message_id":"8a08c8bfb169b1580c086e66fe99a2700b036e64","unresolved":false,"context_lines":[{"line_number":78,"context_line":".. [#f1] April 2016 User Survey"},{"line_number":79,"context_line":"         https://www.openstack.org/assets/survey/April-2016-User-Survey-Report.pdf"},{"line_number":80,"context_line":".. [#f2] MySQL is used as a stand in term in this document for MySQL"},{"line_number":81,"context_line":"         and its close variants, MariaDB, Galera."},{"line_number":82,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of PostgreSQL"},{"line_number":83,"context_line":"         are:"},{"line_number":84,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_5531b4eb","line":81,"in_reply_to":"ff0f0b1f_d25ddae9","updated":"2017-05-21 19:35:50.000000000","message":"\u003e We had angst a little while ago when someone from Oracle attempted to add support for NDB Cluster.\n\nthat\u0027s still happening and is ongoing:\n\nhttps://review.openstack.org/#/c/427970/\n\nthere are also gerrits for most projects (don\u0027t think the contributor knows how to connect them on change-id), here\u0027s nova for example: https://review.openstack.org/#/c/446643/\n\nif you have comments on the feasibility of these changes that would be helpful.","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":6593,"name":"Dirk Mueller","email":"dirk@dmllr.de","username":"dmllr"},"change_message_id":"606016273982b968ce070a465fb15f418bc056fe","unresolved":false,"context_lines":[{"line_number":82,"context_line":".. [#f3] As of the Boston Forum (May 2017) known users of PostgreSQL"},{"line_number":83,"context_line":"         are:"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"         - Suse - (investigating migrating to Galera)"},{"line_number":86,"context_line":"         - Huawei - not actually PostgreSQL, but an internal db based"},{"line_number":87,"context_line":"           on pg."},{"line_number":88,"context_line":"         - Windriver"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_69af6ee3","line":85,"updated":"2017-05-19 08:23:50.000000000","message":"SUSE","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"dc4157fa11a216586d68fdf0ce9941f8ee6942f2","unresolved":false,"context_lines":[{"line_number":89,"context_line":"         - SAP - consuming OpenStack directly, not via vendor. They"},{"line_number":90,"context_line":"           did not respond to the user survey, so their usage is not"},{"line_number":91,"context_line":"           recorded there."},{"line_number":92,"context_line":".. [#f4] Common issues include the fact that MySQL char strings are case"},{"line_number":93,"context_line":"         insensitive by default. That storage defaults for utf8 are"},{"line_number":94,"context_line":"         different between the two."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ff0f0b1f_52ec2af2","line":92,"updated":"2017-05-21 17:23:48.000000000","message":"I\u0027m going to follow up on this on the mailing list- but this is actually (weirdly) a super simple example that strikes to the heart of a fundamental philosophical question that we do not have an answer to (although I believe people on both sides of this PG issue have opinions on)\n\nThat is:\n\n* Should OpenStack assume the existence of an external database tied that it can assume is being managed and treat as an black-box on the other side of a connection string?\n\nOR:\n\n* Should OpenStack take an active and opinionated role in managing the database tier?\n\nI say the MySQL char strings issue is one that is simple and to the heart of the matter, because it\u0027s one that could be managed by the operator DBA role by making some settings in the my.cnf file related to default character set and collation - or it\u0027s one that OpenStack could take an assertive role on making based on its behavioral expectations.\n\nThis manifests in design choices. How do we ensure rolling upgrades of database tables with no downtime is a heavily requested feature - but in a stronger \"external\" world view it could be considered \"a job for the DBA to sort out\". In a stronger \"active\" world view, rather than providing schema upgrade scripts, OpenStack services could notice that they have migrations available, then figure out if they\u0027re on Galera or not. If they are on galera, they could know to set wsrep_OSU_method\u003d\u0027RSU\u0027 then apply the schema update on each of the galera nodes and be set up to read and write data in an appropriate manner during the transition. But OpenStack doing active coordination and remediation of schema upgrades while the server is running is highly dependent on deployment architecture, so without being very opinionated about a specific architecture or set of architectures it\u0027s unpossible.\n\nBecause we don\u0027t have a firm answer on what our philosophy is on the approach, we wind up in a mushy middle ground where we\u0027re sometimes active and prescriptive but sometimes leave things for the deployers to take care of. And back to this review, that lack of shared understanding I believe makes it really hard for people to have the same understanding of what \"the problem\" is around PG. If a human is of the worldview that the database tier will be managed by the capable team of existing PG admins, then questions of the difficulties presented as part of OpenStack design are more than passing strange.\n\nI said I\u0027d follow up on the mailing list then I just word-vomited here. Go me. I\u0027ll go do mailing list now...","commit_id":"aafcc7413f4812df10362a9c95ea8bad23a81dfa"}],"resolutions/20170613-postgresql-status.rst":[{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"5995659dc95f07165457c7ec95fb10e73e8e4a4c","unresolved":false,"context_lines":[{"line_number":40,"context_line":"       be aware that the upstream OpenStack developer community"},{"line_number":41,"context_line":"       is so far focusing on the MySQL family of databases"},{"line_number":42,"context_line":"       as a backend and tests OpenStack against the default"},{"line_number":43,"context_line":"       MySQL [#f2]_ from your Operating System distribution."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"       If you exclusively rely on the OpenStack upstream"},{"line_number":46,"context_line":"       community and not on an OpenStack vendor for support of your"}],"source_content_type":"text/x-rst","patch_set":12,"id":"7f231b9d_2df94c4c","line":43,"updated":"2017-06-14 10:34:08.000000000","message":"because this block is being rendered as something akin to \u003cpre\u003e (because of \u0027preference::\u0027 above) the footnote does not resolve, and in any case, in the context of the block being used verbatim in the docs, there would be no footprint.\n\nNot clear what the desired outcome is for this.","commit_id":"38d3df2567043b7fc26c5b87a3a7bfaa902075e2"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"b5f6503c2c629d9e4026bf86e785859d406cd692","unresolved":false,"context_lines":[{"line_number":40,"context_line":"       be aware that the upstream OpenStack developer community"},{"line_number":41,"context_line":"       is so far focusing on the MySQL family of databases"},{"line_number":42,"context_line":"       as a backend and tests OpenStack against the default"},{"line_number":43,"context_line":"       MySQL [#f2]_ from your Operating System distribution."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"       If you exclusively rely on the OpenStack upstream"},{"line_number":46,"context_line":"       community and not on an OpenStack vendor for support of your"}],"source_content_type":"text/x-rst","patch_set":12,"id":"7f231b9d_eb03a0bd","line":43,"in_reply_to":"7f231b9d_2df94c4c","updated":"2017-06-20 10:38:42.000000000","message":"I think that\u0027s intentional as its the raw data that wants to be added into the docs?","commit_id":"38d3df2567043b7fc26c5b87a3a7bfaa902075e2"},{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"5995659dc95f07165457c7ec95fb10e73e8e4a4c","unresolved":false,"context_lines":[{"line_number":80,"context_line":"         - SUSE: investigating migrating to Galera"},{"line_number":81,"context_line":"         - Huawei: not actually PostgreSQL, but an internal db based on pg."},{"line_number":82,"context_line":"         - Windriver"},{"line_number":83,"context_line":"         - SAP: consuming OpenStack directly, not via vendor. They"},{"line_number":84,"context_line":"         did not respond to the user survey, so their usage is not"},{"line_number":85,"context_line":"         recorded there."}],"source_content_type":"text/x-rst","patch_set":12,"id":"7f231b9d_8d8980cb","line":83,"updated":"2017-06-14 10:34:08.000000000","message":"these items don\u0027t render as a list in the html output, so some other formatting is needed.","commit_id":"38d3df2567043b7fc26c5b87a3a7bfaa902075e2"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"b5f6503c2c629d9e4026bf86e785859d406cd692","unresolved":false,"context_lines":[{"line_number":80,"context_line":"         - SUSE: investigating migrating to Galera"},{"line_number":81,"context_line":"         - Huawei: not actually PostgreSQL, but an internal db based on pg."},{"line_number":82,"context_line":"         - Windriver"},{"line_number":83,"context_line":"         - SAP: consuming OpenStack directly, not via vendor. They"},{"line_number":84,"context_line":"         did not respond to the user survey, so their usage is not"},{"line_number":85,"context_line":"         recorded there."}],"source_content_type":"text/x-rst","patch_set":12,"id":"7f231b9d_cb2ea4f9","line":83,"in_reply_to":"7f231b9d_8d8980cb","updated":"2017-06-20 10:38:42.000000000","message":"+1 needs a fix, but I don\u0027t want to block on that.","commit_id":"38d3df2567043b7fc26c5b87a3a7bfaa902075e2"}]}