)]}'
{"doc/source/placement.rst":[{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"b6afbaab83f5971c0f958a5ee7b15f1944c131b8","unresolved":false,"context_lines":[{"line_number":152,"context_line":"  ``[placement]`` section of nova.conf on the compute is configured. As"},{"line_number":153,"context_line":"  mentioned in the deployment steps above, the Placement service should be"},{"line_number":154,"context_line":"  deployed by this point so the computes can register and start reporting"},{"line_number":155,"context_line":"  inventory, allocation and usage information. If the computes are deployed"},{"line_number":156,"context_line":"  and configured `before` the Placement service, they will continue to try"},{"line_number":157,"context_line":"  and reconnect in a loop so that you do not need to restart the nova-compute"},{"line_number":158,"context_line":"  process to talk to the Placement service after the compute is properly"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_eca787a4","line":155,"updated":"2017-01-31 19:31:42.000000000","message":"Not sure it matters but \u0027usage\u0027 is probably redundant here when you\u0027ve also got \u0027allocation\u0027.","commit_id":"a5b6d7f56807e2dd987ca9b92f9717b770aa1e23"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"54c286a52bec6ce04f5f9cdf560ca80209e04674","unresolved":false,"context_lines":[{"line_number":152,"context_line":"  ``[placement]`` section of nova.conf on the compute is configured. As"},{"line_number":153,"context_line":"  mentioned in the deployment steps above, the Placement service should be"},{"line_number":154,"context_line":"  deployed by this point so the computes can register and start reporting"},{"line_number":155,"context_line":"  inventory, allocation and usage information. If the computes are deployed"},{"line_number":156,"context_line":"  and configured `before` the Placement service, they will continue to try"},{"line_number":157,"context_line":"  and reconnect in a loop so that you do not need to restart the nova-compute"},{"line_number":158,"context_line":"  process to talk to the Placement service after the compute is properly"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_2c37cfff","line":155,"in_reply_to":"3a461143_eca787a4","updated":"2017-01-31 19:42:12.000000000","message":"Good point, I always think of the three because of the objects rather than the tables.","commit_id":"a5b6d7f56807e2dd987ca9b92f9717b770aa1e23"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"872a5da1122a178d3f0010326c3c93d1ae58291d","unresolved":false,"context_lines":[{"line_number":55,"context_line":"Deployment"},{"line_number":56,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"The placement-api service must be deployed at some point after you have"},{"line_number":59,"context_line":"upgraded to the 14.0.0 Newton release but before you can upgrade to the 15.0.0"},{"line_number":60,"context_line":"Ocata release. This is so that the resource tracker in the nova-compute service"},{"line_number":61,"context_line":"can populate resource provider (compute node) inventory and allocation"},{"line_number":62,"context_line":"information which will be used by the nova-scheduler service in Ocata."},{"line_number":63,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_e124cc61","line":60,"range":{"start_line":58,"start_character":0,"end_line":60,"end_character":15},"updated":"2017-02-01 15:43:49.000000000","message":"Now that we merged https://review.openstack.org/#/c/417961/ the Ocata FilterScheduler supports old Newton computes. So, tbc, we should loosen that phrase to \"during the maintenance window\". Even, the scheduler doesn\u0027t use the placement API until all computes are upgraded but ssssshhht :-)","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"9552fdb5e8fd08252932a5d12ce9a16a2365035a","unresolved":false,"context_lines":[{"line_number":55,"context_line":"Deployment"},{"line_number":56,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"The placement-api service must be deployed at some point after you have"},{"line_number":59,"context_line":"upgraded to the 14.0.0 Newton release but before you can upgrade to the 15.0.0"},{"line_number":60,"context_line":"Ocata release. This is so that the resource tracker in the nova-compute service"},{"line_number":61,"context_line":"can populate resource provider (compute node) inventory and allocation"},{"line_number":62,"context_line":"information which will be used by the nova-scheduler service in Ocata."},{"line_number":63,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_320d3638","line":60,"range":{"start_line":58,"start_character":0,"end_line":60,"end_character":15},"in_reply_to":"3a461143_e124cc61","updated":"2017-02-01 20:43:50.000000000","message":"I think it\u0027s probably fine to be a bit forceful about this...","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"872a5da1122a178d3f0010326c3c93d1ae58291d","unresolved":false,"context_lines":[{"line_number":101,"context_line":""},{"line_number":102,"context_line":"**4. Configure and restart nova-compute services**"},{"line_number":103,"context_line":""},{"line_number":104,"context_line":"The 14.0.0 Newton nova-compute service code will begin reporting resource"},{"line_number":105,"context_line":"provider inventory and usage information as soon as the placement API"},{"line_number":106,"context_line":"service is in place and can respond to requests via the endpoint registered"},{"line_number":107,"context_line":"in the service catalog."}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_21afb42c","line":104,"range":{"start_line":104,"start_character":3,"end_line":104,"end_character":17},"updated":"2017-02-01 15:43:49.000000000","message":"we should remove that, or make it clear that :\n- with Newton, it\u0027s optional to do so\n- with Ocata, the compute will fail to start","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"9552fdb5e8fd08252932a5d12ce9a16a2365035a","unresolved":false,"context_lines":[{"line_number":101,"context_line":""},{"line_number":102,"context_line":"**4. Configure and restart nova-compute services**"},{"line_number":103,"context_line":""},{"line_number":104,"context_line":"The 14.0.0 Newton nova-compute service code will begin reporting resource"},{"line_number":105,"context_line":"provider inventory and usage information as soon as the placement API"},{"line_number":106,"context_line":"service is in place and can respond to requests via the endpoint registered"},{"line_number":107,"context_line":"in the service catalog."}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_92334af6","line":104,"range":{"start_line":104,"start_character":3,"end_line":104,"end_character":17},"in_reply_to":"3a461143_21afb42c","updated":"2017-02-01 20:43:50.000000000","message":"We do say below:\n\n\"as soon as the placement API\nservice is in place and can respond to requests via the endpoint registered\nin the service catalog.\n\n``nova.conf`` on the compute nodes must be updated in the ``[placement]``\ngroup to contain credentials for making requests from nova-compute to the\nplacement-api service.\"\n\nSo it\u0027s semi-optional yeah, but again, I tend to think it\u0027s better to be a bit forceful about this stuff before upgrading to Ocata so there are fewer surprises.","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"872a5da1122a178d3f0010326c3c93d1ae58291d","unresolved":false,"context_lines":[{"line_number":171,"context_line":"  ``nova-scheduler`` service to Ocata and restart it, things will still work."},{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_531c9313","line":174,"range":{"start_line":174,"start_character":27,"end_line":174,"end_character":54},"updated":"2017-02-01 15:43:49.000000000","message":"[...] or after upgrading all your compute nodes.\n\nAmending nova.conf for the scheduler process is still mandatory in Ocata, that\u0027s just the fact we only use those values once all the computes are fully new.","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"c64eecfea9c20b44f00fd6615e661ca946e16ed9","unresolved":false,"context_lines":[{"line_number":171,"context_line":"  ``nova-scheduler`` service to Ocata and restart it, things will still work."},{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_87914e55","line":174,"range":{"start_line":174,"start_character":27,"end_line":174,"end_character":54},"in_reply_to":"3a461143_531c9313","updated":"2017-02-01 20:25:56.000000000","message":"\u003e [...] or after upgrading all your compute nodes.\n\nI\u0027m not sure what the context is for that.\n\n\u003e Amending nova.conf for the scheduler process is still mandatory in Ocata, that\u0027s just the fact we only use those values once all the computes are fully new.\n\nThat\u0027s already pointed out in the first upgrade bullet where it says nova-compute won\u0027t start without configuring it for placement.","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"d6100210ce583214f79b8c7b3d7f9698cba2b4a4","unresolved":false,"context_lines":[{"line_number":171,"context_line":"  ``nova-scheduler`` service to Ocata and restart it, things will still work."},{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_07671e26","line":174,"range":{"start_line":174,"start_character":27,"end_line":174,"end_character":54},"in_reply_to":"3a461143_87914e55","updated":"2017-02-01 20:37:45.000000000","message":"I\u0027ll change that last sentence to just this:\n\n\"However, once all computes are upgraded, the scheduler not being able to make requests to Placement will result in **NoValidHost** errors.\"","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"872a5da1122a178d3f0010326c3c93d1ae58291d","unresolved":false,"context_lines":[{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"},{"line_number":178,"context_line":"  computes are reporting into the Placement service, however, and the"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_f3c81fab","line":175,"range":{"start_line":175,"start_character":0,"end_line":175,"end_character":75},"updated":"2017-02-01 15:43:49.000000000","message":"I would put that bullet right after the first one about nova-compute, since it\u0027s related. From an operator perspective, I\u0027d like to know all the tasks I need to perform per service, so I\u0027d appreciate all the things presented by that way. That\u0027s a nit tho and we can expect operators to only upgrade once they fully end up reading documentation :-P","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"a671b4bb6a7be19db5622691766c9c394c3b6782","unresolved":false,"context_lines":[{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"},{"line_number":178,"context_line":"  computes are reporting into the Placement service, however, and the"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_27433a6e","line":175,"range":{"start_line":175,"start_character":0,"end_line":175,"end_character":75},"in_reply_to":"3a461143_e78db22b","updated":"2017-02-01 20:39:10.000000000","message":"(2:29:12 PM) bauzas: mriedem: and about the bullet good place, I meant that since operators need to modify their nova.conf for adding the placement section, it could be nice to let them know quite close that they also need to update ratios\n(2:29:20 PM) bauzas: but that\u0027s a nit","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"c64eecfea9c20b44f00fd6615e661ca946e16ed9","unresolved":false,"context_lines":[{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."},{"line_number":173,"context_line":"  However, the scheduler not being able to make requests to Placement will"},{"line_number":174,"context_line":"  result in a hard failure in the 16.0.0 Pike release."},{"line_number":175,"context_line":"* It is currently possible to exclude the ``CoreFilter``, ``RamFilter`` and"},{"line_number":176,"context_line":"  ``DiskFilter`` from the list of enabled FilterScheduler filters such that"},{"line_number":177,"context_line":"  scheduling decisions are not based on CPU, RAM or disk usage. Once all"},{"line_number":178,"context_line":"  computes are reporting into the Placement service, however, and the"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_e78db22b","line":175,"range":{"start_line":175,"start_character":0,"end_line":175,"end_character":75},"in_reply_to":"3a461143_f3c81fab","updated":"2017-02-01 20:25:56.000000000","message":"I\u0027m not sure how this bullet about the enabled/disabled filters is related to the first bullet about placement being required config for nova-compute.","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"872a5da1122a178d3f0010326c3c93d1ae58291d","unresolved":false,"context_lines":[{"line_number":182,"context_line":"  that type of resource for placement decisions, you will need to adjust the"},{"line_number":183,"context_line":"  corresponding ``cpu_allocation_ratio``, ``ram_allocation_ratio``, and/or"},{"line_number":184,"context_line":"  ``disk_allocation_ratio`` configuration options to be very high values, e.g."},{"line_number":185,"context_line":"  9999.0."},{"line_number":186,"context_line":""},{"line_number":187,"context_line":""},{"line_number":188,"context_line":"REST API"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3a461143_53c0137d","line":185,"updated":"2017-02-01 15:43:49.000000000","message":"Thanks, you better explained the problem than me in the release notes.","commit_id":"438a4075be6cf3a0805be2c7a99f5b31d82205d0"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"5e8d8f2e0188d6b3027416519e2853a5741011ca","unresolved":false,"context_lines":[{"line_number":157,"context_line":"  and reconnect in a loop so that you do not need to restart the nova-compute"},{"line_number":158,"context_line":"  process to talk to the Placement service after the compute is properly"},{"line_number":159,"context_line":"  configured."},{"line_number":160,"context_line":"* The ``nova.scheduler.filter_scheduler.and FilterScheduler`` in Ocata will"},{"line_number":161,"context_line":"  fallback to not using the Placement service as long as there are older"},{"line_number":162,"context_line":"  ``nova-compute`` services running in the deployment. This allows for rolling"},{"line_number":163,"context_line":"  upgrades of the computes to not affect scheduling for the FilterScheduler."}],"source_content_type":"text/x-rst","patch_set":3,"id":"3a461143_0ac8f373","line":160,"range":{"start_line":160,"start_character":40,"end_line":160,"end_character":44},"updated":"2017-02-02 09:53:31.000000000","message":"typo?","commit_id":"e3fec0b9c00e1557255f4e16414d047abe1512ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"340d7570ab554fa4f0f90f2c3714ce3aa29fce37","unresolved":false,"context_lines":[{"line_number":157,"context_line":"  and reconnect in a loop so that you do not need to restart the nova-compute"},{"line_number":158,"context_line":"  process to talk to the Placement service after the compute is properly"},{"line_number":159,"context_line":"  configured."},{"line_number":160,"context_line":"* The ``nova.scheduler.filter_scheduler.and FilterScheduler`` in Ocata will"},{"line_number":161,"context_line":"  fallback to not using the Placement service as long as there are older"},{"line_number":162,"context_line":"  ``nova-compute`` services running in the deployment. This allows for rolling"},{"line_number":163,"context_line":"  upgrades of the computes to not affect scheduling for the FilterScheduler."}],"source_content_type":"text/x-rst","patch_set":3,"id":"3a461143_423498dd","line":160,"range":{"start_line":160,"start_character":40,"end_line":160,"end_character":44},"in_reply_to":"3a461143_0ac8f373","updated":"2017-02-02 13:54:23.000000000","message":"Done","commit_id":"e3fec0b9c00e1557255f4e16414d047abe1512ae"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"5e8d8f2e0188d6b3027416519e2853a5741011ca","unresolved":false,"context_lines":[{"line_number":166,"context_line":"  and the resource providers (compute nodes) registered there. This means if"},{"line_number":167,"context_line":"  the computes are not reporting into Placement by Pike, build requests will"},{"line_number":168,"context_line":"  fail with **NoValidHost** errors."},{"line_number":169,"context_line":"* While the FilterScheduler technically depends on the Placement service"},{"line_number":170,"context_line":"  in Ocata, if you deploy the Placement service `after` you upgrade the"},{"line_number":171,"context_line":"  ``nova-scheduler`` service to Ocata and restart it, things will still work."},{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."}],"source_content_type":"text/x-rst","patch_set":3,"id":"3a461143_2a9ccf68","line":169,"range":{"start_line":169,"start_character":12,"end_line":169,"end_character":27},"updated":"2017-02-02 09:53:31.000000000","message":"``code``?","commit_id":"e3fec0b9c00e1557255f4e16414d047abe1512ae"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"340d7570ab554fa4f0f90f2c3714ce3aa29fce37","unresolved":false,"context_lines":[{"line_number":166,"context_line":"  and the resource providers (compute nodes) registered there. This means if"},{"line_number":167,"context_line":"  the computes are not reporting into Placement by Pike, build requests will"},{"line_number":168,"context_line":"  fail with **NoValidHost** errors."},{"line_number":169,"context_line":"* While the FilterScheduler technically depends on the Placement service"},{"line_number":170,"context_line":"  in Ocata, if you deploy the Placement service `after` you upgrade the"},{"line_number":171,"context_line":"  ``nova-scheduler`` service to Ocata and restart it, things will still work."},{"line_number":172,"context_line":"  The scheduler will gracefully handle the absence of the Placement service."}],"source_content_type":"text/x-rst","patch_set":3,"id":"3a461143_023a90cd","line":169,"range":{"start_line":169,"start_character":12,"end_line":169,"end_character":27},"in_reply_to":"3a461143_2a9ccf68","updated":"2017-02-02 13:54:23.000000000","message":"I don\u0027t think that\u0027s necessary - then I\u0027d have to do it everywhere in here, which would be annoying.","commit_id":"e3fec0b9c00e1557255f4e16414d047abe1512ae"}]}