)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"088b400d9d9e226b1c8f45807b42dd1812a19f52","unresolved":false,"context_lines":[{"line_number":7,"context_line":"[WIP] Remove deprecated cpu,ram,disk claiming in resource tracker"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"A while ago, we deprecated the scheduler filters for CPU, RAM, and Disk in"},{"line_number":10,"context_line":"favor of relying on placement. That means we\u0027re not passing limits down"},{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"},{"line_number":15,"context_line":"need to fix things like failing to claim unrelated root disk space on"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"df7087c5_2f1406a0","line":12,"range":{"start_line":10,"start_character":31,"end_line":12,"end_character":21},"updated":"2018-03-08 22:48:43.000000000","message":"If you\u0027re using the FilterScheduler. If you\u0027re using the CachingScheduler, which doesn\u0027t use Placement, you totally rely on these claims in the compute, so we can\u0027t remove those until we remove CachingScheduler.","commit_id":"c9804fee2c1d0b23a9e06a01b2e47a3c2977e5e4"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"05017ccd14fbcee970cd2af3578ee85bbf741c95","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Remove deprecated CPU, RAM, disk claiming in resource tracker"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"A while ago, we deprecated the scheduler filters for CPU, RAM, and Disk in"},{"line_number":10,"context_line":"favor of relying on placement. That means we\u0027re not passing limits down"},{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_a4727ae2","line":9,"range":{"start_line":9,"start_character":0,"end_line":9,"end_character":11},"updated":"2019-07-23 21:07:08.000000000","message":"I personally dislike stuff like this in commit messages since in 3 years it won\u0027t be awhile ago anymore. I\u0027d prefer to say something like \"In Stein\" or at least just say \"A while ago [1], ...\" and then at the bottom link to [1] Id62136d293da55e4bb639635ea5421a33b6c3ea2.","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"ed547ca2fe46ea60eadf6bbfd756b020cc4f211d","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Remove deprecated CPU, RAM, disk claiming in resource tracker"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"A while ago, we deprecated the scheduler filters for CPU, RAM, and Disk in"},{"line_number":10,"context_line":"favor of relying on placement. That means we\u0027re not passing limits down"},{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_b2a14e98","line":9,"range":{"start_line":9,"start_character":0,"end_line":9,"end_character":11},"in_reply_to":"7faddb67_a4727ae2","updated":"2019-07-24 15:26:48.000000000","message":"Done","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"05017ccd14fbcee970cd2af3578ee85bbf741c95","unresolved":false,"context_lines":[{"line_number":8,"context_line":""},{"line_number":9,"context_line":"A while ago, we deprecated the scheduler filters for CPU, RAM, and Disk in"},{"line_number":10,"context_line":"favor of relying on placement. That means we\u0027re not passing limits down"},{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_647c82b5","line":11,"range":{"start_line":11,"start_character":29,"end_line":11,"end_character":38},"updated":"2019-07-23 21:07:08.000000000","message":"thusly the thusness","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"ed547ca2fe46ea60eadf6bbfd756b020cc4f211d","unresolved":false,"context_lines":[{"line_number":8,"context_line":""},{"line_number":9,"context_line":"A while ago, we deprecated the scheduler filters for CPU, RAM, and Disk in"},{"line_number":10,"context_line":"favor of relying on placement. That means we\u0027re not passing limits down"},{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_d29e4a53","line":11,"range":{"start_line":11,"start_character":29,"end_line":11,"end_character":38},"in_reply_to":"7faddb67_647c82b5","updated":"2019-07-24 15:26:48.000000000","message":"Done","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"05017ccd14fbcee970cd2af3578ee85bbf741c95","unresolved":false,"context_lines":[{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"},{"line_number":15,"context_line":"need to fix things like failing to claim unrelated root disk space on"},{"line_number":16,"context_line":"compute nodes and reporting volumes as used disk space on compute nodes."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"In order to do that, we need to pull out a lot of the handling of these"},{"line_number":19,"context_line":"basic three resources from the compute node\u0027s resource tracker. This patch"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_c439b6f2","line":16,"range":{"start_line":14,"start_character":0,"end_line":16,"end_character":72},"updated":"2019-07-23 21:07:08.000000000","message":"I think this is old isn\u0027t it? With the work in Rocky for fixing volume-backed disk resource tracking right?\n\nhttps://review.opendev.org/#/q/topic:bug/1469179+(status:open+OR+status:merged)\n\nOr is this talking about something else I\u0027m not understanding? Why would report volumes as used disk space on compute nodes? Do you mean if the volumes are on the local compute node but from cinder reporting that usage?","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"ed547ca2fe46ea60eadf6bbfd756b020cc4f211d","unresolved":false,"context_lines":[{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"},{"line_number":15,"context_line":"need to fix things like failing to claim unrelated root disk space on"},{"line_number":16,"context_line":"compute nodes and reporting volumes as used disk space on compute nodes."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"In order to do that, we need to pull out a lot of the handling of these"},{"line_number":19,"context_line":"basic three resources from the compute node\u0027s resource tracker. This patch"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_32b79edb","line":16,"range":{"start_line":14,"start_character":0,"end_line":16,"end_character":72},"in_reply_to":"7faddb67_713c43bf","updated":"2019-07-24 15:26:48.000000000","message":"Removed all of this to simply state this is a no-op now","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"0d29be6385851807eef9ad4bfa11a8bef73174eb","unresolved":false,"context_lines":[{"line_number":11,"context_line":"to the resource tracker, and thus thus treating everything in the claim"},{"line_number":12,"context_line":"process as unlimited."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"As a step towards making BFV instances first class citizens in Nova, we"},{"line_number":15,"context_line":"need to fix things like failing to claim unrelated root disk space on"},{"line_number":16,"context_line":"compute nodes and reporting volumes as used disk space on compute nodes."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"In order to do that, we need to pull out a lot of the handling of these"},{"line_number":19,"context_line":"basic three resources from the compute node\u0027s resource tracker. This patch"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"7faddb67_713c43bf","line":16,"range":{"start_line":14,"start_character":0,"end_line":16,"end_character":72},"in_reply_to":"7faddb67_c439b6f2","updated":"2019-07-23 22:03:18.000000000","message":"Yep, this is just what I had in my head when I was writing this knowing we were a year from being able to merge it.","commit_id":"691239a52f807adf2de9c9042d5faa11a961f7a1"}],"nova/compute/claims.py":[{"author":{"_account_id":5754,"name":"Alex Xu","email":"hejie.xu@intel.com","username":"xuhj"},"change_message_id":"c780143b26d0165d5d37a995db722bc5efee3cd2","unresolved":false,"context_lines":[{"line_number":158,"context_line":"            elif instance_topology:"},{"line_number":159,"context_line":"                self.claimed_numa_topology \u003d instance_topology"},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"    def _test(self, type_, unit, total, used, requested, limit):"},{"line_number":162,"context_line":"        \"\"\"Test if the given type of resource needed for a claim can be safely"},{"line_number":163,"context_line":"        allocated."},{"line_number":164,"context_line":"        \"\"\""}],"source_content_type":"text/x-python","patch_set":5,"id":"7faddb67_f167bab6","line":161,"range":{"start_line":161,"start_character":4,"end_line":161,"end_character":64},"updated":"2019-07-25 07:12:23.000000000","message":"this can be removed also","commit_id":"5406c8bd9b8740a27c60a0ac7983c84e440f0d35"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"cde4f5dd2982bc099ed0d90bf7b22225cee4b4df","unresolved":false,"context_lines":[{"line_number":158,"context_line":"            elif instance_topology:"},{"line_number":159,"context_line":"                self.claimed_numa_topology \u003d instance_topology"},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"    def _test(self, type_, unit, total, used, requested, limit):"},{"line_number":162,"context_line":"        \"\"\"Test if the given type of resource needed for a claim can be safely"},{"line_number":163,"context_line":"        allocated."},{"line_number":164,"context_line":"        \"\"\""}],"source_content_type":"text/x-python","patch_set":5,"id":"7faddb67_1494c443","line":161,"range":{"start_line":161,"start_character":4,"end_line":161,"end_character":64},"in_reply_to":"7faddb67_f167bab6","updated":"2019-07-25 09:00:50.000000000","message":"Can I do this in a follow-up if there\u0027s nothing else to change here, so I don\u0027t need to rebase the whole series?","commit_id":"5406c8bd9b8740a27c60a0ac7983c84e440f0d35"}],"releasenotes/notes/remove-cpu-ram-disk-claims-4f3d28fa3332a426.yaml":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"a6cee55c62d0b9c5b4a61840722474089311e099","unresolved":false,"context_lines":[{"line_number":4,"context_line":"    The resource tracker no longer attempts to claim vCPUs, RAM, or disk."},{"line_number":5,"context_line":"    Accounting for these resources has been the responsibility of the placement"},{"line_number":6,"context_line":"    service for a number of releases now, meaning this claiming has been either"},{"line_number":7,"context_line":"    unnecessary or, in cases like BFV instances, actively harmful."},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"    One side-effect of this change is that it\u0027s no longer possible for virt"},{"line_number":10,"context_line":"    drivers to report overhead on a per-instance basis. This affects the"}],"source_content_type":"text/x-yaml","patch_set":2,"id":"7faddb67_b7e2c79f","line":7,"range":{"start_line":7,"start_character":34,"end_line":7,"end_character":37},"updated":"2019-07-22 17:02:15.000000000","message":"say \"volume-backed\" here, BFV is more or less internal nova dev speak","commit_id":"7ffddf1de8755471eefdab7caf536b334ce4afe8"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"bcb51c5314769e6dc9e3cafc5539bbacaaaf3307","unresolved":false,"context_lines":[{"line_number":6,"context_line":"    service for a number of releases now, meaning this claiming has been either"},{"line_number":7,"context_line":"    unnecessary or, in cases like BFV instances, actively harmful."},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"    One side-effect of this change is that it\u0027s no longer possible for virt"},{"line_number":10,"context_line":"    drivers to report overhead on a per-instance basis. This affects the"},{"line_number":11,"context_line":"    libvirt (``libvirt.LibvirtDriver``), HyperV (``hyperv.HyperVDriver``)"},{"line_number":12,"context_line":"    and XenAPI (``xenapi.XenAPIDriver``) in-tree drivers, and may affect"}],"source_content_type":"text/x-yaml","patch_set":2,"id":"7faddb67_f7535f6b","line":9,"range":{"start_line":9,"start_character":4,"end_line":9,"end_character":34},"updated":"2019-07-22 16:58:45.000000000","message":"Well, technically people should have had these filters disabled for several cycles now, so this change shouldn\u0027t really affect anyone as the overhead wasn\u0027t being factored into the claim or scheduling decision anyway yeah? If the filters are disabled, there will be no limit in the limits dict, which means the RT won\u0027t do claiming for that resource.","commit_id":"7ffddf1de8755471eefdab7caf536b334ce4afe8"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"7a09752ee3e17c359d37ce7e66f59744c77dbb4e","unresolved":false,"context_lines":[{"line_number":6,"context_line":"    service for a number of releases now, meaning this claiming has been either"},{"line_number":7,"context_line":"    unnecessary or, in cases like BFV instances, actively harmful."},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"    One side-effect of this change is that it\u0027s no longer possible for virt"},{"line_number":10,"context_line":"    drivers to report overhead on a per-instance basis. This affects the"},{"line_number":11,"context_line":"    libvirt (``libvirt.LibvirtDriver``), HyperV (``hyperv.HyperVDriver``)"},{"line_number":12,"context_line":"    and XenAPI (``xenapi.XenAPIDriver``) in-tree drivers, and may affect"}],"source_content_type":"text/x-yaml","patch_set":2,"id":"7faddb67_11a71d52","line":9,"range":{"start_line":9,"start_character":4,"end_line":9,"end_character":34},"in_reply_to":"7faddb67_32b4a5f2","updated":"2019-07-23 10:01:11.000000000","message":"I\u0027ve rebased this patch onto the one that removes the CPU, Ram and Disk filters, and copied this into the release note for that change. I\u0027ve also added docs to that change, as above.","commit_id":"7ffddf1de8755471eefdab7caf536b334ce4afe8"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"bebd6d2b4a260207dce7d4d98e9aad3a6a045196","unresolved":false,"context_lines":[{"line_number":6,"context_line":"    service for a number of releases now, meaning this claiming has been either"},{"line_number":7,"context_line":"    unnecessary or, in cases like BFV instances, actively harmful."},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"    One side-effect of this change is that it\u0027s no longer possible for virt"},{"line_number":10,"context_line":"    drivers to report overhead on a per-instance basis. This affects the"},{"line_number":11,"context_line":"    libvirt (``libvirt.LibvirtDriver``), HyperV (``hyperv.HyperVDriver``)"},{"line_number":12,"context_line":"    and XenAPI (``xenapi.XenAPIDriver``) in-tree drivers, and may affect"}],"source_content_type":"text/x-yaml","patch_set":2,"id":"7faddb67_32b4a5f2","line":9,"range":{"start_line":9,"start_character":4,"end_line":9,"end_character":34},"in_reply_to":"7faddb67_f7535f6b","updated":"2019-07-22 17:46:14.000000000","message":"Yeah good point:\n\nhttps://github.com/openstack/nova/blob/d5c67a3d954ddb571645886a23a0f251ae7dd2bb/nova/scheduler/filters/core_filter.py#L53\n\nhttps://github.com/openstack/nova/blob/d5c67a3d954ddb571645886a23a0f251ae7dd2bb/nova/compute/claims.py#L150\n\nhttps://github.com/openstack/nova/blob/d5c67a3d954ddb571645886a23a0f251ae7dd2bb/nova/compute/claims.py#L198\n\nhttps://github.com/openstack/nova/blob/d5c67a3d954ddb571645886a23a0f251ae7dd2bb/nova/compute/claims.py#L246\n\nGiven that, people haven\u0027t been accounting / claiming with overhead for a long time now (probably since at least Pike when we no longer needed those filters but didn\u0027t yet make them deprecated until Stein).\n\nThe information below should probably live in admin docs somewhere, like either per-hypervisor in https://docs.openstack.org/nova/latest/admin/configuration/hypervisors.html or something in https://docs.openstack.org/nova/latest/admin/adv-config.html or https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#allocation-ratios about over committing resources and considerations for reserving room on the host for specific hypervisor drivers.\n\nThe release note is OK but long-term we should have that information in the admin docs as well.","commit_id":"7ffddf1de8755471eefdab7caf536b334ce4afe8"}]}