)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"2b9e661a020bde7cfb1d21960468b005d163c6a3","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"libvirt: Don\u0027t allow \"reserving\" file-backed memory"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"When file-backed memory is configured, it is the only \"memory\" reported"},{"line_number":10,"context_line":"by nova and used by instances, with RAM used in caching capacity. We"},{"line_number":11,"context_line":"should be warning users of this and insisting they explicitly configure"},{"line_number":12,"context_line":"the \u0027[DEFAULT] reserved_host_memory_mb\u0027 config option to 0. However,"},{"line_number":13,"context_line":"doing so now would be breaking change. Instead, start logging a warning"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":3,"id":"ff570b3c_587ccb71","line":10,"range":{"start_line":9,"start_character":39,"end_line":10,"end_character":7},"updated":"2020-06-10 10:36:32.000000000","message":"so as we discussed on irc we need to validate that hte libvirt driver is not reporting any hugepages that are avilabel on the platform in the resouce track.\n\nif we are then the numa toplogy filter could allow this host to be selected resulting in no valid hosts after we hit the retry limit since we dont support hugepage + file backed memory on the same host.\n\nthe reason file baced memroy  was added to libvirt was to allow vhost-user to be used without hugepages\n\nhttps://www.redhat.com/archives/libvir-list/2016-May/msg00689.html\n\n\nso in that configuration hugpeags would be allocated and used by ovs but the vms would not use hugepages allow for live migration without the numa issuse we used to have.\n\nso its not unresoble for the host to have hugepages\n\nyou could use https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.reserved_huge_pages\n\ni guess to also set it to 0 but its getting kind of excessive.","commit_id":"d12114634e9e11c07c62c0c45dce5e4e5e24ca80"}],"doc/source/admin/file-backed-memory.rst":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"2b9e661a020bde7cfb1d21960468b005d163c6a3","unresolved":false,"context_lines":[{"line_number":46,"context_line":"   capability requires libvirt version 4.4.0 or newer."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Qemu"},{"line_number":49,"context_line":"   File-backed memory requires qemu version 2.6.0 or newer. Discard capability"},{"line_number":50,"context_line":"   requires qemu version 2.10.0 or newer."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Memory overcommit"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ff570b3c_ad2207b9","line":49,"range":{"start_line":49,"start_character":59,"end_line":49,"end_character":60},"updated":"2020-06-10 10:36:32.000000000","message":"nit unrealated but i proably would fix it too","commit_id":"d12114634e9e11c07c62c0c45dce5e4e5e24ca80"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"2b9e661a020bde7cfb1d21960468b005d163c6a3","unresolved":false,"context_lines":[{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Reserved memory"},{"line_number":60,"context_line":"   When configured, file-backed memory is reported as total system memory to"},{"line_number":61,"context_line":"   placement, with RAM used as cache. Reserved memory is corresponds to disk"},{"line_number":62,"context_line":"   space not set aside for file-backed memory."},{"line_number":63,"context_line":"   :oslo.config:option:`reserved_host_memory_mb` should be set to ``0`` in"},{"line_number":64,"context_line":"   ``nova.conf``."}],"source_content_type":"text/x-rst","patch_set":3,"id":"ff570b3c_f8015f06","line":61,"range":{"start_line":61,"start_character":54,"end_line":61,"end_character":57},"updated":"2020-06-10 10:36:32.000000000","message":"delete","commit_id":"d12114634e9e11c07c62c0c45dce5e4e5e24ca80"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"2b9e661a020bde7cfb1d21960468b005d163c6a3","unresolved":false,"context_lines":[{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Reserved memory"},{"line_number":60,"context_line":"   When configured, file-backed memory is reported as total system memory to"},{"line_number":61,"context_line":"   placement, with RAM used as cache. Reserved memory is corresponds to disk"},{"line_number":62,"context_line":"   space not set aside for file-backed memory."},{"line_number":63,"context_line":"   :oslo.config:option:`reserved_host_memory_mb` should be set to ``0`` in"},{"line_number":64,"context_line":"   ``nova.conf``."},{"line_number":65,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ff570b3c_184b735f","line":62,"range":{"start_line":61,"start_character":72,"end_line":62,"end_character":8},"updated":"2020-06-10 10:36:32.000000000","message":"hum this raise another point. are we subtracting the file backed memroy size form the host local_GB ?\n\nif the mount point for the file backed memroy is on the same device as the  nova instance dir then both will come form the same overall capasity.\n\nso we shoudl be adjusting the placmenet reservation to account for that too right.\n\nor have the operator add the file backed moery to the reserved_host_disk_gb parmater or whatever its called.","commit_id":"d12114634e9e11c07c62c0c45dce5e4e5e24ca80"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"4bca42e26817ac02e83f2ecac1c8b75bc50da344","unresolved":false,"context_lines":[{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Memory overcommit"},{"line_number":53,"context_line":"   File-backed memory is not compatible with memory overcommit."},{"line_number":54,"context_line":"   :oslo.config:option:`ram_allocation_ratio` must be set to ``1.0`` in"},{"line_number":55,"context_line":"   ``nova.conf``, and the host must not be added to a :doc:`host aggregate"},{"line_number":56,"context_line":"   \u003c/admin/aggregates\u003e` with ``ram_allocation_ratio`` set to anything but"},{"line_number":57,"context_line":"   ``1.0``."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_f34fc067","line":54,"range":{"start_line":54,"start_character":24,"end_line":54,"end_character":44},"updated":"2020-06-10 11:39:32.000000000","message":"this is in the default section so maybe this works un qualified but i was expected to see a section referenced.\n\nin any case https://196ae2c512459368d998-da6d66fdec213ff875550bc735254b25.ssl.cf1.rackcdn.com/734658/1/check/openstack-tox-docs/7e7b2e4/docs/admin/file-backed-memory.html was the last time zuul completed running on this change so im going to hold my +1 until the docs job runs and assert it worked corredtly","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"97dc8eed6eab5fe114e95d553761128a107af974","unresolved":false,"context_lines":[{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Memory overcommit"},{"line_number":53,"context_line":"   File-backed memory is not compatible with memory overcommit."},{"line_number":54,"context_line":"   :oslo.config:option:`ram_allocation_ratio` must be set to ``1.0`` in"},{"line_number":55,"context_line":"   ``nova.conf``, and the host must not be added to a :doc:`host aggregate"},{"line_number":56,"context_line":"   \u003c/admin/aggregates\u003e` with ``ram_allocation_ratio`` set to anything but"},{"line_number":57,"context_line":"   ``1.0``."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_f97b21f1","line":54,"range":{"start_line":54,"start_character":24,"end_line":54,"end_character":44},"in_reply_to":"ff570b3c_f34fc067","updated":"2020-06-10 13:43:03.000000000","message":"this rednders correctly and links properly too\n\nhttps://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_dd6/734658/4/check/openstack-tox-docs/dd6b3ed/docs/admin/file-backed-memory.html#prerequisites-and-limitations","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"4bca42e26817ac02e83f2ecac1c8b75bc50da344","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_536e6cc9","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"updated":"2020-06-10 11:39:32.000000000","message":"as noted in http://paste.openstack.org/show/794561/\nwhen we enable file back memory we dont report the avilable hugepages on the host so we dont actully need to do this the numa toplogy filter wont select the hsot anyway.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"2d596cf1ef5fed9ac62a80f4ba88d3e87de82010","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_eb65041f","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_19ba55e0","updated":"2020-06-10 14:25:16.000000000","message":"so booting a vm with 1 or 2 numa nodes explictly work correctly\n\nhttp://paste.openstack.org/show/794573/\n\nthe memroy backing is correct\n\n  \u003cmemoryBacking\u003e\n    \u003csource type\u003d\u0027file\u0027/\u003e\n    \u003caccess mode\u003d\u0027shared\u0027/\u003e\n    \u003callocation mode\u003d\u0027immediate\u0027/\u003e\n    \u003cdiscard/\u003e\n  \u003c/memoryBacking\u003e\n\nand we generate two backing files one per virtual numa node and they are correctly affinitesed to the host numa nodes too\n\n-object memory-backend-file,id\u003dram-node0,mem-path\u003d/var/lib/libvirt/qemu/ram/libvirt/qemu/2-instance-00000003/ram-node0,discard-data\u003dyes,share\u003dyes,size\u003d268435456,host-nodes\u003d0,policy\u003dbind -numa node,nodeid\u003d0,cpus\u003d0,memdev\u003dram-node0 \n\n-object memory-backend-file,id\u003dram-node1,mem-path\u003d/var/lib/libvirt/qemu/ram/libvirt/qemu/2-instance-00000003/ram-node1,discard-data\u003dyes,share\u003dyes,size\u003d268435456,host-nodes\u003d1,policy\u003dbind -numa node,nodeid\u003d1,cpus\u003d1,memdev\u003dram-node1 \n\nnodeid is the virtual guest numa node and host-nodes is obvious.\n\nso file backed memroy and numa seams to work \n\nif i create a flavor that ask form more numa nodes then the hsot supports e.g. 4 then the numa toplogy filter rejects it\n\nso i think numa and numa affinity which is based on the cpu to pci affinity should work.\n\n\n\nif we use hw:mem_page_size\u003dsmall\n\nwe pass the numa toplogy filter which is operating on the info in the numa toplogy blob and then fail with\n\nnova.exception.MemoryPagesUnsupported\n\nin the virt driver\n\nJun 10 14:10:28 numa-2 nova-compute[113097]: DEBUG nova.compute.manager [None req-349e3d95-7dca-430e-9e15-08bf98899cbd demo admin] [instance: 783f1934-5161-4ef5-aee0-4bf8c6c46d04] Build of instance 783f1934-5161-4ef5-aee0-4bf8c6c46d04 was re-scheduled: Host does not support guests with custom memory page sizes {{(pid\u003d113097) _do_build_and_run_instance /opt/stack/nova/nova/compute/manager.py:2176}}\n\nwhich lead to a retry and then a no valid host because i only have one enabled currently.\n\n\ni get the same with hw:mem_page_size\u003dany\n\nthat in theory should work but we never actully fully implemnted that so currenntly any means do what the image says.\n\nhw:mem_page_size\u003dlarge also fails with the same exception\nso if you set a page size it will fail late on the compute node.\n\n\nlooking at placment\n\nubuntu@numa-1:/opt/repos/devstack$ openstack resource provider inventory list f5448faf-4224-429a-b2b1-3c7edff3ee9c\n+----------------+------------------+----------+----------+----------+-----------+-------+\n| resource_class | allocation_ratio | min_unit | max_unit | reserved | step_size | total |\n+----------------+------------------+----------+----------+----------+-----------+-------+\n| VCPU           |             16.0 |        1 |       12 |        0 |         1 |    12 |\n| MEMORY_MB      |              1.0 |        1 |     4096 |        0 |         1 |  4096 |\n| DISK_GB        |              1.0 |        1 |       19 |        0 |         1 |    19 |\n+----------------+------------------+----------+----------+----------+-----------+-------+\n\n\nwe are correctly reporting the availabel file backed memory 4G in this case but we are not reserving it in DISK_GB automatically.\n\nas i said in other comment that is only correct to do if the mount point is on the same filesystem as the instance state dire and if you are using local storage but we should document that in a follow up path so people can set teh correct value in reserved_host_disk_mb\n\nassuming we dont automate that.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"7bc1a1d402c3813222df8938bfd7eefc56645709","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_f3918048","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_536e6cc9","updated":"2020-06-10 12:08:49.000000000","message":"ignore this we are incorrectly reporting hugepages i had not restarted libvirt","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"4a29b3803ebd5872ea99ac37c22d5c3aae24fefd","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"bf51134e_2daa74d3","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_b6b9a87a","updated":"2020-06-17 10:55:20.000000000","message":"All good then","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":5754,"name":"Alex Xu","email":"hejie.xu@intel.com","username":"xuhj"},"change_message_id":"2b621874294927645206cb76fd22dc8e063975fa","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_b6b9a87a","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_eb65041f","updated":"2020-06-11 01:01:04.000000000","message":"@Sean, thanks for test. Yea, agree with disk reserve. Also I think NUMA topo filter is still using the real ram as the host numa node memory. And we consume the real ram in the host numa topo. I didn\u0027t see we report the file_back_memory resize to the host numa topo.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":5754,"name":"Alex Xu","email":"hejie.xu@intel.com","username":"xuhj"},"change_message_id":"854b0e494a33c5cb2733fdde31105a84aeceed5a","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_fe8ca736","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_f3918048","updated":"2020-06-10 12:55:02.000000000","message":"If the instance boot with NUMA topo and normal page, what will happened? I guess NUMA topology filter may pass the scheduling, but it will fail at virt driver.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"7ae1b774291ea51c78ba1db0e65a3d94963df8e9","unresolved":false,"context_lines":[{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Huge pages"},{"line_number":67,"context_line":"   File-backed memory is not compatible with huge pages. Instances with huge"},{"line_number":68,"context_line":"   pages configured will not start on a host with file-backed memory enabled. It"},{"line_number":69,"context_line":"   is recommended to use host aggregates to ensure instances configured for"},{"line_number":70,"context_line":"   huge pages are not placed on hosts with file-backed memory configured."},{"line_number":71,"context_line":""},{"line_number":72,"context_line":"Handling these limitations could be optimized with a scheduler filter in the"},{"line_number":73,"context_line":"future."}],"source_content_type":"text/x-rst","patch_set":4,"id":"ff570b3c_19ba55e0","line":70,"range":{"start_line":68,"start_character":77,"end_line":70,"end_character":73},"in_reply_to":"ff570b3c_fe8ca736","updated":"2020-06-10 13:41:37.000000000","message":"i will test that now and let you know.\n\nhugepages is not ment too work.\n\nnuma might work although it should just be looking at the cpu and pci divice iguess and not the guest memroy.\n\ni suspect that was not tought about properly the spec was pretty light on detail.\n\nhttps://specs.openstack.org/openstack/nova-specs/specs/rocky/implemented/libvirt-file-backed-memory.html\n\ni kind of wish i had been involved in the review\nhttps://review.opendev.org/#/c/563704/\nit merged when i was getting ready to leave Intel and move to redhat so i was not really following upstream at the time but this should all have been address in the sepc.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"}],"nova/tests/unit/virt/libvirt/test_driver.py":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"4a29b3803ebd5872ea99ac37c22d5c3aae24fefd","unresolved":false,"context_lines":[{"line_number":1352,"context_line":"        drvr \u003d libvirt_driver.LibvirtDriver(fake.FakeVirtAPI(), True)"},{"line_number":1353,"context_line":"        self.assertRaises("},{"line_number":1354,"context_line":"            exception.InternalError, drvr._check_file_backed_memory_support,"},{"line_number":1355,"context_line":"        )"},{"line_number":1356,"context_line":""},{"line_number":1357,"context_line":"    @mock.patch.object(libvirt_driver.LOG, \u0027warning\u0027)"},{"line_number":1358,"context_line":"    def test__check_file_backed_memory_support__has_reserved(self, mock_log):"}],"source_content_type":"text/x-python","patch_set":4,"id":"bf51134e_adb5e433","line":1355,"updated":"2020-06-17 10:55:20.000000000","message":"we already went to an error, so this is fine.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"}],"nova/virt/libvirt/driver.py":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"efca8ad245f76a52b8e97061f16698479790ba96","unresolved":false,"context_lines":[{"line_number":848,"context_line":"                    \"reserved_host_memory_mb\u0027\""},{"line_number":849,"context_line":"                )"},{"line_number":850,"context_line":"                raise exception.InternalError(msg)"},{"line_number":851,"context_line":""},{"line_number":852,"context_line":"            # TODO(stephenfin): Change this to an exception in W or later"},{"line_number":853,"context_line":"            LOG.warning("},{"line_number":854,"context_line":"                \"Reserving memory via \u0027[DEFAULT] reserved_host_memory_mb\u0027 \""}],"source_content_type":"text/x-python","patch_set":1,"id":"ff570b3c_5d1cc8c0","line":851,"updated":"2020-06-09 18:01:14.000000000","message":"so here you need to set CONF.reserved_host_memory_mb\u003d0\n\nalthough we proably shoudl not raise the internal erroor and jsut do this on line 843 and delete the if above","commit_id":"07c4fa84301e6f407c3acdac397eb7b5f9dbc81f"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"a0516b3c26bc9a5030dbb96678694ea618e3b685","unresolved":false,"context_lines":[{"line_number":848,"context_line":"                    \"reserved_host_memory_mb\u0027\""},{"line_number":849,"context_line":"                )"},{"line_number":850,"context_line":"                raise exception.InternalError(msg)"},{"line_number":851,"context_line":""},{"line_number":852,"context_line":"            # TODO(stephenfin): Change this to an exception in W or later"},{"line_number":853,"context_line":"            LOG.warning("},{"line_number":854,"context_line":"                \"Reserving memory via \u0027[DEFAULT] reserved_host_memory_mb\u0027 \""}],"source_content_type":"text/x-python","patch_set":1,"id":"ff570b3c_ea57c525","line":851,"in_reply_to":"ff570b3c_5d1cc8c0","updated":"2020-06-10 09:21:00.000000000","message":"I don\u0027t think I want to override the value behind their back. Even though it\u0027s a nonsensical decision, I\u0027d still rather we report the configured reserved value until the user configures the thing","commit_id":"07c4fa84301e6f407c3acdac397eb7b5f9dbc81f"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"efca8ad245f76a52b8e97061f16698479790ba96","unresolved":false,"context_lines":[{"line_number":852,"context_line":"            # TODO(stephenfin): Change this to an exception in W or later"},{"line_number":853,"context_line":"            LOG.warning("},{"line_number":854,"context_line":"                \"Reserving memory via \u0027[DEFAULT] reserved_host_memory_mb\u0027 \""},{"line_number":855,"context_line":"                \"is not compatible with file-backed memory. Consider \""},{"line_number":856,"context_line":"                \"unsetting \u0027[DEFAULT] reserved_host_memory_mb\u0027. This will be \""},{"line_number":857,"context_line":"                \"an error in a future release.\""},{"line_number":858,"context_line":"            )"},{"line_number":859,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"ff570b3c_dd37384d","line":856,"range":{"start_line":855,"start_character":60,"end_line":856,"end_character":62},"updated":"2020-06-09 18:01:14.000000000","message":"the default is 512 so you will get this error if you dont have it in your config. you need to set it to 0 expclitly.","commit_id":"07c4fa84301e6f407c3acdac397eb7b5f9dbc81f"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"c8ed3572f32d476edee71b33e7814f579c8eddce","unresolved":false,"context_lines":[{"line_number":852,"context_line":"            # TODO(stephenfin): Change this to an exception in W or later"},{"line_number":853,"context_line":"            LOG.warning("},{"line_number":854,"context_line":"                \"Reserving memory via \u0027[DEFAULT] reserved_host_memory_mb\u0027 \""},{"line_number":855,"context_line":"                \"is not compatible with file-backed memory. Consider \""},{"line_number":856,"context_line":"                \"unsetting \u0027[DEFAULT] reserved_host_memory_mb\u0027. This will be \""},{"line_number":857,"context_line":"                \"an error in a future release.\""},{"line_number":858,"context_line":"            )"},{"line_number":859,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"ff570b3c_6d3f4f87","line":856,"range":{"start_line":855,"start_character":60,"end_line":856,"end_character":62},"in_reply_to":"ff570b3c_0a4f9992","updated":"2020-06-10 09:38:14.000000000","message":"the default being 512 is why i want to override it.\n\nif the default was 0 then i would aggree they should remove it from there conf but this might be reporting a warning because of our default value.","commit_id":"07c4fa84301e6f407c3acdac397eb7b5f9dbc81f"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"a0516b3c26bc9a5030dbb96678694ea618e3b685","unresolved":false,"context_lines":[{"line_number":852,"context_line":"            # TODO(stephenfin): Change this to an exception in W or later"},{"line_number":853,"context_line":"            LOG.warning("},{"line_number":854,"context_line":"                \"Reserving memory via \u0027[DEFAULT] reserved_host_memory_mb\u0027 \""},{"line_number":855,"context_line":"                \"is not compatible with file-backed memory. Consider \""},{"line_number":856,"context_line":"                \"unsetting \u0027[DEFAULT] reserved_host_memory_mb\u0027. This will be \""},{"line_number":857,"context_line":"                \"an error in a future release.\""},{"line_number":858,"context_line":"            )"},{"line_number":859,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"ff570b3c_0a4f9992","line":856,"range":{"start_line":855,"start_character":60,"end_line":856,"end_character":62},"in_reply_to":"ff570b3c_dd37384d","updated":"2020-06-10 09:21:00.000000000","message":"Oh, good call","commit_id":"07c4fa84301e6f407c3acdac397eb7b5f9dbc81f"}],"releasenotes/notes/bug-1882821-file-backed-memory-reserved-conflict-3ad4c04ab993ebf8.yaml":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"4a29b3803ebd5872ea99ac37c22d5c3aae24fefd","unresolved":false,"context_lines":[{"line_number":1,"context_line":"---"},{"line_number":2,"context_line":"upgrade:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    When using file-backed memory, the ``nova-compute`` service will now fail"},{"line_number":5,"context_line":"    to start if the amount of reserved memory configured using ``[DEFAULT]"},{"line_number":6,"context_line":"    reserved_host_memory_mb`` is equal to or greater than the total amount of"},{"line_number":7,"context_line":"    memory configured using ``[libvirt] file_backed_memory``. Where reserved"}],"source_content_type":"text/x-yaml","patch_set":4,"id":"bf51134e_cde65845","line":4,"range":{"start_line":4,"start_character":36,"end_line":4,"end_character":77},"updated":"2020-06-17 10:55:20.000000000","message":"since it looks silly to reserve more RAM than what you have, I\u0027m OK with hard-stopping when upgrading from Ussuri.","commit_id":"3b99747b42e460567d52cc6d748396cb349d56b8"}]}