Back to bug 1561961
| Who | When | What | Removed | Added |
|---|---|---|---|---|
| Red Hat Bugzilla Rules Engine | 2018-03-29 09:19:39 UTC | Keywords | ZStream | |
| Target Release | 13.0 (Queens) | --- | ||
| Lee Yarwood | 2018-05-09 10:29:51 UTC | Status | ASSIGNED | POST |
| Lee Yarwood | 2018-07-02 09:09:21 UTC | Status | POST | ASSIGNED |
| Wei Yongjun | 2018-07-11 11:18:36 UTC | CC | weiyongjun | |
| Stephen Finucane | 2018-07-19 10:35:27 UTC | Doc Text | Feature: Added support for PCI device NUMA affinity policies. These are configured as part of the ``[pci]alias`` configuration option(s). There are three policies supported: - ``required`` (must-have) - ``legacy`` (must-have-if-available) (default) - ``preferred`` (nice-to-have) In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity one is willing to forsake before failing to schedule. Reason: These allow you to configure how strict your NUMA affinity should be on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. Result: When the 'preferred' policy is configured for a PCI device, nova will now utilize CPUs on a different NUMA node from the NUMA node of the PCI device if this is all that is available. This will result in increased resource utilization with the downside of reduced performance for these instances. |
|
| Lee Yarwood | 2018-07-19 10:39:06 UTC | Status | ASSIGNED | MODIFIED |
| Target Release | --- | 13.0 (Queens) | ||
| Fixed In Version | openstack-nova-17.0.1-0.20180302144923.9ace6ed.el7ost | |||
| Assignee | lyarwood | sfinucan | ||
| Target Milestone | async | z2 | ||
| Red Hat Bugzilla Rules Engine | 2018-07-19 10:39:17 UTC | Target Release | 13.0 (Queens) | --- |
| Lon Hohberger | 2018-07-20 10:34:31 UTC | Status | MODIFIED | ON_QA |
| Jon Schlueter | 2018-08-20 13:47:51 UTC | Flags | needinfo?(sfinucan) | |
| Joanne O'Flynn | 2018-08-20 20:42:23 UTC | CC | joflynn | |
| Doc Text | Feature: Added support for PCI device NUMA affinity policies. These are configured as part of the ``[pci]alias`` configuration option(s). There are three policies supported: - ``required`` (must-have) - ``legacy`` (must-have-if-available) (default) - ``preferred`` (nice-to-have) In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity one is willing to forsake before failing to schedule. Reason: These allow you to configure how strict your NUMA affinity should be on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. Result: When the 'preferred' policy is configured for a PCI device, nova will now utilize CPUs on a different NUMA node from the NUMA node of the PCI device if this is all that is available. This will result in increased resource utilization with the downside of reduced performance for these instances. | This feature adds support for PCI device NUMA affinity policies. These are configured as part of the `[pci]alias` configuration options. There are three policies supported: - `required` - `legacy` - `preferred` In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity you can forsake before failing to schedule. These policies allow you to configure how strict your NUMA affinity is on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. When the 'preferred' policy is configured for a PCI device, nova now utilizes CPUs on a different NUMA node from the NUMA node of the PCI device if this is all that is available. This results in increased resource utilization with the downside of reduced performance for these instances. |
||
| Lon Hohberger | 2018-08-24 18:09:06 UTC | CC | rhos-flags | |
| Flags | needinfo?(sfinucan) | needinfo?(rhos-flags) | ||
| Scott Lewis | 2018-10-18 18:32:57 UTC | Target Milestone | z2 | z3 |
| Pavan Chavva | 2018-10-18 19:18:05 UTC | CC | pchavva | |
| Flags | needinfo?(lyarwood) | |||
| Erwan Gallen | 2018-11-02 10:56:23 UTC | CC | egallen | |
| PnT Account Manager | 2018-11-02 22:06:51 UTC | CC | sferdjao | |
| Derek Li | 2018-11-07 16:44:03 UTC | CC | derli | |
| Stephen Finucane | 2018-11-07 17:16:49 UTC | Flags | needinfo?(rhos-flags) needinfo?(lyarwood) | |
| Stephen Finucane | 2018-11-07 17:22:10 UTC | Depends On | 1647536 | |
| Stephen Finucane | 2018-11-07 17:25:44 UTC | Summary | [RFE] Optional NUMA affinity for SR-IOV devices | [RFE] Optional NUMA affinity for PCI devices |
| Lee Yarwood | 2018-11-16 16:00:58 UTC | Summary | [RFE] Optional NUMA affinity for PCI devices | [TestOnly][RFE] Optional NUMA affinity for PCI devices |
| Lee Yarwood | 2018-11-16 16:01:59 UTC | Summary | [TestOnly][RFE] Optional NUMA affinity for PCI devices | [TestOnly][13][RFE] Optional NUMA affinity for PCI devices |
| Lee Yarwood | 2018-11-16 16:02:45 UTC | Blocks | 1650606 | |
| Jon Schlueter | 2018-11-19 06:12:53 UTC | CC | jschluet | |
| Target Milestone | z3 | z4 | ||
| Lee Yarwood | 2018-12-05 13:16:35 UTC | Status | ON_QA | ON_DEV |
| Fixed In Version | openstack-nova-17.0.1-0.20180302144923.9ace6ed.el7ost | |||
| Target Milestone | z4 | async | ||
| Lee Yarwood | 2018-12-05 16:36:43 UTC | Keywords | FutureFeature, TestOnly | |
| Summary | [TestOnly][13][RFE] Optional NUMA affinity for PCI devices | [13] Optional NUMA affinity for PCI devices | ||
| Red Hat Bugzilla Rules Engine | 2018-12-05 16:36:59 UTC | Target Release | --- | 13.0 (Queens) |
| David Juran | 2019-03-11 15:54:43 UTC | CC | djuran | |
| Daniel Berrangé | 2019-04-11 13:42:58 UTC | CC | berrange | |
| PnT Account Manager | 2019-04-22 21:40:08 UTC | QA Contact | awaugama | nova-maint |
| PnT Account Manager | 2019-04-22 21:40:18 UTC | CC | awaugama | |
| Matthew Booth | 2019-07-03 15:27:28 UTC | Keywords | FutureFeature | |
| CC | mbooth | |||
| Matthew Booth | 2019-07-03 15:28:05 UTC | Keywords | Patch | |
| Russell Bryant | 2019-07-10 13:29:23 UTC | CC | rbryant | |
| Lee Yarwood | 2019-08-02 11:41:48 UTC | Target Milestone | async | z8 |
| Archit Modi | 2019-08-13 14:52:18 UTC | CC | amodi | |
| QA Contact | nova-maint | smooney | ||
| Archit Modi | 2019-08-13 14:59:02 UTC | QA Contact | smooney | lyarwood |
| Lee Yarwood | 2019-08-27 11:11:09 UTC | Target Milestone | z8 | z9 |
| Matthew Booth | 2019-09-04 13:45:17 UTC | CC | vromanso | |
| Stephen Finucane | 2019-09-30 14:40:55 UTC | Status | ON_DEV | POST |
| Stephen Finucane | 2019-09-30 15:00:11 UTC | Status | POST | MODIFIED |
| Fixed In Version | openstack-nova-17.0.10-9.el7ost | |||
| Joanne O'Flynn | 2019-09-30 15:05:35 UTC | CC | joflynn | |
| Bertrand | 2019-10-02 16:04:57 UTC | Depends On | 1757886 | |
| James Parker | 2019-10-09 12:11:16 UTC | CC | jparker | |
| QA Contact | lyarwood | jparker | ||
| Steve Linabery | 2019-10-17 16:19:04 UTC | CC | slinaber | |
| errata-xmlrpc | 2019-10-17 16:20:43 UTC | Status | MODIFIED | ON_QA |
| PnT Account Manager | 2019-10-25 21:30:15 UTC | CC | atelang | |
| James Parker | 2019-11-02 02:57:25 UTC | Status | ON_QA | VERIFIED |
| Alex McLeod | 2019-11-04 12:05:52 UTC | CC | amcleod | |
| Doc Text | This feature adds support for PCI device NUMA affinity policies. These are configured as part of the `[pci]alias` configuration options. There are three policies supported: - `required` - `legacy` - `preferred` In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity you can forsake before failing to schedule. These policies allow you to configure how strict your NUMA affinity is on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. When the 'preferred' policy is configured for a PCI device, nova now utilizes CPUs on a different NUMA node from the NUMA node of the PCI device if this is all that is available. This results in increased resource utilization with the downside of reduced performance for these instances. | This feature adds support for PCI device NUMA affinity policies. You can configure these policies as part of the `[pci]alias` configuration options. There are three policies supported: - `required` - `legacy` - `preferred` In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity you can forsake before failing to schedule. You can use these policies to configure how strict your NUMA affinity is on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. When you configure the 'preferred' policy for a PCI device, nova now utilizes CPUs on a different NUMA node from the NUMA node of the PCI device if only a different node is available. This results in increased resource utilization with the downside of reduced performance for these instances. |
||
| Alex McLeod | 2019-11-05 10:25:15 UTC | Doc Text | This feature adds support for PCI device NUMA affinity policies. You can configure these policies as part of the `[pci]alias` configuration options. There are three policies supported: - `required` - `legacy` - `preferred` In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is how much NUMA affinity you can forsake before failing to schedule. You can use these policies to configure how strict your NUMA affinity is on a per-device basis or, more specifically, per device alias. This is useful to ensure maximum resource utilization. When you configure the 'preferred' policy for a PCI device, nova now utilizes CPUs on a different NUMA node from the NUMA node of the PCI device if only a different node is available. This results in increased resource utilization with the downside of reduced performance for these instances. | This release includes support for PCI device NUMA affinity policies. You can configure these policies as part of the `[pci]alias` configuration options. The following policies are supported: - `required` - `legacy` - `preferred` In all cases, strict NUMA affinity is provided if possible. The key difference between the policies is the amount of NUMA affinity you can sacrifice before failing to schedule. You can use these policies to configure how strict your NUMA affinity is for each device, or more specifically, for each device alias. This is useful to ensure maximum resource utilization. When you configure the 'preferred' policy for a PCI device, nova now utilizes CPUs on a different NUMA node from the NUMA node of the PCI device, if only a different node is available. This results in increased resource utilization with the downside of reduced performance for these instances. |
| errata-xmlrpc | 2019-11-06 15:15:43 UTC | Status | VERIFIED | RELEASE_PENDING |
| errata-xmlrpc | 2019-11-07 14:14:01 UTC | Status | RELEASE_PENDING | CLOSED |
| Resolution | --- | ERRATA | ||
| Last Closed | 2019-11-07 14:14:01 UTC | |||
| errata-xmlrpc | 2019-11-07 14:14:29 UTC | Link ID | Red Hat Product Errata RHBA-2019:3796 | |
| Lee Yarwood | 2019-11-22 10:43:00 UTC | Depends On | 1775575 | |
| Lee Yarwood | 2019-11-22 10:47:09 UTC | Depends On | 1775576 | |
| Karrar Fida | 2020-01-16 20:06:32 UTC | Depends On | 1783354 | |
| Depends On | 1791991 | |||
| Red Hat One Jira (issues.redhat.com) | 2022-03-13 15:18:18 UTC | Link ID | Red Hat Issue Tracker OSP-13596 | |
| Jeff Fearn 🐞 | 2023-02-22 23:02:48 UTC | CC | rhos-flags |
Back to bug 1561961