[Starlingx-discuss] How to enable cpu_dedicated_set

open infra openinfradn at gmail.com
Mon Nov 14 16:03:38 UTC 2022


Hi,

On Thu, Nov 10, 2022 at 6:42 PM Khalil, Ghada <Ghada.Khalil at windriver.com>
wrote:

> Hello,
>
> Regarding the question below:
>
>        Is there way I can enable application-isolation across all the
> worker nodes including new nodes to be added in future?
>
>
>
> Application-isolated CPUs are configured on a per-host basis. When new
> nodes are added, they will need to be explicitly configured.
>
>
>
> Is there a specific use-case that you have where it would be beneficial
> for new nodes to automatically inherit the config of an existing node?
> There was an old capability where host profiles could be created and
> applied to new nodes. But this proved problematic as exact hardware matches
> were not guaranteed. As a result, this capability was abandoned a long time
> ago.
>


I was looking for easy method for re-using configuration of for group of
nodes and make sure no human errors.
Lets say, all nodes in bladcenter A get specific config.
Right now we have less than 20 workers in a test bed but when it scale to
hundreds there is a high chance we might miss one node.



>
> Regards,
>
> Ghada
>
>
>
> *From:* open infra <openinfradn at gmail.com>
> *Sent:* Wednesday, September 28, 2022 3:32 AM
> *To:* Brito, Thiago <Thiago.Brito at windriver.com>
> *Cc:* starlingx-discuss at lists.starlingx.io
> *Subject:* Re: [Starlingx-discuss] How to enable cpu_dedicated_set
>
>
>
> [Please note: This e-mail is from an EXTERNAL e-mail address]
>
>
>
>
>
> On Wed, Dec 8, 2021 at 3:16 AM Brito, Thiago <Thiago.Brito at windriver.com>
> wrote:
>
> Did you do an application-apply after you changed the processors to
> application-isolated? Can you send the output of the command below for you
> compute pods:
>
>
>
> kubectl exec -n openstack nova-compute-[woker|controller]-X-XXXXX-xxxx --
> cat /etc/nova/nova.conf|grep dedicated_set
>
>
>
>
>
> Is there way I can enable application-isolation across all the worker
> nodes including new nodes to be added in future?
>
>
>
> Don't forget to change the id of the pod for your compute pods.
>
>
>
> Thiago
> ------------------------------
>
> *From:* open infra <openinfradn at gmail.com>
> *Sent:* Tuesday, December 7, 2021 5:45 AM
> *To:* Brito, Thiago <Thiago.Brito at windriver.com>
> *Cc:* starlingx-discuss at lists.starlingx.io <
> starlingx-discuss at lists.starlingx.io>
> *Subject:* Re: [Starlingx-discuss] How to enable cpu_dedicated_set
>
>
>
> [Please note: This e-mail is from an EXTERNAL e-mail address]
>
> Here is the info from STX side
>
>
>
>
>
> *Processors *
>
> 2
>
> *Physical Cores Per Processor *
>
> 32
>
> *Hyper-Threading *
>
> Yes
>
> CPU Assignments
>
> Create Cpu Profile
> <https://urldefense.com/v3/__http:/37.50.100.18:8080/admin/8/addcpuprofile/__;!!AjveYdw8EvQ!Km5xOEiy_wI3YAmney4OpHhqwjOOCiNhB37qanNSizdNpERIM_eDDRfwAzNuQoVun-HI$>
>
> *Function*
>
> *Processor Logical Cores*
>
> Platform
>
> *Processor 0 : *0,64
>
> vSwitch
>
> Shared
>
> Application
>
> *Processor 0 : *17-31,81-95
> *Processor 1 : *56-63,120-127
>
> Application-isolated
>
> *Processor 0 : *1-16,65-80
> *Processor 1 : *32-55,96-119
>
>
>
> On Tue, Dec 7, 2021 at 1:46 PM open infra <openinfradn at gmail.com> wrote:
>
> Hi Thiago,
>
>
>
> I have configured the worker as mentioned in the given link and dedicated
> cpu_policy set in the flavor.
>
> Instance creation failed with the following error.
>
>
>
> No valid host was found. There are not enough hosts available.
>
>
>
> controller-0:~$ openstack flavor show dn.large  -c properties
>
>
> +------------+--------------------------------------------------------------------------------------------------------+
>
> | Field      | Value
>                                             |
>
>
> +------------+--------------------------------------------------------------------------------------------------------+
>
> | properties | hw:cpu_cores='2', hw:cpu_policy='dedicated',
> hw:cpu_sockets='1', hw:cpu_threads='2', hw:numa_nodes='1' |
>
>
> +------------+--------------------------------------------------------------------------------------------------------+
>
>
>
> Regards,
>
> Danishka
>
>
>
> On Tue, Dec 7, 2021 at 1:18 AM Brito, Thiago <Thiago.Brito at windriver.com>
> wrote:
>
> Hi Danishka,
>
>
>
> Whatever you configure on the platform as Application-isolated core for a
> node will get to cpu_dedicated_set on stx-openstack. Please check this doc:
> https://docs.starlingx.io/r/stx.5.0/admintasks/openstack/configure-dedicated-and-shared-cpu-pools-on-hosts.html
> <https://urldefense.com/v3/__https:/docs.starlingx.io/r/stx.5.0/admintasks/openstack/configure-dedicated-and-shared-cpu-pools-on-hosts.html__;!!AjveYdw8EvQ!Km5xOEiy_wI3YAmney4OpHhqwjOOCiNhB37qanNSizdNpERIM_eDDRfwAzNuQv9j-Hlr$>
>
>
>
> Thiago
> ------------------------------
>
> *From:* open infra <openinfradn at gmail.com>
> *Sent:* Monday, December 6, 2021 3:15 PM
> *To:* starlingx-discuss at lists.starlingx.io <
> starlingx-discuss at lists.starlingx.io>
> *Subject:* [Starlingx-discuss] How to enable cpu_dedicated_set
>
>
>
> [Please note: This e-mail is from an EXTERNAL e-mail address]
>
> Hi,
>
>
>
> Is the cpu_dedicated_set is already enabled in stx based openstack?
>
> How to verify and enable it?
>
>
>
> Regards,
>
> Danishka
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.starlingx.io/pipermail/starlingx-discuss/attachments/20221114/3db52354/attachment-0001.htm>


More information about the Starlingx-discuss mailing list