| *** pmannidi|AFK is now known as pmannidi | 00:33 | |
| hgy | Sorry to bother you, I would like to ask whether Ironic counts the resources of bare metal nodes which are managed by ironic into the total available resources | 02:44 |
|---|---|---|
| hgy | Sorry, I know this question is unclear | 02:50 |
| hgy | The reason is I got “exceed quote for cores and ram” | 02:52 |
| hgy | When I deploy a bare metal node | 02:52 |
| hgy | Oh, sorry， I find the cause, I didn't modify the default quote | 05:39 |
| hgy | I will try it | 05:39 |
| *** redrobot2 is now known as redrobot | 05:58 | |
| arne_wiebalck | Good morning, Ironic! | 07:09 |
| arne_wiebalck | hgy: The quota is not managed nor checked by Ironic, so it follows the same rules as for virtual machines. | 07:10 |
| arne_wiebalck | NobodyCam: Good morning! Sorry, I left right after that comment. I will try to have a look today! | 07:12 |
| iurygregory | good morning arne_wiebalck and Ironic o/ | 07:25 |
| MikeCTZA | morning arne_wiebalck and all ironic | 07:27 |
| arne_wiebalck | Good morning iurygregory and MikeCTZA o/ | 07:31 |
| MikeCTZA | openstack hypervisor stats show should show me non 0 values based on the baremetal node that I have in openstack right? and likewise hypervisor show should match (assuming just 1 node), mine are all zero | 07:37 |
| MikeCTZA | I'm having 1 more go at trying to get a box working before our brainstorming meeting in a few hours time | 07:40 |
| MikeCTZA | (an internal team meeting that is) | 07:40 |
| arne_wiebalck | MikeCTZA: I barely use "openstack hypervisor ..." | 08:01 |
| arne_wiebalck | MikeCTZA: let me check what I see ... | 08:01 |
| arne_wiebalck | MikeCTZA: so, on an active hypervisor, the resources are used (vcpus 0, vcpus_used 64) | 08:12 |
| arne_wiebalck | MikeCTZA: on an available hypervisor, the resources are not used but also not there (vpus 0, vcpus_used 0) | 08:16 |
| MikeCTZA | OK so I need to look a bit more at nova scheduler then as when I try deploy a box I get No valid host was found, been working through the docs on what to check but not found anything that solves my issue as yet | 08:23 |
| rpittau | good morning ironic! o/ | 08:34 |
| janders | hey arne_wiebalck iurygregory MikeCTZA rpittau and Ironic o/ | 08:37 |
| iurygregory | hey janders rpittau o/ | 08:37 |
| rpittau | hey janders iurygregory :) | 08:39 |
| dtantsur | morning ironic | 09:14 |
| rpittau | hey dtantsur :) | 09:14 |
| arne_wiebalck | MikeCTZA: this is on which release? | 09:14 |
| arne_wiebalck | MikeCTZA: first thing I would check for "no valid host" is placement | 09:14 |
| arne_wiebalck | hey janders rpittau and dtantsur o/ | 09:14 |
| rpittau | hey arne_wiebalck :) | 09:14 |
| zigo | Hi there! Can someone have a look at this Debian bug and help me fix? | 09:37 |
| zigo | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=997545 | 09:37 |
| dtantsur | hi zigo! stendulker may be able to | 09:41 |
| dtantsur | I wonder if it's related to a combination of newer sushy and older proliantutils? | 09:41 |
| zigo | dtantsur: I can try to upgrade both, sure! :) | 09:43 |
| stendulker | yes, it needs to use newer sushy | 09:46 |
| zigo | stendulker: dtantsur: I'm already on sushy 3.12.0 | 10:15 |
| zigo | And it's also latest proliantutils 2.12.0 ... | 10:15 |
| opendevreview | Merged openstack/ironic-python-agent master: Delete EFI boot entry duplicate labels first https://review.opendev.org/c/openstack/ironic-python-agent/+/815899 | 10:19 |
| opendevreview | Verification of a change to openstack/ironic-python-agent master failed: Add blockdev to re-read the partition table https://review.opendev.org/c/openstack/ironic-python-agent/+/816192 | 10:19 |
| dtantsur | arne_wiebalck: rebase required, I suspect ^^ | 10:19 |
| *** pmannidi is now known as pmannidi|AFK | 10:34 | |
| stendulker | zigo: It looks like some of more calls needs to be fixed in proliantutils | 10:41 |
| arne_wiebalck | dtantsur: thanks, I will have a look! | 10:54 |
| *** dviroel|out is now known as dviroel|rover | 11:21 | |
| janders | see you tomorrow Ironic o/ | 11:30 |
| opendevreview | Arne Wiebalck proposed openstack/ironic-python-agent master: Add blockdev to re-read the partition table https://review.opendev.org/c/openstack/ironic-python-agent/+/816192 | 11:51 |
| arne_wiebalck | dtantsur: rpittau ^^ (I rebased the patch) | 11:51 |
| dtantsur | thanks! | 11:52 |
| * dtantsur is not looking forward towards his rebase after these 2 patches | 11:52 | |
| arne_wiebalck | dtantsur: yeah, I was hoping to get mine in before your 800(?) line patch :-D | 11:52 |
| dtantsur | which pretty much just moves the code around :) | 11:53 |
| dtantsur | but yeah | 11:53 |
| arne_wiebalck | would it be a lot easier if we stop mine and merge yours first? | 11:54 |
| arne_wiebalck | mine is a one-liner, it is mostly the tests | 11:54 |
| arne_wiebalck | but yours was first clearly :) | 11:55 |
| rpittau | I'll hold approving | 11:57 |
| dtantsur | arne_wiebalck: yours will need a backport, so it should probably go first | 11:59 |
| dtantsur | rpittau: ^^^ | 11:59 |
| dtantsur | I think I'll survive a rebase :) | 11:59 |
| rpittau | ack | 11:59 |
| dtantsur | arne_wiebalck: meanwhile, could you check https://review.opendev.org/c/openstack/ironic-python-agent/+/815629? will make rebasing slightly easier | 12:07 |
| opendevreview | Merged openstack/metalsmith master: Update pep8 test requirements https://review.opendev.org/c/openstack/metalsmith/+/814543 | 12:55 |
| timeu | MikeCTZA: regarding your "no valid host", the troubleshooting guide has some infos: https://docs.openstack.org/ironic/queens/admin/troubleshooting.html. Best is to check the nova scheduler log and check what the various filters return. Are you using traits or resource_class for your flavors/nodes ? | 12:57 |
| opendevreview | Verification of a change to openstack/ironic-python-agent-builder master failed: Bump pip for tinyipa to 21.3.1 https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/814894 | 13:27 |
| opendevreview | Merged openstack/ironic-python-agent master: Stop requiring mocking of utils.execute if ironic-lib execute is mocked https://review.opendev.org/c/openstack/ironic-python-agent/+/815629 | 14:19 |
| TheJulia | arne_wiebalck: when you changed the conductor group for a nova compute node, what exactly happened? | 14:42 |
| opendevreview | Aija Jauntēva proposed openstack/ironic master: Fix idrac-wsman deploy with existing non-BIOS jobs https://review.opendev.org/c/openstack/ironic/+/816521 | 14:46 |
| arne_wiebalck | TheJulia: I have to run to a meeting in a moment, but I think the resource provider was (soft)deleted and could not be recreated by the new responsible. | 14:58 |
| TheJulia | resource provider soft deleted because it just completely disappeared right? | 14:59 |
| TheJulia | at least, to nova's handling of partition keys | 14:59 |
| TheJulia | arne_wiebalck: I'm going to add you on a nova change I'm working on. Closely related issues | 15:00 |
| dtantsur | see you tomorrow folks o/ | 16:00 |
| rpittau | bye dtantsur o/ | 16:13 |
| opendevreview | Aija Jauntēva proposed openstack/ironic master: Fix idrac-wsman deploy with existing non-BIOS jobs https://review.opendev.org/c/openstack/ironic/+/816521 | 16:13 |
| opendevreview | Merged openstack/ironic-python-agent-builder master: Bump pip for tinyipa to 21.3.1 https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/814894 | 16:38 |
| rpittau | good night! o/ | 16:59 |
| TheJulia | o/ | 17:03 |
| arne_wiebalck | TheJulia: Meeting done :) | 17:14 |
| arne_wiebalck | TheJulia: So, the issue was that Nova (not placement) soft-deleted the compute node and when it appeared on another nova-compute the insertion with the same UUID fails. | 17:14 |
| arne_wiebalck | *soft-deleted the node/resource provider | 17:15 |
| arne_wiebalck | FWIU, the initial registration is done by nova and in nova as well as in placement. | 17:15 |
| arne_wiebalck | The resource tracker only updates the resource of this resource provider. | 17:16 |
| TheJulia | That seems like an area patched with-in the last couple of cycles | 17:16 |
| arne_wiebalck | TheJulia: this is very possible | 17:16 |
| arne_wiebalck | TheJulia: Our nova is on Stein. | 17:16 |
| arne_wiebalck | Therefore my hesitation to make definitive statements on this, it may have been all fixed already. | 17:17 |
| arne_wiebalck | bye everyone, see you tomorrow o/ | 17:22 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent master: TBD: Add Lenovo SR650 issue note https://review.opendev.org/c/openstack/ironic-python-agent/+/816545 | 17:35 |
| TheJulia | NobodyCam: what if we had a "standby" state | 18:53 |
| TheJulia | which may power things down, or not? | 18:53 |
| TheJulia | or do other "magic" | 18:53 |
| NobodyCam | Oh so a state beyond active | 18:55 |
| NobodyCam | And good morning | 18:55 |
| NobodyCam | How would the system know to “wake” up is powered down or a sleep | 18:56 |
| TheJulia | I think it would need to be an explicit case | 18:56 |
| TheJulia | so the model I know exists in some shops is you have two clusters | 18:57 |
| TheJulia | Like A side and B side | 18:57 |
| TheJulia | B is secondary, but needs to be "on" | 18:57 |
| TheJulia | but not used | 18:57 |
| TheJulia | A, until it is explicitly failed, takes the workload | 18:57 |
| NobodyCam | That’s interesting. So a pool of cold standbys | 19:00 |
| TheJulia | well | 19:01 |
| TheJulia | warm/luke warm | 19:01 |
| NobodyCam | If they are sleeping or powered down. My concern would how are they spun up on demand | 19:02 |
| TheJulia | I think $higher level thing | 19:04 |
| TheJulia | err, there should have been underscores | 19:05 |
| TheJulia | something has to make the decision to fail a cluster | 19:05 |
| TheJulia | and that act may be a set of additional work, where it is not really a reprovision, just a state update | 19:05 |
| TheJulia | and maybe bios reconfiguration | 19:05 |
| NobodyCam | Humm. I see where that could be useful. | 19:07 |
| TheJulia | "reconfigure" | 19:08 |
| TheJulia | almost | 19:08 |
| NobodyCam | I would okay with a standby / state. Concern is would we need to boot ipa to make the bios updates and the. Reboot back into the image | 19:08 |
| NobodyCam | That could add 10 minutes to power on | 19:09 |
| TheJulia | Yeah, I *think* reconfigure may be a totally valid thing there | 19:09 |
| TheJulia | but the case I'm thinking is largely OOB bios settings and reboot | 19:09 |
| TheJulia | If we were to do standby, reconfigure could just make sense at the same time | 19:16 |
| TheJulia | since we're wiring in the same places | 19:16 |
| TheJulia | but... how to know what to do then..., | 19:16 |
| stevebaker[m] | morning | 19:25 |
| TheJulia | good morning stevebaker[m] | 19:27 |
| *** dviroel|rover is now known as dviroel|rover|afk | 19:45 | |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: WIP: Enable improved power/state management https://review.opendev.org/c/openstack/ironic-specs/+/816551 | 19:49 |
| TheJulia | stevebaker[m]: NobodyCam: tzumainn: zer0c00l: rloo: eandersson: arne_wiebalck: thoughts/insight might be useful on https://review.opendev.org/c/openstack/ironic-specs/+/816551 | 19:54 |
| stevebaker[m] | looking | 19:54 |
| tzumainn | TheJulia, at a first glance it seems really cool; I have to admit that I don't really understand what 'standby' would mean for a node | 20:55 |
| tzumainn | I'm going to point the spec out to the MOC people to see if they have thoughts | 20:55 |
| *** dviroel|rover|afk is now known as dviroel|rover | 21:29 | |
| *** dviroel|rover is now known as dviroel|rover|out | 22:18 | |
| TheJulia | tzumainn: thinking like a standby configuration or state, like on reserve | 22:19 |
| TheJulia | Please don’t reuse this machine to mine bitcoin | 22:19 |
| TheJulia | Keep it in low power mode | 22:20 |
| janders | good morning Ironic o/ | 23:43 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!