| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-03-21 | |||
| 08:48:01 | bauzas | venki: looks like an ironic issue to me | |
| 08:48:17 | bauzas | lemme explain why | |
| 08:48:37 | bauzas | ironic creates three fake ironc nodes using the driver you select | |
| 08:49:11 | bauzas | then, n-cpu runs the RT that calls the ironic driver for knowing how many ironic nodes are present | |
| 08:49:22 | bauzas | oops RT = Nova ResourceTracker | |
| 08:49:41 | bauzas | then, Nova creates one hypervisor per ironic node found | |
| 08:50:05 | bauzas | last step, Ironic waits for nova to discover those nodes it created | |
| 08:50:45 | bauzas | once nova creates the according 3 hypervisors with the same UUID that match the ironic node UUID (behind a single n-cpu service), then ironic continues its deployment | |
| 08:50:51 | bauzas | but | |
| 08:51:22 | bauzas | if Ironic can't create those 3 nodes using a specific driver, then Nova doesn't find those nodes, hence the error you see | |
| 08:52:00 | bauzas | when you say the stack is created if using another ironic driver, that makes me think the real problem is that you don't end up having 3 ironic nodes, that's it | |
| 08:53:41 | venki | @bauzas : awww okey .... nice explanation :):):) | |
| 09:03:58 | johnthetubaguy | gmann: I wasn't understanding your comments on the policy stuff totally, are you still around? | |
| 09:35:00 | lyarwood | johnthetubaguy: morning, https://review.openstack.org/#/c/439520/ could use a review if you have any time today. | |
| 09:35:29 | lyarwood | johnthetubaguy: it's now first up in the series given the -1 from mdbooth on landing the db changes before we know if we need the get_by_attachment_id lookups | |
| 09:35:54 | johnthetubaguy | lyarwood ah, yeah, I need to hit that again, will try get to that today | |
| 09:36:10 | lyarwood | johnthetubaguy: cheers | |
| 09:42:41 | openstackgerrit | John Garbutt proposed openstack/nova master: Refactor: remove _items() in nova/api/openstack/compute/attach_interfaces.py https://review.openstack.org/447698 | |
| 09:46:57 | tojuvone | johnthetubaguy, bauzas , jaypipes, oomichi, gibi Would like to get this forward if you have time: https://review.openstack.org/310510 | |
| 09:47:21 | bauzas | tojuvone: either way, I need to make another pass on specs soon | |
| 09:47:28 | bauzas | tojuvone: but will do :) | |
| 09:47:51 | tojuvone | bauzas, thanks :) | |
| 09:49:33 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposed block accepting AZ renames https://review.openstack.org/446446 | |
| 09:49:42 | bauzas | johnthetubaguy: you could be interested in ^ | |
| 09:50:02 | bauzas | I still need to write a backlog spec for commenting out all the AZ issues | |
| 09:51:22 | bauzas | heh :) | |
| 09:52:04 | bauzas | johnthetubaguy: you also have what? 8 open specs ? :p | |
| 09:52:23 | johnthetubaguy | yeah, mostly not being written by me right know though | |
| 09:52:27 | gibi | tojuvone: It is on my queue to look at it again | |
| 09:53:23 | tojuvone | gibi, thanks, tried to update with things we discussed | |
| 09:56:41 | openstackgerrit | John Garbutt proposed openstack/nova master: Add description to policies in agents.py https://review.openstack.org/445208 | |
| 09:59:27 | openstackgerrit | John Garbutt proposed openstack/nova master: Add descripiton to policies in virtual_interfaces.py https://review.openstack.org/445652 | |
| 10:00:58 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposed block accepting AZ renames https://review.openstack.org/446446 | |
| 10:03:53 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tox: Add 'profiler' target https://review.openstack.org/437055 | |
| 10:05:27 | openstackgerrit | Claudiu Belu proposed openstack/nova master: POC: tests: autospecs all the mock.patch usages https://review.openstack.org/447505 | |
| 10:11:47 | openstackgerrit | sahid proposed openstack/nova master: libvirt: return a CPU overhead if isolate emulator threads requested https://review.openstack.org/385364 | |
| 10:11:47 | openstackgerrit | sahid proposed openstack/nova master: numa: fit instance NUMA node with cpus reserved onto host NUMA node https://review.openstack.org/400875 | |
| 10:11:48 | openstackgerrit | sahid proposed openstack/nova master: virt: update overhead to take into account vCPUs https://review.openstack.org/385024 | |
| 10:11:53 | openstackgerrit | sahid proposed openstack/nova master: libvirt: make emulator threads to run on the reserved pCPU https://review.openstack.org/400878 | |
| 10:11:54 | openstackgerrit | sahid proposed openstack/nova master: numa: take into account cpus reserved https://review.openstack.org/400876 | |
| 10:11:55 | openstackgerrit | sahid proposed openstack/nova master: numa: update numa usage incording the cpu reserved https://review.openstack.org/400877 | |
| 10:13:21 | sfinucan | johnthetubaguy: Couldn't we just add the docstrings to https://review.openstack.org/#/c/434842/ and keep everyone happy? | |
| 10:14:19 | sfinucan | I'm assuming the oslo.policy change/release will take a while to make it into nova | |
| 10:14:29 | johnthetubaguy | sfinucan: its a method thats about to be deleted, its already released yesterday | |
| 10:14:45 | sfinucan | Oh, it's been released already? | |
| 10:14:47 | sfinucan | In that case | |
| 10:14:48 | johnthetubaguy | yeah | |
| 10:15:10 | johnthetubaguy | just need the requirements bumps to flow through, then rework stuff to use the new method | |
| 10:16:20 | sfinucan | Gotcha. Does seem more sane to rework after the requirements bump, in that case? | |
| 10:21:29 | johnthetubaguy | sfinucan: so that whole method will be deleted after the bump goes through | |
| 10:21:42 | johnthetubaguy | sfinucan: I don't think the bump has been put up for review yet | |
| 10:22:14 | sfinucan | johnthetubaguy: Right, but all the calls to that function will have to be rewritten? | |
| 10:22:26 | sfinucan | So could we replace that change with a requirement.txt bump change? | |
| 10:22:44 | sfinucan | ...and turn a two step process to one | |
| 10:22:49 | johnthetubaguy | sfinucan: yes, they will all have to get re-writen | |
| 10:24:11 | sfinucan | johnthetubaguy: Right. Any reason not to do that now? | |
| 10:24:29 | sfinucan | Just making sure we're not creating work for ourselves :) | |
| 10:24:32 | johnthetubaguy | sfinucan: it delays getting the harder reviews out of our queue | |
| 10:24:53 | johnthetubaguy | the rename is a single simple patch really, once the requirements stuff is all lined up | |
| 10:35:50 | sfinucan | johnthetubaguy: OK, the method signatures are the same meaning the rename can be done en masse, as you say. +2/+W'd | |
| 10:38:16 | johnthetubaguy | sfinucan: cool, thanks | |
| 10:38:36 | johnthetubaguy | sfinucan: operators been asking for better docs on this stuff for a few years now, glad to be making progress on it | |
| 10:38:58 | sfinucan | johnthetubaguy: Yup, similar idea behind centralize-config-options | |
| 10:39:23 | sfinucan | One question though - why aren't the 'discoverable' policies documented? https://review.openstack.org/#/c/447722/1/nova/policies/suspend_server.py | |
| 10:39:33 | johnthetubaguy | sfinucan: we are about to delete them all | |
| 10:40:17 | sfinucan | Oh, that part of this spec or some other initiative? | |
| 10:40:24 | sfinucan | *is that | |
| 10:40:28 | johnthetubaguy | its the follow on spec | |
| 10:40:53 | johnthetubaguy | now all extensions are hard coded on by default, having policy to hide a few from the list seems pointless really | |
| 10:42:46 | sfinucan | makes sense | |
| 10:44:14 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in suspend_server.py https://review.openstack.org/447722 | |
| 10:44:19 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: WIP/DNM: Destroy the ReqSpec object when deleting the instance https://review.openstack.org/391060 | |
| 10:46:51 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in volumes.py https://review.openstack.org/445637 | |
| 10:50:14 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in volumes_attachments.py https://review.openstack.org/445252 | |
| 10:52:56 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/nova master: Read NIC features in libvirt https://review.openstack.org/444415 | |
| 10:58:10 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in attach_interfaces.py https://review.openstack.org/447696 | |
| 10:59:03 | openstackgerrit | Lee Yarwood proposed openstack/nova master: compute: Detach volumes on _rebuild_default_impl failure https://review.openstack.org/442105 | |
| 11:00:18 | cdent | alex_xu: can you do the same curl that you left in your comment, but with a -v so we can see the request headers too? | |
| 11:02:06 | cdent | alex_xu: I'm pretty sure your request is somehow getting a content-type header | |
| 11:04:39 | alex_xu | cdent: thanks, curl send a "Content-Type: application/x-www-form-urlencoded" | |
| 11:04:53 | alex_xu | I got that with -v | |
| 11:05:58 | cdent | alex_xu: strange, I would expect the result to be a 415 then | |
| 11:06:27 | alex_xu | cdent: yea, I got a 415 | |
| 11:07:00 | cdent | oh, so which code did you remove that allowed the 200? | |
| 11:09:17 | alex_xu | cdent: sorry, I didn't get your quesion | |
| 11:09:42 | alex_xu | I removed the check at here also https://review.openstack.org/#/c/447625/2/nova/api/openstack/placement/handler.py | |
| 11:09:59 | sdague | alex_xu / johnthetubaguy / sfinucan anyone want to hit this easy one - https://review.openstack.org/#/c/447466/ - the docs change is on top | |
| 11:10:19 | sfinucan | sdague: Sure | |
| 11:10:34 | alex_xu | sdague: checking | |
| 11:11:19 | cdent | alex_xu: the reason, in that case, that you didn't get 500 error that the original bug fixed is because when you use -d with curl, it automatically sends a content-type header and the error that was being protected against was when there was no content-type header. do you think there's something wrong with the new patch? | |
| 11:11:58 | sfinucan | sdague: Two small indentation things. Otherwise TrivialFix | |
| 11:16:13 | openstackgerrit | Sean Dague proposed openstack/nova master: remove log translation from nova.api.metadata https://review.openstack.org/447466 | |
| 11:16:13 | openstackgerrit | Sean Dague proposed openstack/nova master: update i18n guide for nova https://review.openstack.org/447528 | |
| 11:16:42 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in agents.py https://review.openstack.org/445208 | |
| 11:17:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add description to policies in aggregates.py https://review.openstack.org/445230 | |
| 11:17:07 | openstackgerrit | Andrey Volkov proposed openstack/nova master: PoC: Pick PCI devices with unique tags https://review.openstack.org/448008 | |
| 11:19:18 | alex_xu | sfinucan: nothing i found from me, you can +w directly as you mention it is trivialfix | |
| 11:20:09 | sfinucan | alex_xu, sdague: Sure. Done. | |
| 11:20:33 | alex_xu | cdent: i didn't finish the review yet, just want to figure why those code are there first | |
| 11:20:47 | Dinesh_Bhor | sfinucan: Hi, I just rebased the patch: https://review.openstack.org/#/c/445782/ could you please take a look at it? | |