| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-04-19 | |||
| 10:58:25 | yikun_ | ^ @gmann I'm not sure it's a best solution or not, but it is a way to solve our problem at least. | |
| 10:59:09 | yikun_ | gmann, you could take a look when u have time. :) https://review.openstack.org/562547 | |
| 11:17:32 | Kevin_Zheng | gibi are you around? | |
| 11:20:38 | gibi | Kevin_Zheng: yes | |
| 11:21:24 | Kevin_Zheng | I almost done the coding, but again, there are some problems with the tests | |
| 11:22:32 | gibi | Kevin_Zheng: you can push the patch up then I can take a look at it and might be able to help with the test | |
| 11:22:39 | Kevin_Zheng | as you guys left in the comments, I suppose to adding req_id for instance update notifications and not include it for periodic tasks | |
| 11:23:07 | Kevin_Zheng | which matt suggested using the judgement of whether context.auth_token is present | |
| 11:23:21 | Kevin_Zheng | since we use noauth middleware | |
| 11:23:36 | Kevin_Zheng | there will always be no context.auth | |
| 11:23:50 | Kevin_Zheng | https://github.com/openstack/nova/blob/324899c621ee02d877122ba3412712ebb92831f2/nova/api/openstack/auth.py#L57 | |
| 11:24:27 | Kevin_Zheng | I will put it up if I cannot come up with ideas before I go | |
| 11:24:46 | gibi | Kevin_Zheng: your understanding seems correct to me | |
| 11:25:55 | Kevin_Zheng | maybe we shouldn't judge by authtoken | |
| 11:26:31 | Kevin_Zheng | because if someone is actually using Noauth middleware, he will always got no req_id in the notification | |
| 11:26:32 | gibi | Kevin_Zheng: then we have to find another way to distinguish between context from a periodic task and another normal context | |
| 11:26:47 | gibi | Kevin_Zheng: let me dig a bit | |
| 11:26:56 | Kevin_Zheng | Thanks | |
| 11:31:53 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Return all resources in provider_summaries https://review.openstack.org/558045 | |
| 11:31:54 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Return all nested providers in tree https://review.openstack.org/559480 | |
| 11:31:55 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Return resource providers without inventories https://review.openstack.org/559554 | |
| 11:34:49 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Add request_id to instance action notifications https://review.openstack.org/553288 | |
| 11:36:51 | kashyap | Are others hitting these IBM zVM CI failures too: http://extbasicopstackcilog01.podc.sl.edst.ibm.com/test_logs/jenkins-check-nova-master-16984/console.html | |
| 11:37:17 | kashyap | It is all: "Exceeded maximum number of retries. Exhausted all hosts available for retrying build failures for instance" | |
| 11:41:20 | Kevin_Zheng | gibi Hi, I got it, maybe I should use (context.is_admin and not context.auth_token) rather than just context.auth_token | |
| 11:43:23 | gibi | Kevin_Zheng: I think that is a partial solution. If the deployer use NoAuth and the user is an admin or the non admin context is elevated somewhere in the code path then your condition will be true | |
| 11:43:53 | Kevin_Zheng | yeah | |
| 11:50:03 | openstackgerrit | Yikun Jiang (Kero) proposed openstack/nova master: Support inherit api samples in ApiSampleTestBase https://review.openstack.org/562547 | |
| 11:51:49 | gibi | the NoAuth set user_id and project_id but the get_admin_context() call used by the periodic task set them to None | |
| 11:52:31 | Kevin_Zheng | Hmm | |
| 11:52:32 | gibi | Kevin_Zheng: so instead of checking for the auth_token you can check the project_id and user_id | |
| 11:52:43 | Kevin_Zheng | Let me try | |
| 11:54:02 | openstackgerrit | Merged openstack/osc-placement master: RP list: member_of and resources parameters (v1.3, v1.4) https://review.openstack.org/511183 | |
| 12:02:36 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Add request_id to instance action notifications https://review.openstack.org/553288 | |
| 12:02:58 | Kevin_Zheng | gibi ^ done | |
| 12:03:17 | gibi | Kevin_Zheng: thanks, looking | |
| 12:05:05 | mriedem | lyarwood: bauzas: today is rocky-1 and i think melwitt wanted to do some stable branch releases too - stable/ocata has a bunch of patches which need a final +2 if you can go through those https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/ocata | |
| 12:05:55 | gibi | Kevin_Zheng: you have to update the comment in https://review.openstack.org/#/c/553288/22/nova/notifications/objects/instance.py@124 | |
| 12:06:30 | bauzas | mriedem: yup, I reviewed the queens branch this morning and I was about to look at the pike and ocata branches this afternoon ;) | |
| 12:08:36 | Kevin_Zheng | gibi maybe shouldn't write about auth_token anymore? | |
| 12:09:46 | gibi | Kevin_Zheng: I'm OK with that too | |
| 12:10:01 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Add request_id to instance action notifications https://review.openstack.org/553288 | |
| 12:10:05 | gibi | Kevin_Zheng: but then we have to explain in the commit message that we could not simply use the auth_token as Matt suggested | |
| 12:10:40 | Kevin_Zheng | ...really? just updated | |
| 12:13:59 | gibi | Kevin_Zheng: or we can tell mriedem on IRC so he will know why we not used context.auth_token ;) | |
| 12:15:38 | Kevin_Zheng | gibi I like this one | |
| 12:15:49 | mriedem | i don't think anyone uses NoAuthMiddleware in production fwiw | |
| 12:15:51 | mriedem | or shouldn't | |
| 12:17:27 | Kevin_Zheng | yeah well, judging by project_id and user_id could work right? | |
| 12:18:09 | Kevin_Zheng | in the tests, we have admin actions so (is_admin and not auth_token) cannot work as well | |
| 12:18:29 | openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient master: [CLI] Fix token auth type https://review.openstack.org/424256 | |
| 12:18:38 | gibi | mriedem: still our functional env uses NoAuth which means if the code checks for the auth_token then our functional tests become useless | |
| 12:19:21 | gibi | mriedem: and checking for project_id and user_id seems meaningful for me as if the request comes from the user, then those should be set on the context and that also means a proper request_id is also set | |
| 12:19:58 | mriedem | yeah that's fine | |
| 12:20:26 | mriedem | an appropriate follow up for this is probably putting a helper method on the RequestContext itself to tell if it's a real user request | |
| 12:21:00 | mriedem | we have RequestContext.is_user_context but that returns False if is_admin is True | |
| 12:21:50 | gibi | mriedem: I think the appropriate follow up is to merge https://review.openstack.org/#/c/524306 which means we have a proper request_id for periodic tasks as well and therefore we can remove the check from the payload generation | |
| 12:22:55 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/ocata: only increment disk address unit for scsi devices https://review.openstack.org/561613 | |
| 12:23:03 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Deduplicate notification samples Rocky - 7 https://review.openstack.org/560748 | |
| 12:23:52 | openstackgerrit | sahid proposed openstack/nova master: compute: introduce cpu_shared_set option https://review.openstack.org/561912 | |
| 12:23:53 | openstackgerrit | sahid proposed openstack/nova master: libvirt: always pass emulator threads policy https://review.openstack.org/561913 | |
| 12:23:54 | openstackgerrit | sahid proposed openstack/nova master: libvirt: place emulator threads on CONF.cpu_shared_set https://review.openstack.org/510897 | |
| 12:23:55 | openstackgerrit | sahid proposed openstack/nova master: libvirt: update emulator threads docstring https://review.openstack.org/562565 | |
| 12:24:23 | openstackgerrit | Merged openstack/nova stable/queens: Add regression test for persisted RequestSpec.retry from failed resize https://review.openstack.org/560142 | |
| 12:24:31 | openstackgerrit | Merged openstack/nova stable/queens: Don't persist RequestSpec.retry https://review.openstack.org/560143 | |
| 12:25:30 | openstackgerrit | sahid proposed openstack/nova master: compute: introduce cpu_shared_set option https://review.openstack.org/561912 | |
| 12:25:31 | openstackgerrit | sahid proposed openstack/nova master: libvirt: always pass emulator threads policy https://review.openstack.org/561913 | |
| 12:25:32 | openstackgerrit | sahid proposed openstack/nova master: libvirt: place emulator threads on CONF.cpu_shared_set https://review.openstack.org/510897 | |
| 12:25:33 | openstackgerrit | sahid proposed openstack/nova master: libvirt: update emulator threads docstring https://review.openstack.org/562565 | |
| 12:26:47 | openstackgerrit | Merged openstack/nova stable/queens: List instances performace optimization https://review.openstack.org/561715 | |
| 12:28:05 | kashyap | mriedem: Morning, I think this is ready for some review: https://review.openstack.org/#/c/558783/ -- libvirt: Bump MIN_{LIBVIRT,QEMU}_VERSION for "Rocky" | |
| 12:28:16 | kashyap | mriedem: I'd like to build other compat code clean-up patches based on the above | |
| 12:28:32 | kashyap | Which is the bulk of the change | |
| 12:30:19 | kashyap | (Also my eyes are a bit hazy after arriving at that patch iterating a couple of weeks. A fresh pair of eyes taking a look at it would be nice.) | |
| 12:31:33 | mriedem | kashyap: i can put it in a tab for now | |
| 12:31:59 | kashyap | Merci. Some look at it and comments will give me a bit more push to fix whatever stuff reviewers might find. | |
| 12:33:45 | lyarwood | mriedem: ack will do | |
| 12:33:48 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Deduplicate notification samples Rocky - 7 https://review.openstack.org/560748 | |
| 12:34:08 | openstackgerrit | sahid proposed openstack/nova-specs master: fix blueprint related to the spec overhead-pin-set https://review.openstack.org/561916 | |
| 12:40:00 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Add request_id to instance action notifications https://review.openstack.org/553288 | |
| 12:44:36 | openstackgerrit | Merged openstack/nova stable/queens: Document how to disable notifications https://review.openstack.org/561954 | |
| 12:44:44 | openstackgerrit | Merged openstack/nova stable/queens: libvirt: Report the allocated size of preallocated file based disks https://review.openstack.org/562074 | |
| 12:45:58 | sahid | mriedem: about your comment on the trusted tag we are going to erase, i let a comment in the review | |
| 12:46:02 | sahid | https://review.openstack.org/#/c/562565/ | |
| 12:46:31 | sahid | i think there are some condition where that could be an issue | |
| 12:52:49 | mriedem | sahid: wrong patch? | |
| 12:57:47 | sahid | mriedem: yep sorry https://review.openstack.org/#/c/458820/ | |
| 13:08:00 | openstackgerrit | Matthew Booth proposed openstack/nova master: Rename recreate to evacuate in driver signatures https://review.openstack.org/560900 | |
| 13:12:46 | openstackgerrit | Lei Zhang proposed openstack/nova master: Improve check capacity sql https://review.openstack.org/561862 | |
| 13:20:09 | maciejjozefczyk | Guys, maybe you remember any problem related to metadata-agent and cloud-init during openstack upgrade? I heard about bugs like cleaning of instance ssh service keys during upgrades but I'm not able to track any details. | |
| 13:25:45 | mriedem | sahid: replied | |
| 13:25:57 | mriedem | cdent: i guess we already had a postgresql job in nova's experimental queue | |
| 13:26:00 | mriedem | but that doesn't help you | |
| 13:27:52 | cdent | mriedem: yeah, looks like both jay and eric are happy to merge my fix without test proof | |
| 13:28:11 | cdent | which would require a functional environment running pg or a new tempest test | |
| 13:28:42 | cdent | I explored doing a functional pg environment locally but it quickly fell apart for reasons I couldn't care to figure out (at least not last night) | |
| 13:29:21 | sahid | mriedem: ok so you may have right, i thought about that hard-reboot thing but it seems i'm wrong i will build that env and re-put my head in that whole serie | |
| 13:29:50 | jaypipes | what did I mess up now? | |
| 13:32:01 | cdent | jaypipes: you're safe this time | |
| 13:32:05 | openstackgerrit | Chris Dent proposed openstack/nova master: Add root and parent provider uuid to group by clause https://review.openstack.org/562379 | |