Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2020-02-20
02:33:49 openstackgerrit melanie witt proposed openstack/nova master: Move nova-manage placement sync_aggregates to nova-audit https://review.opendev.org/708787
07:32:54 openstack bug 1862633 in OpenStack Compute (nova) "unshelve leak allocation if update port fails" [Medium,In progress] https://launchpad.net/bugs/1862633 - Assigned to Balazs Gibizer (balazs-gibizer)
07:32:54 openstackgerrit Merged openstack/nova master: Reproduce bug 1862633 https://review.opendev.org/706867
08:56:54 stephenfin efried: It would be better if johnthetubaguy can handle vTPM, but if he can't then I can
09:19:34 openstackgerrit Stephen Finucane proposed openstack/python-novaclient master: setup.cfg: Various Python 3 fixes https://review.opendev.org/708824
09:40:05 bauzas efried: I can't attend this meeting but I'll leave notes in the etherpad
09:40:10 bauzas thanks for the ping either way
09:46:13 kashyap bauzas: "So, given a tempest.log, how do I find which was the last test to run?"
09:47:44 bauzas kashyap: link of the change?
09:49:17 kashyap bauzas: Actually, I forgot it is a local deployment that a QEMY developer was trying
10:06:03 openstackgerrit Stephen Finucane proposed openstack/python-novaclient master: Remove six https://review.opendev.org/708831
10:06:03 openstackgerrit Stephen Finucane proposed openstack/python-novaclient master: tox: Configure 'ignore_basepython_conflict' https://review.opendev.org/708830
10:06:04 openstackgerrit Stephen Finucane proposed openstack/python-novaclient master: Bump to hacking 2.x https://review.opendev.org/708832
10:13:57 openstackgerrit Lee Yarwood proposed openstack/nova master: WIP libvirt: Provide the backing file format when creating qcow2 disks https://review.opendev.org/708745
11:07:01 openstackgerrit Stephen Finucane proposed openstack/python-novaclient master: Don't print user_data for 'nova show' https://review.opendev.org/708850
11:16:07 ivve hello, i have a nova block device attachments stuck (i have been trying to get it removed with nova and openstack commands but with no avail) i can see it still in the nova.block_device_mapping table, any recommendations on setting it properly to deleted?
11:16:35 ivve other than removing the entire row
11:53:04 openstackgerrit Merged openstack/nova stable/queens: Mask the token used to allow access to consoles https://review.opendev.org/707845
12:50:56 gibi efried: I will attend the nova meeting today to talk about FFE but I hope we can make it quick as I have a parallel meeting at the same time.
12:56:45 sean-k-mooney gibi: is that for the local cache or delete on terminate spec or both
12:57:13 sean-k-mooney i reviewd the delete on terminate spec and other then some minor wording issue i think its good to go
12:57:34 sean-k-mooney i have not looked at the local caching one in a few days but ill try to take a look before the meeting
12:58:06 gibi sean-k-mooney: efried collecs opinions here https://etherpad.openstack.org/p/nova-ussuri-planning
12:58:39 gibi sean-k-mooney: for the local cache I think we can support that but we might not need to commit to it yet as there are os-brick and cinder dependencies to be implemnted first
12:58:42 sean-k-mooney ah righit i just comemnted on the spec ill add mine
12:58:57 sean-k-mooney gibi: ya
12:59:15 sean-k-mooney gibi: i think cinder were planning to approve there FFE today before the nova meeting
12:59:18 gibi sean-k-mooney: for the delte on terminate I saw an API problem raised from gmann that wasnot resolved
12:59:55 sean-k-mooney ah i did not know he had raised an issue
13:01:39 gibi sean-k-mooney: I'm trying to find the description of the problem
13:01:43 sean-k-mooney gibi: ya so i realy dont know hwy we dont just support setting delete on terminate via a put to /servers/{server_id}/os-volume_attachments/{volume_id}
13:02:27 gibi sean-k-mooney: https://review.opendev.org/#/c/580336/28/specs/ussuri/approved/destroy-instance-with-datavolume.rst@43
13:02:39 sean-k-mooney ^ was the open question i had. i assume PUT to that endpoing is how we trigger swap volumn?
13:02:49 gibi sean-k-mooney: because that PUT is for swap volume not for updating existing volume
13:03:22 sean-k-mooney ya that is what i was assuming
13:03:58 sean-k-mooney so we are just modifying the put then so that it only swaps the value if you alter the volume id
13:04:00 gibi so we made a suboptimal decision in the past to use PUT as swap and now we have no good place to implement a simple update as PUT is already used
13:04:06 sean-k-mooney which is why it is no optional in the spefc
13:04:24 sean-k-mooney gibi: well we could use patch
13:04:43 gibi sean-k-mooney: I could accept to have a conditional swap as you described above
13:04:49 gibi sean-k-mooney: I'm not sure gmann likes that
13:05:04 sean-k-mooney right well the spec currently had the conditional swap
13:05:20 sean-k-mooney was useing PATCH discussed?
13:06:20 sean-k-mooney PATCH is intended for a partial update where you jsut sent the data to be updated so PATCH would just contain delete_on_terminate=True|False in the body and use the volumn id for the url
13:06:38 sean-k-mooney gmann: ^ would you be ok with that?
13:06:55 sean-k-mooney gmann: it would allow put to be kept just for swap volumn
13:07:15 sean-k-mooney *volume
13:07:15 gibi sean-k-mooney: as far as I remember PATCH was not discussed
13:07:28 gibi sean-k-mooney: I looked at the current version of the spec. I'm OK wiht the conditional swap
13:09:01 gibi but I would not like to +2 the patch without hearing back from gmann
13:09:13 sean-k-mooney sure
13:34:17 openstackgerrit Balazs Gibizer proposed openstack/nova stable/pike: Mask the token used to allow access to consoles https://review.opendev.org/708876
13:35:05 openstackgerrit Balazs Gibizer proposed openstack/nova stable/pike: Mask the token used to allow access to consoles https://review.opendev.org/708876
13:35:20 openstackgerrit Victor Coutellier proposed openstack/nova master: Non-Admin user can filter their instances by more filters https://review.opendev.org/701609
13:47:32 efried stephenfin: regarding vTPM, is it something you would have time and desire to do?
13:50:55 gmann gibi: sean-k-mooney sounds good to me. my main concern is not to over-scope the existing swap API which makes our API hard/confusing to use. PATCH option looks good to me
13:51:00 gibi stephenfin: if you do the vTPM you can sign me up to review your patches
13:51:30 gibi gmann: so the current conditional PUT is not good for you but a separate PATCH would be good.
13:52:36 gmann gibi: yeah. because i also looked and none of the existing API i find suitable where we can add this feature.
13:53:05 stephenfin efried: Someone needs to do it, so yeah
13:53:12 gibi gmann: noted. Could you state that in the review with a -1?
13:53:15 gmann and as long as we keep that API very clear about updating only nova related/owned bits of volume not proxy to cinder
13:53:19 stephenfin I mean, I should have the bandwidth. It's just the reviews I'd need
13:53:23 gmann gibi: sure.
13:53:32 gibi gmann: thanks
13:53:35 stephenfin Story of my OpenStack life :)
13:54:09 gibi :)
13:55:16 efried stephenfin: okay, thanks for that. johnthetubaguy said it was important, and obviously jroll and the vmg folk want it, so...
13:56:38 sean-k-mooney efried: stephenfin for what its worth we want vtpm downstream as a medium to long term feature
13:57:01 efried sean-k-mooney: "we" == RH?
13:57:10 kashyap I think it might be the "royal we" :D
13:57:19 stephenfin in this case yeah
13:57:30 sean-k-mooney yep, we as in redhat. we dont need it in the next release but we expect customer two ask for it in the future
13:57:43 efried noted
13:58:03 efried Nova meeting in #openstack-meeting in ~2 mins
13:58:55 sean-k-mooney we have spoke to some goverment funded organisation in the eu where they were considering making tpm supprot a requirement in some cases. but its will be a whiel before anything like that comes into effect
14:00:42 efried lyarwood: fyi there's an agenda item to follow up about rocky EM
14:02:19 lyarwood efried: yup still not found time to do it sorry
14:02:26 lyarwood efried: rocky is at least passing now
14:02:36 efried no worries, just wanted an update for the record, will ask again officially :)
14:19:15 openstackgerrit Merged openstack/nova master: Use tempest-full-py3 as base job https://review.opendev.org/708548
14:21:37 openstackgerrit Merged openstack/nova stable/ocata: Use stable constraint for Tempest pinned stable branches https://review.opendev.org/706872
14:44:34 jroll stephenfin: <3 thanks for picking that up
14:51:58 alex_xu gmann: gibi lyarwood sean-k-mooney, I guess we need to show the 'delete_on_termination' in GET method also, otherwise I don't know how the user know the current value for delete_on_termination.
14:52:06 alex_xu brinzhang__: ^
14:52:40 sean-k-mooney am yes we could add that in the same microverion bump
14:52:55 alex_xu for PATCH, I just feel our existing PUT is behavior like PATCH, now we add another PATCH, then we add another mess :)
14:52:57 brinzhang__ alex_xu: agree, if the volume was speficied delete_on_termination, it shuold show
14:53:38 alex_xu I also remember ironic's patch method use json-patch in the body, gmann do you know is there any defintion from the api-wg for the patch's body?
14:53:42 sean-k-mooney alex_xu: well PUT required the entire resoce to be updated. its not sematiclly correct to only set some filed in a PUT
14:53:53 sean-k-mooney partail updates should have been using PATCH
14:54:15 alex_xu sean-k-mooney: oh, yea, right.
14:54:19 sean-k-mooney so the current way we use PUT for swap volume is not correct
14:54:29 alex_xu right
14:54:30 sean-k-mooney but we cant really change that at this point
14:54:35 gmann alex_xu: GET server has that.
14:55:32 gmann https://github.com/openstack/nova/blob/08b98d75cf4b77563e373709a84685f7d0b72ba3/nova/api/openstack/compute/views/servers.py#L716
14:55:43 sean-k-mooney kevinz: hi do you want to talk about arm testing here or on #openstack-qa
14:55:43 gmann i think that is from microversion 2.3
14:56:08 sean-k-mooney #openstack-qa might be better to let the volume discussion happen here

Earlier   Later