Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-06-08
14:53:02 johnthetubaguy mriedem: any reason not to cache the availablity of a microversion, once we find its available?
14:54:53 cfriesen johnthetubaguy: couldn't they reboot the cinder service at any time?
14:55:11 cfriesen johnthetubaguy: or are you expecting monotonic increases
14:55:21 johnthetubaguy cfriesen: yeah, but we don't really support downgrades
14:55:40 johnthetubaguy cfriesen: your right in general though, just wondering if we can be simpler
14:56:14 johnthetubaguy actually, yeah, downgrading past this point just isn't going to work, as far as I understand it
14:58:35 openstackgerrit Stephen Finucane proposed openstack/nova master: objects: Add NovaComparableObject mixin https://review.openstack.org/472285
14:59:04 sfinucan bauzas, dansmith: What about that? ^
14:59:06 mriedem johnthetubaguy: replied inline
14:59:18 sfinucan johnthetubaguy: Oh hai
14:59:18 mriedem johnthetubaguy: in general because caching of versions gets complicated
14:59:32 mriedem especially when nova as the client might need to request multiple different versions based on which operation we're doing
14:59:42 mriedem like if we need 3.27 today for something but 3.33 tomorrow for something else
14:59:55 dansmith sfinucan: why not use equal_prims for that?
14:59:56 mriedem johnthetubaguy: for now i'm just keeping it simple
15:00:10 dansmith sfinucan: more "is this equal" helpers seems bad to me
15:00:16 johnthetubaguy mriedem: I was thinking caching was simpler, but you are right
15:00:26 mriedem it's simple in the simple case :)
15:00:29 johnthetubaguy mriedem: what you have is clearly correct
15:00:51 sfinucan dansmith: I was hoping you'd tell me :) I'm not confident enough in that code to say they're interchangeable
15:00:57 sfinucan If they are, I can take that route
15:01:07 mriedem johnthetubaguy: when talking about this last week, mordred was also saying that at some point i think ksa can provide this info, so we don't need the extra call to cinder
15:01:11 mriedem to get the version doc
15:01:15 mriedem ksa would just provide it
15:01:35 sfinucan (by "that code", I mean "all things object'y")
15:01:47 johnthetubaguy mriedem: good point, he is cooking things in that area, they can do the caching
15:01:49 mriedem johnthetubaguy: i.e. when we get the ksa session (which is cached...)
15:01:51 dansmith sfinucan: I'm saying if you want to have a mixin that implements __eq__, use the existing equal_prims thing to determine equality in your handler, don't define another one
15:02:08 mriedem johnthetubaguy: but thanks for reviewing, and welcome back
15:02:53 mordred mriedem: yes
15:02:56 mordred mriedem: this is correct
15:07:23 openstackgerrit Rick Bartra proposed openstack/nova master: Add policy granularity to the Flavors API https://review.openstack.org/449288
15:10:13 openstackgerrit Chris Friesen proposed openstack/python-novaclient master: 2.47: Show flavor info in server details https://review.openstack.org/435141
15:10:24 openstackgerrit Stephen Finucane proposed openstack/nova master: objects: Add NovaComparableObject mixin https://review.openstack.org/472285
15:10:28 mordred johnthetubaguy, mriedem: the ksa patches for all of this are under review, so hopefully should be able to take legit forward aciton in the next couple of weeks
15:10:30 sfinucan dansmith: Like that?
15:12:40 openstackgerrit Matt Riedemann proposed openstack/nova master: Add docstring for test_limit_check_project_and_user_zero_values https://review.openstack.org/472287
15:12:53 mriedem cdent: bauzas: melwitt: ^
15:16:23 cfriesen kashyap: apparently we have booted UEFI on Newton by specifying ‘--property hw_firmware_type=uefi” on the image and then booting as normal.
15:16:36 kashyap cfriesen: Thank you for confirming
15:16:45 kashyap cfriesen: I just checked with Fedora 25, and current Nova Git
15:17:01 kashyap cfriesen: I thought the Q35 machine type was _mandatory_ for OVMF
15:17:37 kashyap But, apparently it worked even with stock F25; which was QEMU 2.7 or something
15:19:09 cfriesen kashyap: we're using CentOS 7's qemu-kvm-ev-2.6.0-28.el7_3.9.1, so who knows what all they've done. :)
15:19:39 openstackgerrit sean mooney proposed openstack/os-vif master: set mtu on all code paths https://review.openstack.org/469951
15:20:13 mnaser stable/newton is broken with the new release of oslo.db that drops sqlite_db (it was removed in stable/ocata and beyond but not in stable/newton apparently): https://review.openstack.org/#/c/472290/
15:20:25 mnaser not sure why a notification didn't pop up here but if a small eye could be kept on that :)
15:20:39 kashyap cfriesen: Heh
15:21:49 mriedem mnaser: i'm not sure why something dropped in ocata would break newton,
15:21:55 mriedem since we should be using upper-constraints
15:22:55 mnaser mriedem: maybe it was my mistake, but pretty much i did a pip install from the stable/newton repo and there doesn't seem to be a limit for oslo.db in requirements.txt
15:23:11 mnaser so it downloaded the latest version of the library which has this commit from a few days ago => https://github.com/openstack/oslo.db/commit/cc64f7010522e812a696159009a72e59e2b017a8
15:23:16 mriedem mnaser: it'd be in tox.ini
15:23:43 mnaser mriedem ok i see it
15:23:56 mriedem mnaser: yeah that's 4.19.0, but we're not using that in newton https://github.com/openstack/requirements/blob/stable/newton/upper-constraints.txt#L215
15:24:15 mriedem https://github.com/openstack/nova/blob/stable/newton/tox.ini#L12 should have you covered
15:24:23 mnaser mriedem: sorry about that then, should i abandon that patch in that case?
15:24:46 openstackgerrit Alex Szarka proposed openstack/nova master: Refactor create_delete_server_with_instance_update https://review.openstack.org/466296
15:25:12 mriedem mnaser: yes please
15:25:47 mnaser mriedem: done, thanks for the lesson in requirements and sorry for noise
15:25:56 mriedem not a problem
15:28:01 cfriesen speaking of requirements...I'm trying to run current master devstack and it's giving me a series of version issues: ContextualVersionConflict: (os-brick 1.12.0 (/usr/local/lib/python2.7/dist-packages), Requirement.parse('os-brick>=1.13.1'), set(['cinder']))
15:28:11 cfriesen isn't it supposed to handle upgrading packages as needed automatically?
15:32:45 openstackgerrit Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for aggregates https://review.openstack.org/470932
15:39:52 mriedem melwitt: can we use an IN clause to delete all server group members using the list of instance uuids at once? https://review.openstack.org/#/c/457858/16
15:52:46 sfinucan melwitt: Could you take another look at https://review.openstack.org/#/c/379524/ today, please? :)
15:54:28 openstackgerrit Hironori Shiina proposed openstack/nova master: Ironic: Support boot from Cinder volume https://review.openstack.org/215385
15:56:29 openstackgerrit Merged openstack/python-novaclient master: Better handle key=value pair not being key=value https://review.openstack.org/469988
16:04:27 cfriesen re: my comment above, I just hit another one...neutron wants os-xenapi>=0.2.0. Shouldn't devstack handle the pip upgrade itself?
16:08:20 sdague johnthetubaguy: https://dogpilecache.readthedocs.io/en/latest/api.html#memory-backends if we wanted to go down that path
16:08:52 sdague cfriesen: RECLONE=True ?
16:09:10 sdague because if not you can hit requirements files out of sync with upper-constraints
16:12:29 cfriesen sdague: thanks, will give that a try. I thought that was the default, for some reason.
16:13:20 sdague no, because then it hard resets people's code
16:13:50 cfriesen wonder if I should try PIP_UPGRADE=True
16:17:13 sdague I don't think that does anything any more with constraints in play
16:19:34 mriedem kashyap: is there any reason that nova "needs" to set the 'serial' field in the disk config for the libvirt domain xml?
16:19:41 mriedem reading https://libvirt.org/formatdomain.html#elementsDisks
16:19:48 mriedem it says,
16:19:50 mriedem "If present, this specify serial number of virtual hard drive. For example, it may look like <serial>WD-WMAP9A966149</serial>. Not supported for scsi-block devices, that is those using disk type 'block' using device 'lun' on bus 'scsi'. Since 0.7.1"
16:20:04 mriedem however, we make no distinction of when we set that in the disk config, block type or not
16:20:17 mriedem and it's always the cinder volume id i believe
16:20:27 mriedem does qemu actually care about it?
16:21:31 johnthetubaguy sdague: yeah, good call, spied your comment on that
16:22:49 mriedem except avoid pickle
16:27:35 cfriesen sdague: with RECLONE=True I still hit it: ContextualVersionConflict: (pyroute2 0.4.13 (/usr/local/lib/python2.7/dist-packages), Requirement.parse('pyroute2>=0.4.15'), set(['neutron']))
16:28:31 sdague cfriesen: I don't know
16:28:51 sdague definitely going to need more than a single error log to know how you got there
16:29:10 sdague go through the log and see why there is a disconnect there
16:29:37 cfriesen sdague: http://paste.openstack.org/show/612007/
16:30:24 sdague cfriesen: and the value in /opt/stack/requirements/upper-constraints.txt
16:30:54 cfriesen sdague: pyroute2===0.4.15
16:31:47 sdague hmmm... idk
16:32:00 sdague I thought that was always supposed to override
16:35:03 cfriesen sdague: it choked on pyroute2 again even with PIP_UPGRADE=True
16:36:36 cfriesen I'm tempted to just blow everything away and start from scratch
16:37:37 sdague yeh, I'm rerunning locally to try to figure out if I see anything
16:38:02 sdague there could be a pip regression
16:45:03 openstackgerrit Merged openstack/nova master: Default to 0 when merging values in limit check https://review.openstack.org/471973
16:49:25 openstackgerrit Alex Szarka proposed openstack/nova master: Refactor create_delete_server_with_instance_update https://review.openstack.org/466296

Earlier   Later