| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-23 | |||
| 14:44:12 | openstackgerrit | Eric Fried proposed openstack/nova master: Monkey patch the blockdiag extension https://review.openstack.org/476159 | |
| 14:57:06 | openstackgerrit | Merged openstack/nova master: Allocate resources on forced dest host during live migration https://review.openstack.org/496031 | |
| 14:59:56 | clarkb | mriedem: did you see my comment about the cell1 cond logs? | |
| 15:05:59 | openstackgerrit | Merged openstack/nova master: Remove source node allocation after live migration completes https://review.openstack.org/496032 | |
| 15:06:32 | mriedem | clarkb: nope | |
| 15:07:17 | clarkb | mriedem: lwft it as acomment on the change but the format is funny in that gile so indexing wont work as is | |
| 15:07:27 | clarkb | *file | |
| 15:07:59 | mriedem | it's different from screen-n-cond? | |
| 15:08:43 | mriedem | oh it's just screen-n-cond-cell1? | |
| 15:08:44 | mriedem | weird | |
| 15:19:34 | openstackgerrit | Merged openstack/nova master: Restrict live migration to same cell https://review.openstack.org/496419 | |
| 15:20:07 | openstackgerrit | Merged openstack/nova master: Clarifying node_uuid usage in ironic driver. https://review.openstack.org/485803 | |
| 15:29:35 | cfriesen_ | is there a novaclient interface to do a PUT on /os-quota-class-sets ? | |
| 15:34:13 | dims | cfriesen_ : this one? http://git.openstack.org/cgit/openstack/python-novaclient/tree/novaclient/v2/quota_classes.py#n47 | |
| 15:37:28 | openstackgerrit | Merged openstack/nova master: Pull out code that builds VIF in _build_network_info_model https://review.openstack.org/466449 | |
| 15:37:29 | cfriesen_ | dims: thanks, that's the one. for some reason I'm blind. incidentally that API missing from openstackclient | |
| 15:38:08 | dims | cfriesen_ : hound is indispensable for me :) http://codesearch.openstack.org/ | |
| 15:42:13 | cfriesen_ | dims: oooh, I like it. | |
| 15:51:58 | stephenfin | mriedem, dansmith: ooh, should we deprecate the cells v1 nova-manage commands? https://github.com/openstack/nova/blob/master/nova/cmd/manage.py#L1030 | |
| 15:53:47 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add missing unit tests for FilterScheduler._get_all_host_states https://review.openstack.org/496803 | |
| 15:54:02 | mriedem | stephenfin: yes, should have done that when i deprecated cells v1 | |
| 15:54:18 | stephenfin | mriedem: I'll do that now so. RC2 material? | |
| 15:55:10 | mriedem | i don't know that it needs to be rc2, but realize it's kind of weird for a deprecation timer if the server side code is deprecated in pike and the CLI is deprecated in queens | |
| 15:55:42 | mriedem | you could still theoretically run the cli against an older installation that still has cells v1 code | |
| 15:55:47 | mriedem | in a venv or something | |
| 15:56:55 | stephenfin | True. I'll shove it up anyway. Can "backport" if it makes sense | |
| 15:57:01 | mriedem | yeah, thanks | |
| 16:01:48 | clarkb | mriedem: ya just that one service has a weird log, once its fixed I can approve the change to index it | |
| 16:02:19 | mriedem | ok i'll investigate | |
| 16:08:49 | openstackgerrit | Stephen Finucane proposed openstack/nova master: nova-manage: Deprecate 'cell' commands https://review.openstack.org/496815 | |
| 16:09:14 | stephenfin | mriedem, dansmith: ^ | |
| 16:09:57 | stephenfin | I'd opt for bringing it into Pike, but only because it makes cliff migration in Queens simpler. Good to get rid of it either way | |
| 16:11:32 | mriedem | stephenfin: comments inline | |
| 16:11:34 | mriedem | dansmith: ^ | |
| 16:16:47 | openstackgerrit | Stephen Finucane proposed openstack/nova master: nova-manage: Deprecate 'cell' commands https://review.openstack.org/496815 | |
| 16:17:00 | stephenfin | mriedem: I think that should do it ^ | |
| 16:19:04 | sean-k-mooney | stephenfin: regarding portbinding i dont think there is a neutron that is still supported that does not use portbining | |
| 16:19:35 | stephenfin | sean-k-mooney: Yeah, seems the decision is that we expect portbinding to be there be default now | |
| 16:20:09 | sean-k-mooney | stephenfin: i can ask on the neutorn channel but if you are using ml2 it is required. it may be posible to not use it if you use a monolitic plugin | |
| 16:20:31 | sean-k-mooney | monolitic plugins are effectivly deprected | |
| 16:21:06 | sean-k-mooney | to be part of the neutron stadium i belive you have to intergrate via ml2. that was a requirement for ovn anyway | |
| 16:30:23 | openstackgerrit | Merged openstack/os-vif master: Using assertIsNone() instead of assertEqual(None) https://review.openstack.org/496066 | |
| 16:56:09 | openstackgerrit | Lajos Katona proposed openstack/nova master: Add functional migrate force_complete test https://review.openstack.org/496202 | |
| 16:58:09 | openstackgerrit | Merged openstack/nova master: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | |
| 17:01:30 | openstackgerrit | Lajos Katona proposed openstack/nova master: Add functional migrate force_complete test https://review.openstack.org/496202 | |
| 17:47:30 | mriedem | stephenfin: +2 on the cell v1 nova-manage command deprecations, nice and generic wording there | |
| 17:48:02 | mriedem | would like dansmith's opinion on backporting for rc2, but i don't really think it's all that necessary, but we could if we want | |
| 17:56:40 | dansmith | mriedem: don't we need it in pike to call it actually deprecated? or are you just assuming your blanket cellsv1 deprecation covers it, even if not quite specific enough? | |
| 18:01:45 | mriedem | the latter | |
| 18:02:18 | mriedem | you could point the cli at an older deployment that's using cellsv1 still | |
| 18:02:29 | mriedem | not sure why anyone would, but i don't see a huge reason to backport | |
| 18:02:51 | mriedem | i also don't know if we'll remove cells v1 in queens or not | |
| 18:05:26 | dansmith | sure, okay | |
| 18:14:57 | zen369 | ERROR nova NotAllowed: Connection.open: (530) NOT_ALLOWED - access to vhost 'nova' refused for user 'nova' | |
| 18:15:01 | zen369 | any help? | |
| 18:46:39 | knikolla | hi all. We have a instance in error state after being live migrated to a server which doesn't have enough memory to start it. how can i migrate that to another host? | |
| 18:52:35 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add missing tests for _remove_deleted_instances_allocations https://review.openstack.org/496847 | |
| 18:53:43 | mriedem | knikolla: cold migrate it? | |
| 18:53:58 | mriedem | you can cold migrate from stopped state | |
| 18:54:30 | knikolla | mriedem: i stopped it and ran `openstack server migrate <id>` | |
| 18:54:34 | knikolla | after that it goes back to error | |
| 18:54:37 | knikolla | saying it's not running | |
| 18:55:21 | mriedem | what says it's not running? | |
| 18:55:27 | mriedem | the api, compute service, other? | |
| 18:55:53 | knikolla | mriedem: http://paste.openstack.org/show/nfBmtEhE7SScwd0EimVQ/ | |
| 18:59:02 | mriedem | that doesn't tell much - but assuming something failed on the compute service side | |
| 18:59:08 | mriedem | which is why it's recorded as an instance fault | |
| 19:00:49 | knikolla | checking compute logs | |
| 19:01:23 | knikolla | there's a InterfaceDetachFailed: Failed to detach network adapter device from c7703251-96d6-4a2e-b055-7c204ab104e7 | |
| 19:12:20 | openstackgerrit | Chris Dent proposed openstack/nova-specs master: Add a spec for minimal cache headers in placement https://review.openstack.org/496853 | |
| 19:15:36 | cdent | mriedem: do you recall when sdague comes back on the scene? I’d like to get his input on the microversion-parse adjustments I’ve got in progress | |
| 19:15:55 | mriedem | cdent: next week | |
| 19:16:02 | cdent | ✔ | |
| 19:16:11 | mriedem | his wood paneled station wagon can tour the country only so long | |
| 19:16:45 | mriedem | https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSIjcZl5KWDWlIYAkdWMwyv-yuZYuBUFN8RRc0MzWUwOfn_Uwfs | |
| 19:16:56 | mriedem | ^ is sdague last week | |
| 19:17:32 | cdent | ah the grand ol family truckster | |
| 19:17:35 | cdent | a fine fine auto | |
| 19:18:47 | dansmith | mriedem: he is in a station wagon, but I'm sure it doesn't have tree flesh on it | |
| 19:18:56 | dansmith | mriedem: and it might be powered by love and french fry oil | |
| 19:19:14 | mriedem | ha | |
| 19:19:28 | mriedem | subaru? | |
| 19:19:33 | knikolla | mriedem: the easy fix was to live-migrate another vm which was fine to make enough space for that to start, haha | |
| 19:19:35 | dansmith | yeah | |
| 19:19:50 | mriedem | knikolla: whatever works :) | |
| 19:46:36 | mriedem | now that i'm looking at the cold migration code, i see something odd | |
| 19:46:46 | mriedem | besides the fact it's called resize | |
| 19:47:15 | exarr | Hi all. I have applied the patch as noted here: https://review.openstack.org/#/c/428481/ to allow me to create volumes now, which is great, and seems to work. | |
| 19:47:19 | mriedem | i don't see how this would ever happen https://github.com/openstack/nova/blob/f1f55d48cc753ef3b6ab9b626148dec06e02f8cd/nova/compute/manager.py#L3773 | |
| 19:47:26 | mriedem | added in https://review.openstack.org/#/c/73387/ | |
| 19:47:33 | exarr | However, i'm stuck now with the following error: Exception during message handling: CantStartEngineError: No sql_connection parameter is establ | |
| 19:47:36 | exarr | ished | |
| 19:47:48 | mriedem | that exception is raised from driver.migrate_disk_and_power_off which is called from resize_instance, which is on the source host, after the dest host does and rpc cast from _prep_resize | |
| 19:51:46 | mriedem | there were like 10 people reviewing that change though so i'm not sure how they would all miss that | |
| 19:57:58 | exarr | mriedem: I assume you're not talking to me? :-) | |
| 19:58:03 | mriedem | correct | |
| 20:00:37 | cdent | mriedem: maybe something else (non-local) changed in the interim? | |
| 20:01:11 | mriedem | it has been 3 years, but looking at the code at the time that change was added, it's basically the same code | |
| 20:06:29 | dansmith | mriedem: did you figure out anything with the oslo timeutils thing yesterday? | |
| 20:06:52 | dansmith | I'm getting a weird native/naive comparison failure in unit tests when loading an object | |