| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-11-09 | |||
| 13:24:28 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/ocata: WIP: Functional test for regression bug #1713783 https://review.openstack.org/505160 | |
| 13:24:29 | openstack | bug 1713783 in OpenStack Compute (nova) ocata "After failed evacuation the recovered source compute tries to delete the instance" [High,Triaged] https://launchpad.net/bugs/1713783 | |
| 13:26:56 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/ocata: Functional test for regression bug #1713783 https://review.openstack.org/505160 | |
| 13:26:57 | openstack | bug 1713783 in OpenStack Compute (nova) ocata "After failed evacuation the recovered source compute tries to delete the instance" [High,Triaged] https://launchpad.net/bugs/1713783 | |
| 13:42:40 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/ocata: Set error state after failed evacuation https://review.openstack.org/518733 | |
| 13:43:43 | jianghuaw | jaypipes, are you around? | |
| 14:00:46 | jianghuaw | jaypipes, do you have time to have a quick look at the comment at here:https://review.openstack.org/#/c/518294/1/nova/compute/resource_tracker.py@795 | |
| 14:01:33 | jianghuaw | basically I think we can remove _resource_change() which is to check if resources changed in compute_node. | |
| 14:02:28 | openstackgerrit | Marco Chiappero proposed openstack/nova master: Add a communication channel for power management support https://review.openstack.org/512815 | |
| 14:06:50 | jianghuaw | jaypipes, I guess we defer the checking to self._provider_tree.has_inventory_changed: https://github.com/openstack/nova/blob/master/nova/scheduler/client/report.py#L602 | |
| 14:07:24 | jianghuaw | If my understand is correct, I think we can totally resource_tracker.py@remove _resource_change() | |
| 14:08:06 | jianghuaw | totally resource_tracker.py@remove _resource_change() | |
| 14:09:12 | jianghuaw | sigh! => totally remove resource_tracker.py@_resource_change() | |
| 14:33:49 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Functional test: cold migrate to compute down https://review.openstack.org/496280 | |
| 14:34:26 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Functional test: evacuate with no compute https://review.openstack.org/498482 | |
| 14:34:35 | efried | jaypipes https://github.com/jaypipes/articles/pull/17 | |
| 14:36:28 | jaypipes | jianghuaw: I'm here now. | |
| 14:38:18 | jianghuaw | jaypipes, do you think we can remove resource_tracker.py@_resource_change()? | |
| 14:38:22 | jianghuaw | and defer the checking to self._provider_tree.has_inventory_changed: https://github.com/openstack/nova/blob/master/nova/scheduler/client/report.py#L602 | |
| 14:38:51 | jaypipes | jianghuaw: we should be able to do that, yes. | |
| 14:39:38 | jianghuaw | jaypipes, cool. I will rework this patch https://review.openstack.org/#/c/518294 to cover this part. | |
| 14:40:09 | jaypipes | jianghuaw: sounds good. | |
| 14:42:37 | jianghuaw | jaypipes, is nested resource provider supported by allocation_candidate? | |
| 14:42:41 | jianghuaw | I guess not. | |
| 14:43:14 | jaypipes | jianghuaw: no, not yet. I am trying my best to get this series merged before having n-r-p do allocation candidates: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:refactor-alloc-cands-get-by-filters | |
| 14:43:17 | jianghuaw | But I'm not if we already have in-progress patches for that. | |
| 14:43:29 | jaypipes | jianghuaw: I would hold off a week | |
| 14:43:49 | jaypipes | jianghuaw: besides reviews from alex_xu, haven't been able to get many reviews on the series this week due to the summit. | |
| 14:44:01 | jaypipes | alex_xu: ty, btw for your reviews and your patience with me :) | |
| 14:44:26 | jianghuaw | jaypipes, understood. I will have a look the series of patches also. thanks. | |
| 14:44:34 | jaypipes | jianghuaw: ty sir! | |
| 14:45:49 | jianghuaw | :-) | |
| 14:49:52 | openstackgerrit | David Wahlstrom proposed openstack/nova stable/pike: Fix changes-since search filter https://review.openstack.org/518740 | |
| 15:00:22 | efried | jaypipes Would like to discuss a couple of design points with you when you have a sec. | |
| 15:00:31 | jaypipes | efried: shoot. | |
| 15:00:56 | jaypipes | efried: hangout? | |
| 15:01:11 | efried | jaypipes Could do. | |
| 15:01:36 | efried | jaypipes You'll have to set it up, though. | |
| 15:01:42 | jaypipes | efried: np, one sec. | |
| 15:02:34 | jaypipes | efried: https://hangouts.google.com/call/NveOHfR-Vd1PJXV8V2SEAAEE | |
| 15:02:48 | efried | jaypipes http://paste.openstack.org/show/625946/ | |
| 15:09:58 | efried | jaypipes Here's what that guy is *currently* returning: http://paste.openstack.org/show/625947/ | |
| 15:26:02 | efried | jaypipes https://review.openstack.org/#/c/518123/ <= bottom of test refactor & adds | |
| 15:46:52 | snasir | Hi is there a way to delete a nova hypervisor manually ? I deleted a node but something went wrong and the hypervisor did not get removed... Can anybody assist . Thanks | |
| 17:07:26 | efried | jaypipes About to push that test case into https://review.openstack.org/#/c/517119/6 -- you got local revs on that series? | |
| 17:07:54 | jaypipes | efried: nope, go for it. | |
| 17:08:10 | openstackgerrit | Eric Fried proposed openstack/nova master: handle traits with sharing providers https://review.openstack.org/517119 | |
| 17:08:13 | efried | jaypipes hecho | |
| 17:08:55 | jaypipes | efried: -1 on the bottom patch of your test series... | |
| 17:11:21 | efried | jaypipes Responded - I didn't change it. | |
| 17:12:36 | jaypipes | efried: ugh... | |
| 17:12:53 | efried | Do I get the sense that the base class shouldn't have been NoDB? | |
| 17:13:05 | jaypipes | efried: that reminds me of why I didn't like how these func tests were written in the first place... | |
| 17:13:12 | jaypipes | efried: yes. :( | |
| 17:13:20 | jaypipes | efried: ok, never mind. change in future patch. | |
| 17:13:26 | efried | ack | |
| 17:14:39 | efried | jaypipes I got the impression that the fixture was mocking the database-ness in some way different than the main TestCase base . | |
| 17:15:14 | jaypipes | efried: I hate the whole USES_DB_SELF crap. | |
| 17:15:28 | jaypipes | along with NoDBTestCase... just doesn't make logical sense. | |
| 17:17:22 | efried | jaypipes Well, FWIW, NoDB+USES_DB_SELF seems to run 19-23% faster (CPU vs. wallclock) | |
| 17:17:37 | efried | (sample size: 1) | |
| 17:17:58 | jaypipes | efried: I just hate the incongruence. | |
| 17:18:02 | efried | mm | |
| 17:20:52 | efried | jaypipes Appears same effect could be achieved by inheriting from TestCase, setting USES_DB = False, and also setting USES_DB_SELF = True. | |
| 17:21:14 | openstackgerrit | Murali Annamneni proposed openstack/nova master: Enables MySQL Cluster Support for Nova https://review.openstack.org/446643 | |
| 17:21:29 | jaypipes | efried: USES_DB_SELF is a terribly-named class variable. | |
| 17:21:37 | jaypipes | efried: I mean, wtf does that even mean? :( | |
| 17:22:15 | efried | USES_a_DB_that_i_will_configure_mySELF_so_dont_do_it_for_me ? | |
| 17:24:08 | efried | The code that consumes those vars looks pretty simple; *and* they appear to be mutually exclusive, so the base class(es) could toggle the one based on the other. Then we could come up with better names, and jaypipes is a happy bear. | |
| 17:24:20 | efried | bbiab | |
| 17:40:18 | jaypipes | efried_afk: number of requests for you in https://review.openstack.org/#/c/518124/ | |
| 18:13:46 | snasir | Hi Nova people. I have this error in Nova-conductor logs (using Newton Release), when I try to spawn a baremetal node . Changed the compute_driver to ironic driver and ironic_host_manager in nova.conf to do so. https://thepasteb.in/p/Vmh0yPQz56GC4 | |
| 18:14:15 | snasir | Desperately need some help. I dont see any other logs to indicate why nova cannot schedule.. Any help would be greatly appreciated, Thanks | |
| 18:18:22 | snasir | Is anyone here ? | |
| 18:33:56 | snasir | Anybody ? | |
| 18:36:36 | openstackgerrit | Mike Lowe proposed openstack/nova master: live-mig: keep disk device address same https://review.openstack.org/518022 | |
| 18:39:37 | jaypipes | snasir: not many people here due to people being in Sydney at the OpenStack summit... | |
| 18:40:01 | jaypipes | snasir: have you asked on the #openstack-ironic channel and/or the openstack@ mailing list? | |
| 18:41:35 | jaypipes | snasir: compute-0: (ComputeCapabilitiesFilter) capability is missing | |
| 18:41:52 | jaypipes | snasir: there is a capability that compute-0 doesn't have that the flavor you are booting requires. | |
| 19:00:42 | snasir | jaypipes: oh ok, No the problem is with the other node, which just fails at Retry | |
| 19:07:57 | jaypipes | snasir: I'm telling you that the logs are pointing out why a NoValidHost is being returned. Compute-0 is being filtered out of the results because it's missing a required capability. | |
| 19:41:38 | openstackgerrit | Elod Illes proposed openstack/nova master: Transform instance-live_migration_pre notification https://review.openstack.org/482070 | |
| 20:13:48 | efried | jaypipes Module-level using rp.context -- does rp have a .context or just a ._context? | |
| 20:14:03 | jaypipes | efried: yeah, sorry, rp._context | |
| 20:14:19 | efried | jaypipes You're okay with me using a private member for the sake of making the method module-level? | |
| 20:14:27 | jaypipes | efried: yes sir | |
| 20:14:29 | efried | ight. | |
| 20:18:42 | openstackgerrit | Merged openstack/nova master: Extract allocation candidates functional tests https://review.openstack.org/518123 | |
| 20:28:23 | jmlowe | jaypipes: got a sec? | |
| 20:28:36 | jaypipes | jmlowe: yessir | |
| 20:31:13 | jmlowe | so I gave that test another try, I think it should work but it doesn't seem to update the xml like I expect it to, is there some obvious gross conceptual error on my part? | |
| 20:31:15 | jmlowe | http://logs.openstack.org/22/518022/7/check/openstack-tox-py27/104cb93/testr_results.html.gz | |
| 20:37:24 | jaypipes | jmlowe: looking.... | |
| 20:45:28 | jaypipes | jmlowe: gawd I hate dealing with XML... :( the only thing I'm guessing is maybe because the host list is a sub-element one level further down from the address element, that maybe the copy-destination-element thing isn't working properly? Try changing the alias element, which *is* at the same level as the address element and seeing if that works. | |
| 20:46:50 | jmlowe | on the upside if it can't do nested then it's probably a bug, there are certainly storage systems where the sub elements can and should change over time | |
| 20:55:58 | jaypipes | jmlowe: ack. I'm just grasping at straws really :) | |
| 20:56:22 | jmlowe | I'm trying to find where the alias is generated | |
| 21:08:00 | jmlowe | so I can't actually locate where the alias is generated, but I'm thinking it's generated from the address which isn't supposed to change | |
| 21:36:49 | openstackgerrit | Marco Chiappero proposed openstack/nova master: Add a communication channel for power management support https://review.openstack.org/512815 | |