| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-22 | |||
| 09:20:04 | openstackgerrit | wanghao proposed openstack/nova master: Set min_disk in the image meta based on the root volume's size https://review.openstack.org/407397 | |
| 09:32:19 | ratailor | Can anybody pl help regarding this issue, can it be resolved by implementing command in nova-manage https://launchpad.net/bugs/1452641 | |
| 09:32:19 | openstack | Launchpad bug 1452641 in Cinder "Static Ceph mon IP addresses in connection_info can prevent VM startup" [Medium,Confirmed] - Assigned to Jon Bernard (jbernard) | |
| 09:39:03 | kashyap | stewie925: DevStack questions can be discussed on #openstack-qa or #openstack-dev | |
| 12:19:49 | openstackgerrit | Matt Riedemann proposed openstack/python-novaclient master: Remove deprecated image commands/API bindings https://review.openstack.org/436494 | |
| 12:19:49 | openstackgerrit | Matt Riedemann proposed openstack/python-novaclient master: Add functional test for "nova boot --image-with" https://review.openstack.org/436608 | |
| 12:30:17 | openstackgerrit | Matt Riedemann proposed openstack/python-novaclient master: Remove deprecated image commands/API bindings https://review.openstack.org/436494 | |
| 13:42:17 | ildikov | mriedem: when do you start this morning? | |
| 13:44:48 | mriedem | ildikov: 9ish? | |
| 13:45:06 | ildikov | mriedem: ok :) | |
| 13:45:14 | ildikov | Tnx | |
| 13:48:13 | openstackgerrit | Dan Smith proposed openstack/nova master: Make conductor create InstanceAction in the proper cell https://review.openstack.org/436649 | |
| 13:48:14 | openstackgerrit | Dan Smith proposed openstack/nova master: Make scheduler get hosts from all cells https://review.openstack.org/436648 | |
| 13:56:31 | openstackgerrit | Wang Qing wu proposed openstack/nova master: PowerVM Driver: localdisk https://review.openstack.org/409401 | |
| 14:02:15 | mriedem | https://etherpad.openstack.org/p/nova-ptg-pike | |
| 14:06:32 | ganesan | is it possible to boot vm from PXE server? | |
| 14:16:05 | gibi | mriedem, gmann, bauzas: I added our measured execution times for the proposed tempest tests in https://review.openstack.org/#/c/374887/. It about 35 seconds addition time for tempest | |
| 14:20:00 | ganesan | is it possible to boot instance from PXE server? | |
| 14:26:07 | bauzas | gibi: 35 seconds for testing a single filters seems uber-expensive to me | |
| 14:26:43 | bauzas | imagine that we have to test N filters multiplied by the number of tests we want to cover | |
| 14:27:00 | bauzas | and then I could be very worried by the order of magnitude it would be | |
| 14:27:50 | gmann | bauzas: gibi apart from timing, i need to know does scheduler filter promise exact behavior to user? | |
| 14:28:04 | bauzas | either way, I have that idea in my pipe for too long, I'll really work on implementing functional tests for that filter | |
| 14:28:52 | gmann | bauzas: gibi as bauzas mentioned yesterday if there re multiple filter and they will be applied as they appears in list so there is no guarantee that any filter in last will enforce the scheduling right? | |
| 14:29:26 | bauzas | gmann: you know how filters work, right? it's black boxes I/Oing a list of nodes and reducing it per filter | |
| 14:29:46 | bauzas | gmann: how you reduce that list is depending on the ordering you make | |
| 14:30:18 | bauzas | for example, say I have a 1000 nodes cloud | |
| 14:30:26 | bauzas | 999 of them are disabled | |
| 14:30:37 | gmann | bauzas: right and there is no full guarantee that aggregate filter will give you expected results if there are more other filters are before that | |
| 14:31:00 | bauzas | if I'm not putting ComputeFilter ahead, and I have 15 filters in my conf, I'll iterate over those 1000 nodes instead of just iterating once | |
| 14:31:19 | bauzas | gmann: exactly, because there is a dependency chain based on the order of your filters | |
| 14:31:33 | bauzas | that's what I said 'no isolation' yesterday | |
| 14:31:43 | bauzas | s/what/why | |
| 14:31:59 | gmann | bauzas: yea, i got your point now. | |
| 14:32:28 | bauzas | an ideal world would be in-tree Tempest testing like I said in the review | |
| 14:32:41 | gmann | and tempest can get false results till it go and tests each filter one by one and that is something less feasible | |
| 14:32:45 | bauzas | that would be triggered on purpose | |
| 14:32:51 | bauzas | gmann: yeah | |
| 14:33:10 | gmann | bauzas: in tree tempest tests you mean with tempest plugin? | |
| 14:33:46 | bauzas | I'm saying "in Nova functional tests that would run Tempest" | |
| 14:33:52 | gmann | bauzas: cinder has one and have some of the tempest like tests there | |
| 14:34:21 | gmann | bauzas: run tempest means? run with tempest ecosystem ? | |
| 14:35:33 | gmann | bauzas: anyways i agree with your idea, let's catch in room sometime during break time. | |
| 14:35:43 | bauzas | gmann: nevermind my comment, that's something I'm not expert on and I'd like to leave that for later | |
| 14:38:43 | gibi | bauzas, gmann: I get your point that in general testing every filter in tempest is not feasible and I don't want to do that. I just want to be able to prove about a deployed cloud if it separate tenants or not. | |
| 14:39:24 | gibi | we already have the tempesttest-plugin internally for this I just wanted to get rid of that downstream code | |
| 14:39:28 | diana_clarke | Is this the etherpad nova folks are using right now? https://etherpad.openstack.org/p/nova-ptg-pike Asking b/c I'm not seeing a lot of typing in it. | |
| 14:41:31 | gmann | gibi: sure, let discuss if we can solve both things here. | |
| 14:42:14 | cdent | diana_clarke: a) yes, b) lots of unstructurd talking happening right now, about hierarchical quotas | |
| 14:42:27 | diana_clarke | cdent: thx! | |
| 14:42:40 | cdent | bummer you're not here | |
| 14:43:08 | openstackgerrit | Daniel Pawlik proposed openstack/nova master: Changed network_info variable in test_create_configdrive https://review.openstack.org/436953 | |
| 14:44:40 | gibi | bauzas, gmann: how hard would it be to create and maintain a tempesttest-plugin for this kind of nova specific tests upstream? | |
| 14:45:54 | diana_clarke | cdent: I was going to send beer or cookies to the nova room, but I lost my visa last weekend ;( | |
| 14:46:01 | gmann | gibi: very easy :), i can setup one in 20 min and we keep adding tests there. | |
| 14:46:35 | cdent | diana_clarke: :( | |
| 14:46:39 | gmann | only things is from tempest side we enforce/encourage plugin to be in separate repo than project | |
| 14:47:18 | jsheeren | hi all, has anyone else noticed that for windows 2012 images, the spice vdagent uses up a full core? | |
| 14:47:21 | gmann | and that might be something we should think and discuss. | |
| 14:47:25 | bauzas | mriedem: am I blind or I can't see where we log the real exception as said in https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L3292-L3294 ? | |
| 14:47:32 | jsheeren | the vdagent log shows: write_completion::vio_serial write completion error 554 | |
| 14:47:36 | bauzas | mriedem: wrt https://bugs.launchpad.net/nova/+bug/1666148 | |
| 14:47:36 | openstack | Launchpad bug 1666148 in OpenStack Compute (nova) "nova set-password instancesname return 409 error" [Undecided,Incomplete] | |
| 14:48:00 | jsheeren | which shows me following bug: https://bugs.launchpad.net/nova/+bug/1634495 | |
| 14:48:01 | openstack | Launchpad bug 1634495 in OpenStack Compute (nova) "nova generate wrong spice channel type 'pty', which should be 'spicevmc'" [Undecided,In progress] - Assigned to m4cr0v (m4cr0v) | |
| 14:48:50 | gibi | gmann: thanks for the info. I agree that we need to discuss this further | |
| 14:49:54 | mriedem | bauzas: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L3289 would log the original exception | |
| 14:50:04 | mriedem | LOG.exception will log the trace for the original exception | |
| 14:52:34 | bauzas | mriedem: oh right | |
| 14:52:55 | bauzas | okay, we can ask for this trace then | |
| 14:53:26 | mriedem | bauzas: yeah i already did in the bug report :) | |
| 14:57:16 | bauzas | mriedem: I think the problem is that the reporter is using an internationalized version of Nova and tries to google translate the exception he has in order to have English back | |
| 15:06:35 | BobBall | mriedem: Are there expectations on nova drivers regarding ACTIVE status after image upload? (see my revert at https://review.openstack.org/435908) | |
| 15:09:51 | openstackgerrit | Diana Clarke proposed openstack/python-novaclient master: WIP - fix python-novaclient warning https://review.openstack.org/436975 | |
| 15:15:10 | mriedem | BobBall: how is that kwarg used in tempest? | |
| 15:15:30 | BobBall | mriedem: Default is true so if we are waiting for ACTIVE image we also wait for ACTIVE server | |
| 15:16:08 | BobBall | The only case that we set it to False by passing it in was where there was an explicit wait for ACTIVE server in the test case - so we didn't slow the test case down | |
| 15:17:03 | mriedem | totally missed that the default was True | |
| 15:19:11 | BobBall | I'm open on the ways we should try and fix this - if there should be an expectation on Nova drivers to ensure the server is ACTIVE before the image goes ACTIVE then we could look at that (although ATM I'm not sure how the implementation would work totally unless there was an artificial delay in Nova around setting the image ACTIVE until we had changed the server?) | |
| 15:21:31 | mriedem | +1 | |
| 15:21:36 | mriedem | on the patch | |
| 15:22:04 | BobBall | Ah, awesome, thanks. | |
| 15:26:55 | gmann | mriedem: BobBall thanks, yea i too did not notice the default value is true. | |
| 15:28:59 | claudiub | sfinucan: hello. I've reported a bug for the PCI resize thingy we've talked about: https://bugs.launchpad.net/nova/+bug/1666949 | |
| 15:28:59 | openstack | Launchpad bug 1666949 in OpenStack Compute (nova) "PCI devices cannot be detached if all the PCI devices are claimed" [Undecided,New] | |
| 15:33:31 | mriedem | dansmith: question inline https://review.openstack.org/#/c/434563/ | |
| 15:36:36 | mriedem | talking about https://review.openstack.org/#/c/433362/11/nova/context.py | |
| 15:43:39 | mriedem | sean-k-mooney: https://docs.openstack.org/developer/nova/cells.html#setup-of-cells-v2 | |
| 15:43:47 | mriedem | https://docs.openstack.org/developer/nova/cells.html#step-by-step-for-common-use-cases | |
| 15:45:00 | sean-k-mooney | mriedem: thanks ill crosspost this to the kolla etherpad for cells and multiregoin support | |
| 16:02:10 | openstackgerrit | Maciej Szankin proposed openstack/nova master: Transform instance.reboot notifications https://review.openstack.org/382959 | |
| 16:06:08 | openstackgerrit | Chris Friesen proposed openstack/python-novaclient master: Add support for showing flavor details (v2.43) https://review.openstack.org/435141 | |
| 16:21:59 | openstackgerrit | Eric Fried proposed openstack/nova master: PowerVM Driver: power_on/off and reboot https://review.openstack.org/427380 | |
| 16:36:27 | openstackgerrit | Dan Smith proposed openstack/nova master: Make CellDatabases fixture work over RPC https://review.openstack.org/434563 | |
| 16:36:27 | openstackgerrit | Dan Smith proposed openstack/nova master: Make servers API use cell-targeted context https://review.openstack.org/433362 | |
| 16:36:28 | openstackgerrit | Dan Smith proposed openstack/nova master: Make conductor create InstanceAction in the proper cell https://review.openstack.org/436649 | |
| 16:36:28 | openstackgerrit | Dan Smith proposed openstack/nova master: Make scheduler get hosts from all cells https://review.openstack.org/436648 | |
| 16:43:11 | openstackgerrit | Sean Dague proposed openstack/nova master: Verify project_id when quotas are checked https://review.openstack.org/435010 | |
| 16:44:32 | openstackgerrit | Roman Podoliaka proposed openstack/nova master: Make eventlet hub use a monotonic clock https://review.openstack.org/434327 | |
| 16:45:51 | openstackgerrit | Takashi NATSUME proposed openstack/nova-specs master: Abort Cold Migration https://review.openstack.org/334732 | |