| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-10-04 | |||
| 06:50:13 | pooja_jadhav | gmann: yes, same point i am able to see that, I need to use Cinder Fixture for the same. Tried same thing which is done at [1]https://github.com/openstack/nova/blob/master/nova/tests/functional/api_sample_tests/test_block_device_mapping_boot.py#L23-L27 | |
| 06:51:31 | gmann | pooja_jadhav: yeah, any error you are getting? | |
| 06:51:35 | gmann | you have patch up | |
| 06:53:48 | pooja_jadhav | gmann: I have added same stubs in the https://github.com/openstack/nova/blob/master/nova/tests/functional/api_sample_tests/test_simple_tenant_usage.py and trying to add 4th as BFV instance. | |
| 06:55:34 | pooja_jadhav | gmann: here is the error log http://paste.openstack.org/show/731449/ | |
| 07:00:14 | gmann | pooja_jadhav: this is going on cinder call? did you use Cinder Fixture ? | |
| 07:00:36 | gmann | pooja_jadhav: if you can push your patch in gerrit (WIP though), i can take a look | |
| 07:01:07 | pooja_jadhav | gmann: yes, In setup method, I have added this line : self.useFixture(nova_fixtures.CinderFixture(self)) | |
| 07:01:07 | pooja_jadhav | self.useFixture(nova_fixtures.CinderFixtureNewAttachFlow(self)) | |
| 07:01:58 | gmann | pooja_jadhav: ok, and you are just calling server create with bdmv2 right | |
| 07:02:20 | pooja_jadhav | gmann: yes, right | |
| 07:04:24 | pooja_jadhav | gmann: Also I checked, v2.32 and v2.42, in https://github.com/openstack/nova/blob/master/nova/tests/functional/api_sample_tests/test_servers.py#L189-L196 they are created BFV instances. they are working perfectly fine. Same things trying to use. but not getting success | |
| 07:04:56 | bauzas | good morning nova | |
| 07:06:31 | pooja_jadhav | gmann: How to use this Cinder Fixture?? Am I using correctly in the setup method of the class where I need it ?? | |
| 07:07:10 | gibi | mriedem, efried: I've read back on yesterday's concurrent update. I think one of the goals of the consumer gen patches to make the concurrent update visible and not overwrite allocations so I feel what mriedem saw was a result of that | |
| 07:09:26 | gibi | mriedem, efried: also I agree with efried that if the last patch that updates claim_resources is merged then that parallel allocation request case will stop eariler in the claim_resources when the second claim_resources call for the same consumer fails on the assumption that the consumer does not exists | |
| 07:11:44 | gmann | pooja_jadhav: i think so. but somehow it is not considering it. | |
| 07:12:57 | gmann | pooja_jadhav can you paste the test_simple_tenant_usage.py or if you can push patch up in gerrit then it is easy to debug | |
| 07:22:30 | pooja_jadhav | gmann: http://paste.openstack.org/show/731450/ | |
| 07:29:17 | gmann | pooja_jadhav: you do not need to use both, use CinderFixture only | |
| 07:31:06 | pooja_jadhav | gmann: I have trued by using used CinderFixture only. But still no success L( | |
| 07:31:08 | pooja_jadhav | :( | |
| 07:31:15 | pooja_jadhav | tried* | |
| 07:38:00 | gmann | pooja_jadhav: and the sample file path you changed ? | |
| 07:39:38 | pooja_jadhav | gmann: I am adding sample files inside https://github.com/openstack/nova/tree/85b36cd2f82ccd740057c1bee08fc722209604ab/nova/tests/functional/api_sample_tests/api_samples/os-simple-tenant-usage/v2.40 | |
| 07:40:26 | pooja_jadhav | and passing use_common_server_api_samples=False to _post_server() method so that, It will get the samples from api_samples/os-simple-tenant-usage/v2.40 only. | |
| 07:43:33 | gmann | pooja_jadhav: that's all you are doing perfectly | |
| 07:44:51 | pooja_jadhav | gmann: yes, I will try my best for this. If anything u think I am missing something then please let me know. :) | |
| 07:46:07 | gmann | pooja_jadhav: ok, but if you can push WIP patch up then it will be easy otherwise, it is difficult to debug | |
| 07:50:28 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Consider nested allocations during allocation cleanup https://review.openstack.org/606050 | |
| 07:50:28 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Ignore forcing of evacuation for nested instance https://review.openstack.org/606111 | |
| 07:50:29 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Run negative server moving tests with nested RPs https://review.openstack.org/604125 | |
| 08:16:26 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Update --max-rows parameter description for archive_deleted_rows https://review.openstack.org/606995 | |
| 08:43:13 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Modify get_by_cell_and_project() to get_not_qfd_by_cell_and_project() https://review.openstack.org/607663 | |
| 08:43:13 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Return a minimal construct for nova list when a cell is down https://review.openstack.org/567785 | |
| 08:55:58 | tssurya | sean-k-mooney: good morning, around ? | |
| 08:58:11 | tssurya | its about the exception handling for the scatter utility stuff we discussed yesterday, would you be modifying your existing patch or should I put a new one for the simple approach ? asking just to rebase the series | |
| 09:08:56 | gmann | pooja_jadhav: seems like you need to override the nova.volume.cinder.create which is not in conderFixture | |
| 09:09:12 | gmann | t i tested with that and it worked fine | |
| 09:12:02 | gmann | pooja_jadhav: i mean stub | |
| 09:13:28 | tssurya | sean-k-mooney: nvm, my bad just saw http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2018-10-03.log.html#t2018-10-03T19:17:25 | |
| 09:16:23 | sean-k-mooney | am i am actully im just making coffee | |
| 09:16:59 | sean-k-mooney | this is around the time in normally wake up and start being online | |
| 09:18:31 | sean-k-mooney | tssurya: for the simple approch we agreed to just retrun the exception by construction a new copy and leave a comment saying we are doing that because of https://www.python.org/dev/peps/pep-0344/#open-issue-garbage-collection | |
| 09:19:13 | sean-k-mooney | tssurya: so expect execption e: retrun e.class(e.args) | |
| 09:19:44 | tssurya | sean-k-mooney:ack thanks :) sorry for the early ping then | |
| 09:19:56 | tssurya | I was calculating 1hr from behing here | |
| 09:20:19 | tssurya | 1hr behind the time in CEST I mean | |
| 09:20:33 | sean-k-mooney | haha no worries. i shift my day to give me more overlap with the us also im not a morning person :) | |
| 09:21:34 | tssurya | I'll put something up then for the spec's usecase and add you as a reviewer :) | |
| 09:21:37 | tssurya | thanks again | |
| 09:29:27 | sean-k-mooney | bauzas: are you online? | |
| 09:33:05 | bauzas | sean-k-mooney: I am, but I'll need to get my daughters in 10 mins | |
| 09:33:36 | sean-k-mooney | no worries we can chat when you get back its about vgpus | |
| 09:36:11 | bauzas | sean-k-mooney: ok, for what ? | |
| 09:36:59 | sean-k-mooney | i want to know how libvirt/nova detect something is a gpu | |
| 09:37:15 | sean-k-mooney | also i have some good news for you | |
| 09:37:47 | bauzas | sean-k-mooney: we don't really detect whether it's a gpu | |
| 09:38:07 | bauzas | sean-k-mooney: we just say "heh, look, there are some pci devices that support mdevs" | |
| 09:38:10 | sean-k-mooney | bauzas: mnaser: is amazing and is going to provide 1-2 vm instance with gpus for testing in the gate https://review.openstack.org/#/c/607686/ | |
| 09:38:22 | bauzas | cool | |
| 09:38:30 | bauzas | so I could write a functional test | |
| 09:38:42 | sean-k-mooney | bauzas: ok that is a bug but also cool because i think i konw how to fake them in the gate also | |
| 09:38:54 | bauzas | sean-k-mooney: what's a bug ? | |
| 09:39:03 | bauzas | to say a pci device supporting mdevs is a gpu ? | |
| 09:39:12 | sean-k-mooney | nic and acclerator can expose mdevs | |
| 09:39:13 | bauzas | sure, but we have a spec fixing this | |
| 09:40:03 | bauzas | sean-k-mooney: in this spec, you explicitly say which PCI IDs *are* GPUs | |
| 09:40:06 | bauzas | sean-k-mooney: https://review.openstack.org/#/c/602474/ | |
| 09:40:22 | bauzas | I need to go, bbiabn | |
| 09:40:33 | sean-k-mooney | bauzas: cool that will work for the terible hack i want to do. ill let you know when you get back | |
| 09:51:10 | gmann | pooja_jadhav: hope that worked. | |
| 10:12:00 | bauzas | sean-k-mooney: just able to discuss for 5 mins | |
| 10:12:20 | bauzas | sean-k-mooney: but also, without the spec, it also only works for GPU devices | |
| 10:12:45 | bauzas | sean-k-mooney: because libvirt provides us all the mdev capable devices, but we look at which supports the provided GPU type | |
| 10:12:51 | bauzas | say 'nvidia-11' | |
| 10:13:27 | bauzas | so, in case you have mdev capable devices that are not GPU devices (say a nic one), they don't support the type | |
| 10:13:33 | bauzas | so we don't use them | |
| 10:14:03 | bauzas | also, say you have an Intel GPU not supporting the 'nvidia-11' type, then we don't have inventory for it too | |
| 10:15:06 | sean-k-mooney | ok so your assumeing good nameing of the mdev types without the spec | |
| 10:15:16 | sean-k-mooney | e.g. not small medium large | |
| 10:15:28 | sean-k-mooney | but ya i think that is fine | |
| 10:15:59 | sean-k-mooney | basically i found out that in kernel 4.16 there is a new driver called netdevsim that simulates a nic | |
| 10:16:15 | sean-k-mooney | it supprots sriov without any hardware and mdevs | |
| 10:16:41 | sean-k-mooney | so i think i can use it to test pci passhtrouhg, sriov, and other things in the gate | |
| 10:17:00 | sean-k-mooney | i also may be able to use the mdev support to pretend to be a vgpu | |
| 10:17:32 | sean-k-mooney | not sure about the last point but i was planning to look into it more for sriov testing | |
| 10:22:34 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: api-ref: Move the evacuate action to admin action https://review.openstack.org/607896 | |
| 11:46:42 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: consumer gen: support claim_resources https://review.openstack.org/583667 | |
| 11:46:43 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Enable nested allocation candidates in scheduler https://review.openstack.org/585672 | |
| 11:46:44 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Use provider tree in virt FakeDriver https://review.openstack.org/604083 | |
| 11:49:34 | gibi | jaypipes, efried: I replied in https://review.openstack.org/583667 for both of you fixed your comments. The runway slot for this patch ends today so I appreciate any feedback | |
| 11:49:42 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: [WIP] Refactor scatter-gather utility to return exception objects https://review.openstack.org/607934 | |
| 11:52:29 | jaypipes | gibi: yup, on it. | |
| 11:52:41 | gibi | jaypipes: thanks a lot | |
| 11:54:17 | bauzas | sean-k-mooney: back there | |
| 11:54:25 | bauzas | sean-k-mooney: if you wanna continue to discuss... | |
| 11:54:54 | bauzas | gibi: don't worry about the runway slot, now that I'm done with the reshaper change, your series is my top prio | |
| 11:55:10 | gibi | bauzas: cool, thanks | |
| 11:55:42 | bauzas | gibi: so, the above means that I can review it, right? | |