Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-07-12
12:37:41 openstackgerrit Alex Xu proposed openstack/nova master: Remove the test for the route '/resources.:(format)' https://review.openstack.org/478900
12:37:42 openstackgerrit Alex Xu proposed openstack/nova master: Use plain routes list for image-metadata instead of stevedore https://review.openstack.org/478499
12:37:42 openstackgerrit Alex Xu proposed openstack/nova master: Use plain routes list for server-security-groups instead of stevedore https://review.openstack.org/478902
12:37:43 openstackgerrit Alex Xu proposed openstack/nova master: Use plain routes list for images instead of stevedore https://review.openstack.org/478498
12:37:43 openstackgerrit Alex Xu proposed openstack/nova master: Use plain routes list for os-volumes instead of stevedore https://review.openstack.org/479497
12:37:44 openstackgerrit Alex Xu proposed openstack/nova master: Use plain routes list for os-security-group-default-rules instead of stevedore https://review.openstack.org/478904
12:45:31 alex_xu nova api meeting is in 15 mins at #openstack-meeting-4
12:50:32 claudiub moshele: hello. around? :)
12:51:04 moshele claudiub: hi
12:51:34 claudiub moshele: great. :) If you have a couple of minutes, I have a couple of questions regarding sr-iov
12:51:48 moshele claudiub: ok
12:51:53 claudiub and more explicitly, some implementation details.
12:52:45 claudiub moshele: so, whatever guides i've found online are kind of outdated, or do not contain all the information i need. :)
12:53:56 claudiub moshele: well, first of all: let's say I have a NIC with SR-IOV capabilities, and I have a couple of free VFs. I'm already reporting them to nova in the virt driver as PCI devices with the type type-VF.
12:54:32 moshele claudiub: yes if they are in the whitelist
12:54:41 claudiub and they're whitelisted in the compute node's conf and nova-api's conf
12:54:45 claudiub up to this point, that should be fine IMO.
12:55:08 moshele whitelisted in the compute conf not need in the nova api
12:55:19 claudiub moshele: aliased, sorry. :)
12:55:29 claudiub aliased in the api conf file
12:55:45 moshele the alise should be in nova-api and nova-cmpute
12:56:13 claudiub oh, ok. i only knew about the alias in the nova-api. how does the alias help in the compute node?
12:56:50 claudiub now... I was thinking that I could claim those VFs the same way I would claim other PCI devices: through flavor extra_specs pci_passthrough:alias="the_alias:1"
12:56:58 moshele claudiub: for resize and migration
12:57:56 claudiub moshele: hmm, meaning that the nova compute manager will check if the destination has those aliases as well?
12:58:23 moshele claudiub: yes
12:58:36 claudiub moshele: hm, interesting. i did not know that. :) good to know.
13:00:31 claudiub anyways, continueing my idea. so, I would end up spawning an instance, and that instance would have in the instance.pci_requests.requests an VF. Now, I've been told before that all I have to do is just plug that VF into the VM and that's it
13:02:11 claudiub but that VF would not be tied to any neutron port, meaning that it there won't be normally any dhcp server serving it. I've been told that "the users would have to take care of that". is that ok?
13:02:35 claudiub nor would it have security groups applied to it.
13:07:05 openstackgerrit Rawan Herzallah proposed openstack/nova master: Adding NVMEoF for libvirt driver https://review.openstack.org/482640
13:10:42 claudiub moshele: also, the alias is quite limiting, IMO. for example, I have 2 identical SR-IOV NICs, which means same vendor id and same product id, but they're actually connected to different networks. the alias config option only allows the following fields: vendor_id, product_id, dev_type, name (alias). which means that I don't have any way to specify which network I actually want. :)
13:10:49 moshele claudiub: see this https://docs.openstack.org/mitaka/networking-guide/config-sriov.html
13:12:59 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'configuration' section https://review.openstack.org/478473
13:13:00 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Switch to openstackdocstheme https://review.openstack.org/477751
13:13:00 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'user' section https://review.openstack.org/478475
13:13:01 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Enable pep8 on doc generation code https://review.openstack.org/480469
13:13:01 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'contributor' section https://review.openstack.org/478474
13:13:02 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Start using oslo_policy.sphinxext https://review.openstack.org/479358
13:13:02 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rework README to reflect new doc URLs https://review.openstack.org/480074
13:13:03 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'reference' section https://review.openstack.org/478484
13:13:03 openstackgerrit Stephen Finucane proposed openstack/nova master: policies: Fix Sphinx issues https://review.openstack.org/480516
13:13:04 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rework index page per new sections https://review.openstack.org/478485
13:13:04 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Remove dead plugin https://review.openstack.org/478469
13:13:05 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Remove dead files https://review.openstack.org/478470
13:13:05 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Remove cruft from conf.py https://review.openstack.org/478468
13:13:06 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Start using oslo_config.sphinxext https://review.openstack.org/482961
13:13:06 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Use consistent author, section for man pages https://review.openstack.org/482960
13:13:41 moshele claudiub: there is SR-IOV for networking solution is work with the sriovnicswitch mechanism driver which is linux ip link based so maybe you need to add support there to windows
13:14:25 cdent avolkov`++
13:14:38 moshele claudiub: what you talking is pci passthrough when neutron is not aware of the SR-IOV network interface
13:14:56 stephenfin jaypipes: IF you get a chance to look through that doc-migration series again, I'd be mighty appreciative. Fixed the issue efried pointed out
13:15:07 stephenfin Sorry - gibi
13:15:24 jaypipes stephenfin: you mean the one you just rebased and pushed above?
13:15:37 claudiub moshele: are both cases viable?
13:16:18 stephenfin jaypipes: Yup https://review.openstack.org/#/q/topic:doc-migration+project:openstack/nova+status:open
13:16:58 jaypipes stephenfin: sure thing :)
13:17:01 efried stephenfin So that wasn't intended to be to my attn?
13:17:03 moshele claudiub: yes both of them should work
13:17:17 stephenfin efried: Nope, as always you can ignore me :)
13:18:00 claudiub moshele: aham. interesting. well, i'll look into the neutron side as well then. :) thanks for the help. :)
13:18:19 moshele claudiub: np :)
13:21:39 bauzas jaypipes: dansmith: sorry folks, was (again) on PTO yesterday
13:22:08 avolkov` stephenfin: placement-api-ref should also be moved to doc/source/api, I'll prepare patch when it's accomplished
13:22:27 stephenfin avolkov`: I thought we weren't moving that yet?
13:22:31 stephenfin Did plans change?
13:22:42 cdent stephenfin: dhellman wants that to happen asap
13:22:55 cdent so that we don't have to have a situation where there are publishing jobs that need to get changed
13:23:05 cdent (there's a placement-api-ref publishing job pending)
13:23:14 stephenfin cdent: Ah, hadn't heard that update
13:23:15 stephenfin Sounds fair
13:23:15 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'configuration' section https://review.openstack.org/478473
13:23:16 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'user' section https://review.openstack.org/478475
13:23:16 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'cli' section https://review.openstack.org/478472
13:23:17 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'contributor' section https://review.openstack.org/478474
13:23:17 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Switch to openstackdocstheme https://review.openstack.org/477751
13:23:18 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Start using oslo_config.sphinxext https://review.openstack.org/482961
13:23:18 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Enable pep8 on doc generation code https://review.openstack.org/480469
13:23:19 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Start using oslo_policy.sphinxext https://review.openstack.org/479358
13:23:19 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rework README to reflect new doc URLs https://review.openstack.org/480074
13:23:20 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Populate the 'reference' section https://review.openstack.org/478484
13:23:20 openstackgerrit Stephen Finucane proposed openstack/nova master: policies: Fix Sphinx issues https://review.openstack.org/480516
13:23:21 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Remove dead files https://review.openstack.org/478470
13:23:21 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rework index page per new sections https://review.openstack.org/478485
13:23:26 jaypipes efried: stephenfin just wanted to wake you out of your slumber.
13:24:30 jaypipes ralonsoh: I'm really liking the direction of your native OVSDB os-vif work.
13:29:16 stephenfin bauzas: Could I ask you to look at the doc-migration series starting with https://review.openstack.org/#/c/478468/ too?
13:29:26 bauzas stephenfin: roger this.
13:29:34 stephenfin Shouldn't take long, and it's easy re-arrange stuff later if we don't like the end result
13:29:42 stephenfin bauzas: Ta :)
13:30:03 ralonsoh jaypipes: thanks! Also I added functional tests. YOu only need to have a live OVS running in the system to execute them
13:30:37 jaypipes ralonsoh: I saw those. ++
13:36:26 openstackgerrit Stephen Finucane proposed openstack/nova master: Always set MTU when plugging devices https://review.openstack.org/408781
13:41:02 gibi stephenfin: thanks for the rebase
13:46:03 openstackgerrit Chason Chan proposed openstack/python-novaclient master: switch to openstackdocstheme https://review.openstack.org/477757
13:48:20 mriedem o/
13:49:35 jaypipes cdent: thx for your comments on https://review.openstack.org/#/c/482383/1. Answered. Please have a gander at the latest revision.
13:51:35 gibi mriedem: good morning! I fixed the last outstanding issue with the BDM in notification patch https://review.openstack.org/#/c/448779/
13:51:46 mriedem jaypipes: should i just remove this sentence? https://review.openstack.org/#/c/454322/35/nova/api/openstack/compute/rest_api_version_history.rst@598
13:51:54 mriedem jaypipes: i think that would probably clear up the confusion in that paragraph

Earlier   Later