| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-06-01 | |||
| 14:39:04 | openstackgerrit | Alex Szarka proposed openstack/nova master: Transform instance.exists notification https://review.openstack.org/403660 | |
| 14:41:10 | mriedem | sean-k-m1: at some point i wanted to ask you about https://review.openstack.org/#/c/441183/ which came up in the nova meeting today | |
| 14:41:11 | mriedem | you and jay | |
| 14:41:40 | mriedem | i.e. is this fine as a plugin way of doing things? i think everything else nova has supported for os-vif types has been in the os-vif library, so this would be the first out of tree plugin | |
| 14:42:11 | mriedem | egonzalez: hi | |
| 14:42:27 | openstackgerrit | Alex Szarka proposed openstack/nova master: Add method for verify multiple versioned notifications https://review.openstack.org/465526 | |
| 14:43:07 | mriedem | "THE place for all your vrouter needs!" | |
| 14:43:14 | egonzalez | mriedem, just wanted to explain the use_neutron/dns bug | |
| 14:43:37 | takashin | dansmith: Would you review https://review.openstack.org/#/c/408955/ again? | |
| 14:43:42 | sfinucan | egonzalez: I left comments on said bug | |
| 14:44:56 | takashin | dansmith: Are you around? | |
| 14:45:24 | egonzalez | mriedem, sfinucan thanks will wait until the metadata issue is fixed to try re verify | |
| 14:45:43 | openstackgerrit | Alex Szarka proposed openstack/nova master: Refactor create_delete_server_with_instance_update https://review.openstack.org/466296 | |
| 14:45:51 | sfinucan | egonzalez: which metadata issue? | |
| 14:46:20 | egonzalez | sfinucan, opened another bug for that https://bugs.launchpad.net/nova/+bug/1694666 | |
| 14:46:20 | openstack | Launchpad bug 1694666 in OpenStack Compute (nova) "metadata service PicklingError: Can't pickle <class 'sqlalchemy.orm.session.Session'>" [Undecided,New] | |
| 14:47:11 | sean-k-m1 | mriedem: looking at it now | |
| 14:47:12 | sfinucan | egonzalez: Ah, well I guess that answers my question - you _are_ using the metadata service so? | |
| 14:47:20 | sfinucan | ...presumably with cloud-init? | |
| 14:47:26 | mriedem | mhenkel: i commented on this in the patch and bp, but i'd like to see documentation about how to use this plugin somewhere, which should probably at least be in the readme for the plugin itself https://github.com/Juniper/contrail-nova-vif-driver | |
| 14:47:49 | mhenkel | yes, got the point | |
| 14:47:55 | egonzalez | sfinucan, no cloud-init at all, both issues are different | |
| 14:48:18 | mriedem | i thought we've had this pickling issue before, or reported before at least | |
| 14:48:36 | egonzalez | sfinucan, this last one is, fresh deploy. create a cirros instance and metadata raising those errors | |
| 14:48:48 | sfinucan | egonzalez: and the other one? | |
| 14:49:24 | sean-k-mooney | mriedem: so i would like to make it so that https://review.openstack.org/#/c/441183 would not be required by having neutron pass the plugin to use to nova when we update neutron to pass os-vif objects | |
| 14:50:41 | mriedem | sean-k-mooney: that's the long-term goal but no one is working on that, correct? | |
| 14:50:55 | mriedem | i think ian wells said he'd work on it at both the austin and barcelona summits but it hasn't happened | |
| 14:51:13 | sean-k-mooney | well i am if i can get to it. | |
| 14:51:29 | mriedem | egonzalez: do you know if that metadata pickling issue also happens in ocata? | |
| 14:51:37 | mhenkel | bauzas: on the relevance to dpdk | |
| 14:51:51 | sean-k-mooney | mriedem: i want to do a poc of that before the end of pike | |
| 14:51:57 | mhenkel | this patch is only needed for dpdk | |
| 14:52:17 | sfinucan | egonzalez: as a quick hack, what happens if you change [1] to 'return "%s.%s" % (self.instance.hostname, "bananas")' [1] https://github.com/openstack/nova/blob/a1eca94d89ba49eaa521303dfa8d991237cb5eb7/nova/api/metadata/base.py#L535-L538 | |
| 14:52:21 | egonzalez | mriedem, last manual test worked in ocata yesterday. gates are happy at too | |
| 14:52:27 | sfinucan | easiest way to check if that's where the problem lies ;) | |
| 14:52:35 | sean-k-mooney | mriedem: but regardign that patch was there something specific you wanted my input on? | |
| 14:52:54 | bauzas | mhenkel: sure, I understand that, I just wanted to explain that we only need to name the BP as what it's provided by the plugin | |
| 14:53:04 | mriedem | egonzalez: is that what you mean by "Only source images from master fail, binary(rdo or ubuntu packages) works for now" - are the binary packages from nova 15.0.0 (ocata)? | |
| 14:53:24 | mriedem | sean-k-mooney: in general are you ok with it? besides needing to probably change the vif type name to something less generic, and needing a release note | |
| 14:53:49 | egonzalez | mriedem, nope, probably packages are not build or propagated to main delorean repo | |
| 14:53:50 | sean-k-mooney | mriedem: yes i have previously revied this and i just reviewed it again | |
| 14:54:19 | sean-k-mooney | mriedem: it looks fine to me bar the name as you said | |
| 14:54:28 | mhenkel | bauzas: something like: nova-vif-plugin-for-opencontrail-vrouter? | |
| 14:55:53 | mriedem | sfinucan: egonzalez: so basically it sounds like if you're using neutron, we should set dhcp_domain to '' as the default so you'll just use the instance.hostname https://github.com/openstack/nova/blob/a1eca94d89ba49eaa521303dfa8d991237cb5eb7/nova/api/metadata/base.py#L537 | |
| 14:56:33 | mriedem | maybe people getting dns from neutron/designate just knew to change the dhcp_domain config option? i think miguel lavalle added that support in mitaka | |
| 14:56:35 | mriedem | for neutron dns i mean | |
| 14:57:35 | mlavalle | mriedem: hey | |
| 14:57:44 | sfinucan | mriedem: Yeah, we should probably ignore the nova-network-only option | |
| 14:58:18 | mriedem | mlavalle: trying to sort out this bug https://bugs.launchpad.net/nova/+bug/1694417 | |
| 14:58:18 | openstack | Launchpad bug 1694417 in OpenStack Compute (nova) "dhcp_domain used when use_neutron is not set" [Undecided,Incomplete] | |
| 14:58:37 | mriedem | and wondering if neutron + dns means you need to configure nova's dhcp_domain option to '' when using neutron and the metadata service | |
| 14:58:53 | mriedem | mlavalle: feel free to reply on the bug later if you're busy | |
| 14:59:28 | mlavalle | mriedem: at first glance, it should't be related | |
| 14:59:51 | mlavalle | mriedem: at look at the bug later, though | |
| 14:59:58 | mriedem | mlavalle: your spec https://specs.openstack.org/openstack/nova-specs/specs/mitaka/implemented/internal-dns-resolution.html just talks about cloud-init, not about the metadata service, so i think that might be the issue | |
| 14:59:59 | mriedem | thanks | |
| 15:00:01 | egonzalez | going to add config file as reference in the bug | |
| 15:00:11 | mriedem | egonzalez: can you also show the port details? | |
| 15:00:19 | mriedem | would like to see if dns_name is set on the port | |
| 15:01:03 | egonzalez | mriedem, sure | |
| 15:12:14 | openstackgerrit | Ed Leafe proposed openstack/nova master: Have scheduler claim resources https://review.openstack.org/463993 | |
| 15:12:15 | openstackgerrit | Ed Leafe proposed openstack/nova master: Pass the number of hosts to return explicitly https://review.openstack.org/465767 | |
| 15:12:15 | openstackgerrit | Ed Leafe proposed openstack/nova master: Pass a list of instance UUIDs to scheduler https://review.openstack.org/465171 | |
| 15:12:16 | openstackgerrit | Ed Leafe proposed openstack/nova master: Return multiple hosts from select_destinations https://review.openstack.org/465768 | |
| 15:12:42 | edleafe | cdent: bauzas: ^^ rebased and updated | |
| 15:12:57 | cdent | yay! | |
| 15:14:12 | edleafe | cdent: You'll be happy to learn that your precious blank line has been returned unharmed | |
| 15:14:29 | cdent | doubleyay! | |
| 15:17:14 | hongbin | mriedem: hi matt, after discussing with neutron ptl, we would like to ban neutronclient 6.3.0 due to an issue: http://lists.openstack.org/pipermail/openstack-dev/2017-June/117763.html | |
| 15:17:40 | hongbin | mriedem: i saw nova team proposed to bump the version to 6.3.0, so want to check with you: https://github.com/openstack/requirements/commit/210b33e01658374ac248a188cfdcb35a64a35c80 | |
| 15:18:06 | mriedem | hongbin: that was for the global request id stuff sdague was working on | |
| 15:18:36 | hongbin | mriedem: yes, i couldn't find sdague in the irc right now | |
| 15:18:43 | mriedem | he's on vacation this week | |
| 15:19:00 | hongbin | ok | |
| 15:19:25 | mriedem | shouldn't https://bugs.launchpad.net/zun/+bug/1694897 be targeted at python-neutronclient rather than zun? | |
| 15:19:25 | openstack | Launchpad bug 1694897 in Zun "neutron tag xxx.xxx doesn't work correctly" [Undecided,New] | |
| 15:19:52 | hongbin | yes, i should change it | |
| 15:20:01 | alexchadin | hi Nova team | |
| 15:20:08 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: Fix missing marker functions https://review.openstack.org/469920 | |
| 15:21:01 | alexchadin | how can I get compute.node metrics from nova-compute? I set compute_monitors=cpu.virt_driver but still don't get any compute-node related metrics | |
| 15:21:08 | alexchadin | Ocata version | |
| 15:29:10 | hongbin | mriedem: fyi: https://review.openstack.org/469923 , hope it doesn't break anything on nova side | |
| 15:30:36 | openstackgerrit | Michael Henkel proposed openstack/nova master: contrail: add vrouter VIF plugin type support https://review.openstack.org/441183 | |
| 15:34:28 | openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova master: xen: pass Xen console in cmdline https://review.openstack.org/469826 | |
| 15:40:33 | mriedem | hongbin: we're good https://review.openstack.org/#/c/468356/ the nova change didn't land yet | |
| 15:40:55 | hongbin | mriedem: great. thanks for confirming | |
| 15:41:39 | openstackgerrit | Michael Henkel proposed openstack/nova master: contrail: add vrouter VIF plugin type support https://review.openstack.org/441183 | |
| 15:51:18 | openstackgerrit | Chris Dent proposed openstack/nova master: Sync os-traits to Traits database https://review.openstack.org/469578 | |
| 15:52:09 | cdent | edleafe: that ^ adds some locks and other checks but I haven't added a test that alex_xu asked for. if you feel like taking it on after the api-wg meeting go for it. i'll be gone after that | |
| 15:57:13 | edleafe | cdent: I'll try, but I have a pretty full plate today. If not today, then definitely tomorrow | |
| 15:57:31 | cdent | thanks | |
| 15:58:43 | mriedem | artom: have you seen something like this? https://bugs.launchpad.net/nova/+bug/1694666 | |
| 15:58:43 | openstack | Launchpad bug 1694666 in OpenStack Compute (nova) "metadata service PicklingError: Can't pickle <class 'sqlalchemy.orm.session.Session'>" [Undecided,New] | |
| 15:58:55 | mriedem | or know of someone working on reducing InstanceMetadata to a smaller primitive for pickling? | |
| 15:59:49 | artom | mriedem, whoa | |
| 16:00:15 | artom | mriedem, no, haven't seen anything like it | |
| 16:00:25 | artom | Why is it trying to pickle the session? o_O | |
| 16:00:42 | mriedem | yeah, it pickles everything in InstanceMetadata, which includes a bunch of versioned objects | |
| 16:00:44 | mriedem | like Instance | |