| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-28 | |||
| 00:02:11 | mikal | And its not coming to me instantly | |
| 00:02:41 | dansmith | well, it is the same behavior for lxc and parallels, and doesn't require needing to know what filesystem is there | |
| 00:03:01 | mikal | I do recall someone else wanting directories, it might have been LXC | |
| 00:03:05 | mikal | I'm a bit fuzzy on the details | |
| 00:03:11 | mikal | That said, it didn't merge | |
| 00:03:13 | tonyb | dansmith: I don't see any open reviews, just the old 2015 blueprint | |
| 00:03:18 | mikal | So how does LXC do config drives at the moment? | |
| 00:03:23 | dansmith | tonyb: they had a patch I think | |
| 00:03:26 | dansmith | mikal: it doesn't AFAIK | |
| 00:03:33 | mikal | Oh, ok | |
| 00:04:21 | dansmith | actually, | |
| 00:04:31 | dansmith | tonyb: maybe I'm thinking of the other libvirt image type thing | |
| 00:04:51 | dansmith | I think we're to expect them to have a spec though, IIRC | |
| 00:05:10 | tonyb | dansmith: Yeah could be ... I can't see an configdrive patch and we discussed them at the same time late on Friday soe memory==fuzzy | |
| 00:05:22 | dansmith | yeah | |
| 00:05:48 | dansmith | tonyb: he should be around during your day today I think, so maybe ping him | |
| 00:05:58 | mikal | Was there any discussion of vendordata? I kind of consider it complete at this point, but I don't know if people showed up with pitchforks with further complaints. | |
| 00:06:20 | tonyb | mikal: not near me | |
| 00:06:22 | dansmith | mikal: we declared it a success given the lack of any pitchforks (or other representation) | |
| 00:07:11 | mikal | dansmith: good. I shall pretend it never happened then. | |
| 00:07:34 | dansmith | well you might miss out on eventual praise if you do | |
| 00:07:52 | mikal | dansmith: oh no, I shall tell random people on the street of my exploits | |
| 00:07:58 | dansmith | heh | |
| 00:08:14 | tonyb | dansmith: he'll have a minion pass on the praise in mime | |
| 00:08:17 | mikal | dansmith: and it will go on my application form for Fellow. "I invented the shit out of config drive v2, and the vendordata re-write!" | |
| 00:08:24 | dansmith | heh | |
| 00:08:31 | cburgess | lol | |
| 00:08:31 | mikal | Both fundamental internet technologies | |
| 00:08:58 | mikal | Actually, a tshirt that said "I invented the shit out of config drive v2" would probably be a thing I'd wear | |
| 00:09:48 | dansmith | and most people would not know it's not that impressive, so, sounds like a good idea | |
| 00:10:14 | mikal | I would like to note we intended to have deleted config drive by now | |
| 00:10:20 | mikal | But yet, here we are | |
| 00:10:27 | mikal | Its the gift that keeps on giving | |
| 00:12:07 | tonyb | mikal: we shoudl just fix the metadata service so configdrive isn't needed | |
| 00:12:14 | cburgess | Yeah I don't see how we can ever get rid of it at this poing given we can't do metadata over IPv6 | |
| 00:12:15 | tonyb | mikal: how hard can it be ;P | |
| 00:12:27 | cburgess | and yeah metadata service is... umm... horrible | |
| 00:12:44 | tonyb | cburgess: it's only code from being less horrible | |
| 00:12:54 | mikal | cburgess: that's probably what I'd do next if I had any time | |
| 00:12:58 | openstackgerrit | John Griffith proposed openstack/nova master: Add Cinder V3 Detach calls https://review.openstack.org/438750 | |
| 00:13:02 | mikal | But instead, I write project plans | |
| 00:13:17 | cburgess | tonyb mikal Yeah I wish I had the time to fix metadata as well. Its caused us a lot of scaling problems | |
| 00:13:43 | mikal | cburgess: you just need better users... | |
| 00:13:53 | cburgess | mikal Don't get me started | |
| 00:14:21 | cburgess | Somewhere.. someone on the internet created a chef plugin of some type that spams the hell out of metatata with request from every conciviable cloud format every seconds. | |
| 00:14:32 | cburgess | I have users making 350+ metadata requests a second. | |
| 00:14:35 | cburgess | Its... insane | |
| 00:15:04 | dansmith | cburgess: yeah, it's super good for that | |
| 00:15:10 | dansmith | cburgess: users love hitting it hard | |
| 00:15:33 | mikal | cburgess: you're really going to hate vendordata then | |
| 00:15:46 | mikal | Cause we fan out that request to N other services | |
| 00:16:13 | cburgess | dansmith Is that supposed to be in the sarcasm font? | |
| 00:16:18 | dansmith | cburgess: no | |
| 00:16:24 | dansmith | cburgess: I have 'em too | |
| 00:16:31 | cburgess | mikal I still have to configure that though right? | |
| 00:16:33 | dansmith | cburgess: run puppet every minute and slam the metadata server | |
| 00:16:38 | mikal | cburgess: yes, its opt in | |
| 00:16:50 | mikal | cburgess: but if any of those users want windows active directory registration... | |
| 00:16:52 | cburgess | dansmith Ahh ok yeah, its insane to me, and they seem to not understand how its a problem | |
| 00:16:57 | dansmith | yeah | |
| 00:16:59 | dansmith | but alas | |
| 00:17:02 | cburgess | mikal Hmmm I might need that | |
| 00:17:10 | mikal | cburgess: good news! | |
| 00:17:15 | mikal | cburgess: and by good I mean bad | |
| 00:17:18 | cburgess | dansmith Yeah so... like I said.. if I had the time I would propose a re-write of metadata because its horrible. | |
| 00:17:35 | mikal | I could write a project plan for that | |
| 00:17:39 | mikal | I'm getting quite good at them | |
| 00:17:46 | mikal | Not good at getting things resourced | |
| 00:17:54 | mikal | But I have nice Microsoft Word documents | |
| 00:17:58 | cburgess | LOL | |
| 00:19:51 | openstackgerrit | Li Yingjun proposed openstack/nova master: Error message should not include SQL command https://review.openstack.org/436773 | |
| 03:31:43 | openstackgerrit | Li Yingjun proposed openstack/nova-specs master: Spec for support set vgamem for video https://review.openstack.org/438787 | |
| 03:51:26 | openstackgerrit | Li Yingjun proposed openstack/nova master: libvirt: Enable custom video vgamem setting https://review.openstack.org/411191 | |
| 04:12:51 | openstackgerrit | Huan Xie proposed openstack/nova master: Make xenserver driver compatible with assert_can_migrate https://review.openstack.org/438792 | |
| 07:26:39 | openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova master: Remove mox from nova.tests.unit.virt.xenapi.test_xenapi[1] https://review.openstack.org/438825 | |
| 07:28:20 | openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova master: Remove mox from nova/tests/unit/virt/xenapi/test_xenapi.py[1] https://review.openstack.org/438825 | |
| 07:38:35 | openstackgerrit | Takashi NATSUME proposed openstack/nova-specs master: List/show all server migration types https://review.openstack.org/407237 | |
| 07:59:38 | Dinesh_Bhor | Hi all, I have fresh OpenStack installed on Ubuntu 14.04 but I am unable to boot vm. Logs in n-cpu says that: "InternalError: Nova requires libvirt version 1.2.9 or greater." | |
| 08:01:52 | Dinesh_Bhor | I tried to update libvirt version but it only installs 1.2.2. I have ubuntu 14.04 LTS. | |
| 08:02:09 | Dinesh_Bhor | I found some bug here: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1532460 | |
| 08:02:09 | openstack | Launchpad bug 1532460 in libvirt (Ubuntu) "Please upgrade to allow virsh blockcommit in Ubuntu 14. libvirt-1.2.9 (and above)" [Undecided,Fix released] | |
| 08:43:26 | gmann | alex_xu: mriedem: this looks good but new error code bug me - https://review.openstack.org/#/c/429476/4 | |
| 09:00:57 | vsaienk0 | Hello nova team I'm wondered why multinode tempest test is skipped on grenade multinode in base tests? http://logs.openstack.org/76/429476/4/check/gate-grenade-dsvm-neutron-multinode-ubuntu-xenial/f6613ff/logs/grenade.sh.txt.gz#_2017-02-26_23_56_30_146 | |
| 09:05:39 | vsaienk0 | I meant is it expected behavior? | |
| 09:08:34 | frickler | can someone explain the difference between default_availability_zone and default_schedule_zone? both descriptions look roughly the same to me here: https://docs.openstack.org/draft/config-reference/compute/config-options.html#id3 | |
| 09:25:12 | openstackgerrit | Gao Yuan proposed openstack/nova master: Stop using mox stubs in nova.tests.unit.api.openstack.compute.test_serversV21 https://review.openstack.org/438873 | |
| 09:38:23 | johnthetubaguy | frickler: I did look into all that AZ stuff when we tidied up the descriptions, but you are right, its not very clear from that | |
| 09:39:22 | frickler | johnthetubaguy: I just opened https://bugs.launchpad.net/openstack-manuals/+bug/1668542 because looking at the code I became even more confused | |
| 09:39:22 | openstack | Launchpad bug 1668542 in openstack-manuals "nova.conf - az configuration options in Configuration Reference" [Undecided,New] | |
| 09:39:39 | johnthetubaguy | frickler: that probably needs to be a nova bug, the text lives in nova | |
| 09:40:29 | frickler | johnthetubaguy: added. probably the "open bug" link on the bug page could need amending then, too | |
| 09:40:50 | frickler | s/bug page/doc page/ | |
| 09:42:35 | johnthetubaguy | frickler: yeah, it probably does, not sure if you can change that per page, probably worth a manuals bug for that! | |
| 09:44:10 | johnthetubaguy | frickler: ah, so one is about where the host lives by default, and one is about where instances go, I think... | |
| 09:50:08 | frickler | johnthetubaguy: yes, that is my interpretation, but would be nice to get it confirmed, then I could make a patch for the docs | |
| 09:52:29 | frickler | johnthetubaguy: hmm, the source for the page seems to be ./doc/config-reference/source/compute/config-options.rst in openstack-manuals, though | |
| 09:57:40 | johnthetubaguy | frickler: that thing is auto-generated by a script, so its a two part ifx | |
| 09:57:42 | johnthetubaguy | fix | |
| 09:58:19 | openstackgerrit | John Garbutt proposed openstack/nova master: Fix default_availability_zone docs https://review.openstack.org/438887 | |
| 09:59:09 | johnthetubaguy | frickler: does that patch look like it makes it cleaer? | |