| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-07-20 | |||
| 18:43:16 | sean-k-mooney | trinaths1: just so you are aware without also installing libnuma-devel you will not be able to use hugepages or cpu pinning | |
| 18:43:58 | trinaths1 | sean-k-mooney: agree. | |
| 18:45:06 | melwitt | dansmith: heh. I shall sanity check that. I expect they don't need to be remotable | |
| 18:45:11 | trinaths1 | sean-k-mooney: its numactl-devel for fedora 23 | |
| 18:46:11 | dansmith | melwitt: cool | |
| 18:46:49 | sean-k-mooney | trinaths1: ah i just have centos and ubunut vms to hand currently so i assumed it would be the same as centos | |
| 18:47:53 | trinaths1 | sean-k-mooney: oh okay. | |
| 18:48:12 | trinaths1 | sean-k-mooney: even yum is moved to dnf | |
| 18:49:13 | sean-k-mooney | trinaths1: yes i think you can use dnf on centos too. they have a compaitblity shim on fedora to allow yum to still work but dnf is faster at installing things | |
| 18:50:07 | trinaths1 | sean-k-mooney: agree | |
| 18:54:33 | sean-k-mooney | trinaths1: anyway, that is my best quess of how to fix the issue. if you reconfig and recomiple then restart libvirtd then it might solve your issue | |
| 18:55:33 | trinaths1 | sean-k-mooney: ok. re-configuring and re-compiling libvirt now. | |
| 18:55:57 | trinaths1 | sean-k-mooney: i will share the configure summary in a while. | |
| 18:56:32 | trinaths1 | sean-k-mooney: http://paste.openstack.org/show/616077/ | |
| 18:57:25 | sean-k-mooney | trinaths1: well that looks a little better numa and yajl are now enabled. | |
| 18:57:53 | trinaths1 | sean-k-mooney: yes. any more options/packages i need to enable to compile ? | |
| 18:59:30 | sean-k-mooney | trinaths1: from what i can see you should have enough libvires installed for it to work with openstack. | |
| 19:00:07 | trinaths1 | sean-k-mooney: did i miss other libraries ? | |
| 19:00:10 | sean-k-mooney | trinaths1: if you want to use glsterfs or ceph you will need to install extra packages | |
| 19:01:21 | sean-k-mooney | trinaths1: but for a normal install i dont think you need any other packages | |
| 19:01:29 | mriedem_afk | ildikov: so i think we can remove handling VolumeUnattached from the rest api controller for delete | |
| 19:01:34 | mriedem_afk | (detach_volume) | |
| 19:04:25 | trinaths1 | sean-k-mooney: we did not more to ceph end. i have started re-compile of libvirt. I will update when its completed. | |
| 19:08:45 | mriedem | ildikov: ok comments inline | |
| 19:08:50 | mriedem | which are hopefully clear | |
| 19:13:53 | openstackgerrit | Jay Pipes proposed openstack/nova master: style-only: s/context/ctx/ https://review.openstack.org/485791 | |
| 19:13:54 | openstackgerrit | Jay Pipes proposed openstack/nova master: use os_traits.MISC_SHARES_VIA_AGGREGATE https://review.openstack.org/485792 | |
| 19:17:18 | ildikov | mriedem: ah ok, I didn't get to the point of fully removing it from the previous comments as it wasn't introduced in check_detach | |
| 19:17:28 | ildikov | mriedem: thanks for the comments, I'll update the patch | |
| 19:17:45 | ildikov | mriedem: cleaning up is always good, so yeah, let's do it properly | |
| 19:17:46 | mriedem | jaypipes: i for one love trailing parans | |
| 19:17:50 | mriedem | *parens | |
| 19:18:02 | mriedem | 'know': { | |
| 19:18:09 | mriedem | 'what': 'i mean?' | |
| 19:18:09 | mriedem | } | |
| 19:19:54 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova master: Updated from global requirements https://review.openstack.org/485634 | |
| 19:19:59 | jangutter | mriedem: fun fact, when I signed my current employment contract, there was an unmatched parenthesis. I still am uncertain whether I am actually employed. | |
| 19:20:25 | mriedem | ha | |
| 19:21:10 | openstackgerrit | Sean Dague proposed openstack/nova master: DNM: WIP adding tests for requestlog https://review.openstack.org/485793 | |
| 19:49:58 | jaypipes | mriedem: :) | |
| 20:09:01 | openstackgerrit | Robert Ellis proposed openstack/nova master: Changed code or added clarifying documentation where nodename is used in ironic driver. https://review.openstack.org/485803 | |
| 20:13:23 | openstackgerrit | Robert Ellis proposed openstack/nova master: Clarifying node_uuid usage in ironic driver. https://review.openstack.org/485803 | |
| 20:52:00 | openstackgerrit | melanie witt proposed openstack/nova master: Detach volume after deleting instance with no host https://review.openstack.org/340614 | |
| 20:52:22 | mriedem | nova team meeting in 8 minutes | |
| 20:53:20 | jaypipes | supernovas... UNITE! | |
| 20:53:53 | openstackgerrit | Sean Dague proposed openstack/nova master: WIP: request_log addition for running under uwsgi https://review.openstack.org/485602 | |
| 20:56:17 | jaypipes | dansmith: Voltron. :) | |
| 20:56:26 | jangutter | hey, I just got that. | |
| 20:56:35 | jangutter | coffee levels at 10% | |
| 20:56:40 | dansmith | I' | |
| 20:56:48 | dansmith | jaypipes: I'm proud to have no idea what that is | |
| 20:56:52 | jaypipes | heh | |
| 20:58:02 | melwitt | I always thought power rangers were a rip off voltron | |
| 20:58:16 | mriedem | they are all a rip off of some thing that started in japan | |
| 20:58:19 | mriedem | and are all terrible | |
| 20:58:38 | jangutter | mriedem: "Voltron - The End" is amazing though. | |
| 20:59:14 | melwitt | nova meeting in one minute | |
| 20:59:28 | mriedem | is the end of voltron better then the end of seinfeld? | |
| 20:59:30 | mriedem | because that's not hard | |
| 20:59:41 | jangutter | mriedem: it's a 4 minute youtube video. | |
| 20:59:59 | mriedem | no time | |
| 21:00:07 | mriedem | that's 3:45 too long | |
| 21:02:09 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Cleanup unnecessary logic in os-volume_attachments controller code https://review.openstack.org/485823 | |
| 21:03:27 | openstackgerrit | melanie witt proposed openstack/nova master: Make Quotas object favor the API database https://review.openstack.org/410945 | |
| 21:04:24 | openstackgerrit | melanie witt proposed openstack/nova master: Add online migration to move quotas to API database https://review.openstack.org/410946 | |
| 21:22:28 | cfriesen | If anyone wants a short break from what they're doing, I'd welcome some additional reviewers for https://review.openstack.org/#/c/468203 and https://review.openstack.org/#/c/461456 . Sahid doesn't really like the proposed tweaks to cpu_realtime_mask but sfinucan is fine with it...would be useful to get a few more opinions. | |
| 21:45:20 | melwitt | dansmith: I removed the remotables from the Quotas object patch. those methods are only called in the quota classes API | |
| 21:45:30 | melwitt | thanks for pointing that out | |
| 21:45:39 | dansmith | melwitt: oh sweet | |
| 21:45:58 | melwitt | I must have had remotable fever when I first added them | |
| 21:46:12 | dansmith | a common affliction | |
| 21:46:34 | melwitt | haha, yup!! | |
| 21:47:03 | mriedem | cfriesen: so you mean, take a break from reviewing code to review some code? | |
| 21:47:40 | cfriesen | mriedem: but it's different code! | |
| 21:47:46 | mriedem | OH! | |
| 21:48:01 | dansmith | less critical for the release code? | |
| 21:48:03 | dansmith | sign me up. | |
| 21:48:13 | cfriesen | less stressful, more laid-back | |
| 21:48:14 | mriedem | cfriesen: did you miss this while you were on vacation? | |
| 21:48:21 | mriedem | this abuse | |
| 21:48:43 | cfriesen | all the time. had to go search out random strangers | |
| 21:49:16 | mriedem | random...canadian strangers? | |
| 21:49:22 | cfriesen | no, they're too polite | |
| 21:49:25 | mriedem | right | |
| 21:50:13 | cfriesen | so are multi-cells actually going to be completely usable in pike? anything that needs some reviews to make sure that happens? | |
| 21:50:57 | mriedem | the last big thing is moving quotas to the api | |
| 21:51:06 | mriedem | starts here https://review.openstack.org/#/c/410945/ | |
| 21:52:01 | mriedem | cfriesen: and this doesn't have an owner https://bugs.launchpad.net/nova/+bug/1702935 | |
| 21:52:02 | openstack | Launchpad bug 1702935 in OpenStack Compute (nova) "Unshelving an offloaded server does not restrict to the original cell" [High,Confirmed] | |
| 21:52:15 | mriedem | otherwise i | |
| 21:52:37 | mriedem | otherwise i'd like to see someone doing some performance testing, maybe with rally, comparing stable/ocata to current master | |
| 21:53:06 | mriedem | now that we have the counting quotas stuff in and some changes to the filter scheduler for using placement | |
| 21:53:21 | mriedem | really just try to break things | |
| 21:53:50 | cfriesen | our arch guys have decided to skip ocata and jump from newton right to pike (yeah, I know), so I won't be able to compare against ocata. We should never have made kilo->mitaka work. | |
| 21:54:56 | melwitt | at least you're upgrading? :P | |
| 21:55:08 | cfriesen | there is that | |
| 21:55:53 | melwitt | skipping releases the right way just means you pull the middle code and just run migrations, then move to the next I thought. I would think that shouldn't be too painful or is there more to it? | |
| 21:55:56 | cfriesen | has anyone looked at allowable latency between the API cell and the compute cells? | |
| 21:56:25 | cfriesen | melwitt: we're going to have computes running newton live migrating to computes running pike | |
| 21:56:42 | melwitt | oh, live, okay | |