| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-09-13 | |||
| 12:21:42 | sean-k-mooney | donnyd: which im hoping to run as a perodic job intially and eventuly move to check after its stableised. | |
| 12:22:09 | sean-k-mooney | donnyd: so for now relax, you have done more then enough to help with this effort | |
| 12:23:04 | donnyd | I'm very glad I could be helpful. | |
| 12:26:34 | sean-k-mooney | donnyd: for what its worth this is the set of ci topic we talked about last PTG https://etherpad.openstack.org/p/nova-ptg-train-ci | |
| 12:28:28 | sean-k-mooney | donnyd: line 36 multi numa guests and basically everytin form line 31 to 45 is what you have allowed us to start testing | |
| 12:30:43 | donnyd | Well if fn can do anything else, never be afraid to ask | |
| 12:32:22 | sean-k-mooney | efried: speaking of that etherpad line 29 https://etherpad.openstack.org/p/nova-ptg-train-ci can i anounce/propose the deprectation fo libvirt with virt-type uml | |
| 12:33:35 | sean-k-mooney | efried: there are other virt types that also shoudl be marked experimenal or deprecated but too slince they have no testing. lxc is almost workign again but there are defiently gaps | |
| 12:34:01 | sean-k-mooney | so i kind of want to hold off on that until Ussuri | |
| 12:35:28 | sean-k-mooney | efried: mriedem has an experimal lxc job and we have patches to fixed some of the short comings with cloud init https://review.opendev.org/#/c/667976/ | |
| 12:42:52 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Follow up for I220fa02ee916728e241503084b14984bab4b0c3b https://review.opendev.org/680810 | |
| 12:59:49 | donnyd | sean-k-mooney: I'm pretty interested in lxc functionality due to my resources not being as much as some others out there | |
| 13:00:34 | sean-k-mooney | ya it would be nice to get it working fully again | |
| 13:00:58 | sean-k-mooney | i was able to boot a contianer and inject keys with that cloud init patch | |
| 13:01:08 | donnyd | And knocking down targets from the ptg is a good thing too | |
| 13:02:14 | sean-k-mooney | the next gap for lxc that i hit while testing it is we are not correctly resizing the filesystme in the image to match the route disk on the flaovr | |
| 13:02:23 | sean-k-mooney | and you cant do it inside the contianer | |
| 13:02:43 | sean-k-mooney | so we would have to fix that before i would be comfortable useing it in production | |
| 13:09:57 | openstackgerrit | Lee Yarwood proposed openstack/nova master: Use os-brick locking for volume attach and detach https://review.opendev.org/614190 | |
| 13:10:31 | lyarwood | ^ any cores around, is it too late in the cycle to land bugfixes that bump requirements.txt? | |
| 13:10:46 | lyarwood | or is that not even a thing until rc? | |
| 13:17:32 | stephenfin | gibi: You should probably hit https://review.opendev.org/#/c/682025/ | |
| 13:18:00 | stephenfin | lyarwood: I've no idea. Might be a good thing to ask on #openstack-requirements ? | |
| 13:18:50 | stephenfin | or wait for mriedem, I guess | |
| 13:20:39 | stephenfin | speak of the devil | |
| 13:21:48 | mriedem | muwahahahaha? | |
| 13:22:46 | stephenfin | <lyarwood> ^ any cores around, is it too late in the cycle to land bugfixes that bump requirements.txt? | |
| 13:22:46 | stephenfin | ||
| 13:22:49 | stephenfin | mriedem: ^ | |
| 13:23:14 | mriedem | seems a bit risky to land that right now, | |
| 13:23:20 | mriedem | that's something i'd rather land early in U and let it bake | |
| 13:23:34 | mriedem | b/c fallout from that would probably be weird and hard to determine | |
| 13:25:35 | lyarwood | yup that's fair and on me for not pushing this earlier in the cycle | |
| 13:26:06 | mriedem | i think it's also a low severity issue, | |
| 13:26:10 | mriedem | i left some comments in the patch | |
| 13:26:23 | lyarwood | ack thanks | |
| 13:31:36 | efried | sean-k-mooney: "deprectation fo libvirt with virt-type uml" I don't know what that means. | |
| 13:31:54 | efried | So, sure, you can propose it, but someone who knows more than me might reject your proposal :) | |
| 13:32:27 | sean-k-mooney | the libvirt virt dirver support multiple virt-types that match to different dirver in libvirt | |
| 13:32:32 | sean-k-mooney | UML is user mode linux | |
| 13:33:06 | sean-k-mooney | i dont think we have ever had ci for that and dont know of anyone that support it in a product | |
| 13:34:02 | sean-k-mooney | we talked about deprecating libvirt virt types that do not have ci at the ptg in the test/ci session | |
| 13:34:26 | sean-k-mooney | since we are going to do that for xenserver i was wondering if we should do it for uml | |
| 13:35:51 | sean-k-mooney | again its more of a shot across the bow. | |
| 13:35:57 | dansmith | sean-k-mooney: I'd be fine with deprecation or straight removal, but I'm not sure it's worth it | |
| 13:36:21 | dansmith | the xen deprecation brings benefit, I'm not sure deprecating or removing uml is anything other than more work | |
| 13:36:36 | sean-k-mooney | we have a few conditional in the libvirt dirver to handel it | |
| 13:36:39 | dansmith | I know | |
| 13:36:40 | sean-k-mooney | so we could remove those | |
| 13:36:50 | sean-k-mooney | but it not a huge benifit | |
| 13:36:52 | dansmith | which is work for minimal benefit :) | |
| 13:37:12 | sean-k-mooney | ya | |
| 13:37:15 | dansmith | but if you want to do it I don't think there's any reason to keep it | |
| 13:37:51 | sean-k-mooney | well its not like anyone has filed a bug for it in years | |
| 13:38:07 | sean-k-mooney | whicih is proably the primary inicator no one is useing it | |
| 13:38:46 | sean-k-mooney | well there is one https://bugs.launchpad.net/nova/+bug/1597754 | |
| 13:38:47 | openstack | Launchpad bug 1597754 in OpenStack Compute (nova) "Unable to boot instance using UML" [Low,Triaged] | |
| 13:39:39 | sean-k-mooney | that seams like an imporant action for a virt type to support :) | |
| 13:41:17 | efried | Hell, even powervm could do THAT. | |
| 13:41:42 | sean-k-mooney | :) | |
| 13:42:05 | dansmith | I'd guess that was someone bored on a thursday evening, just trying to see if it still works | |
| 13:42:38 | dansmith | I'd bet it's less flexible than ironic even, probably fails the support matrix test | |
| 13:43:14 | sean-k-mooney | it was reported by mdbooth so i suspect it was filed when he was working on https://review.opendev.org/#/q/status:abandoned+topic:libvirt-imagebackend | |
| 13:43:37 | dansmith | heh | |
| 13:44:50 | sean-k-mooney | he fixed it in his image backend erfactor which we never merged. | |
| 13:45:35 | efried | So I'm having kind of a hard time believing that nova's gate throughput is on par with everyone else's, even in absolute numbers. Are small projects literally getting like two results every three hours? | |
| 13:46:10 | efried | i.e. the algorithm is broken | |
| 13:46:46 | sean-k-mooney | efried: we are effectivly down 2 ci providres at the momemnt i think | |
| 13:46:56 | efried | oh? | |
| 13:47:21 | sean-k-mooney | i think FN is still scaled down and we are having issue with vexxhost and boot form volume | |
| 13:47:34 | sean-k-mooney | and by we i mean infra | |
| 13:47:57 | sean-k-mooney | so maybe | |
| 13:49:21 | mriedem | today is then a docs and education friday! | |
| 13:49:40 | efried | Okay, well, I'm going to more or less take the day off, just watch and recheck. | |
| 13:50:11 | efried | I think mriedem has me on some messaging app or other if I'm needed. | |
| 13:51:39 | mriedem | how about commenting on the highlights patch which is due today https://review.opendev.org/#/c/681943/ | |
| 13:51:51 | efried | have I not done that? | |
| 13:51:53 | mriedem | dansmith: ^ words on the live migration + numa + cpu pinning + huge pages thing? | |
| 13:52:04 | mriedem | efried: oh i guess you did | |
| 13:52:36 | mriedem | or artom i guess ^ | |
| 13:52:41 | efried | I'm basically +1 on whatever you wind up with. alex_xu or luyao could add some words about vpmem, but I'm not sure if that's supposed to be a tech preview or a trumpet-from-rooftops, or whatever | |
| 13:53:04 | efried | also, I think "first draft" was due today, so we're done :P | |
| 13:54:20 | artom | mriedem, I'm happy with it, left a note | |
| 14:07:06 | stephenfin | dansmith/mriedem/efried: This is going to cause gate issues and should be merged, right? https://review.opendev.org/682025 | |
| 14:11:39 | mriedem | sob | |
| 14:12:27 | mriedem | no wonder this didn't go down :/ http://status.openstack.org/elastic-recheck/#1843615 | |
| 14:21:56 | sean-k-mooney | mriedem: i did get a passing test run with numa+pcpus+vpmem last night by the way | |
| 14:22:24 | mriedem | how are you testing vpmem in the gate? | |
| 14:22:26 | sean-k-mooney | https://review.opendev.org/#/c/681771/ https://zuul.opendev.org/t/openstack/build/6a67b28d6e9e4b739e7575df7943a69d | |
| 14:22:28 | mriedem | doesn't that require special hardware? | |
| 14:22:36 | sean-k-mooney | it just had the code merged | |
| 14:22:40 | sean-k-mooney | didnt test the feature | |
| 14:22:50 | mriedem | stephenfin: thanks for calling that out - infra just pushed it to the gate | |
| 14:23:08 | stephenfin | np | |
| 14:23:21 | sean-k-mooney | but we did test with pinning+hugepage+multi numa https://review.opendev.org/#/c/681771/2/playbooks/nfv/nfv.yaml | |
| 14:24:43 | mriedem | sean-k-mooney: iow, vpmem didn't obviously regress any RT stuff with numa+pcpu in the same run | |
| 14:25:01 | sean-k-mooney | mriedem: right | |
| 14:25:06 | sean-k-mooney | no idea if it works | |
| 14:25:16 | sean-k-mooney | but when its not used it does not seam to break anything | |
| 14:26:09 | sean-k-mooney | unfortuetly it looks like the intel pmem ci did not run on that | |