| *** k0da has quit IRC | 00:17 | |
| *** edmondsw has joined #openstack-powervm | 01:32 | |
| *** edmondsw has quit IRC | 02:32 | |
| *** apearson has joined #openstack-powervm | 03:49 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 03:59 | |
| *** edmondsw has joined #openstack-powervm | 04:01 | |
| *** AlexeyAbashkin has quit IRC | 04:03 | |
| *** edmondsw has quit IRC | 04:06 | |
| *** adi_____ has quit IRC | 04:10 | |
| *** AndyWojo has quit IRC | 04:10 | |
| *** catmando has quit IRC | 04:11 | |
| *** adi_____ has joined #openstack-powervm | 04:13 | |
| *** catmando has joined #openstack-powervm | 04:17 | |
| *** AndyWojo has joined #openstack-powervm | 04:18 | |
| *** edmondsw has joined #openstack-powervm | 05:49 | |
| *** apearson has quit IRC | 05:52 | |
| *** chhagarw has joined #openstack-powervm | 05:52 | |
| *** edmondsw has quit IRC | 05:54 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 06:16 | |
| *** AlexeyAbashkin has quit IRC | 06:29 | |
| *** edmondsw has joined #openstack-powervm | 07:38 | |
| *** edmondsw has quit IRC | 07:42 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 08:22 | |
| *** k0da has joined #openstack-powervm | 08:52 | |
| *** efried has joined #openstack-powervm | 09:06 | |
| *** edmondsw has joined #openstack-powervm | 09:26 | |
| *** edmondsw has quit IRC | 09:30 | |
| openstackgerrit | Merged openstack/networking-powervm master: use common agent topics from neutron-lib https://review.openstack.org/548701 | 09:54 |
|---|---|---|
| *** AlexeyAbashkin has quit IRC | 11:03 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 11:09 | |
| *** edmondsw has joined #openstack-powervm | 12:15 | |
| *** edmondsw has quit IRC | 12:20 | |
| *** edmondsw has joined #openstack-powervm | 13:06 | |
| *** apearson has joined #openstack-powervm | 13:43 | |
| efried | edmondsw: Do we have an interest in "live resize"? | 13:57 |
| efried | I imagine PVC supports this via backdoor. Would it be interesting/desirable or allow us to remove tech debt if it was supported in nova proper? | 13:58 |
| edmondsw | efried yes | 14:13 |
| efried | Okay, I'll add us to specs etc. | 14:13 |
| edmondsw | I think I'm already on at least one of them | 14:13 |
| efried | Oh, look, you are. And so am I. | 14:14 |
| efried | https://review.openstack.org/#/c/141219/ | 14:14 |
| edmondsw | yep... I prompted them on 2/23 to repropose for Rocky :) | 14:15 |
| edmondsw | need to go back and reread it though | 14:15 |
| *** k0da has quit IRC | 14:55 | |
| *** chhagarw has quit IRC | 15:02 | |
| *** esberglu has joined #openstack-powervm | 15:11 | |
| efried | edmondsw: Have we ever heard of "the metadata service"? | 15:34 |
| edmondsw | yes | 15:34 |
| edmondsw | it's an alternative to config_drive | 15:34 |
| edmondsw | where the instance talks back over the network to nova's metadata service | 15:34 |
| edmondsw | has security issues | 15:35 |
| efried | So not a solution for our slot manager? Swift is... somehow better than that? | 15:35 |
| edmondsw | and assumes that guests have network connectivity to the metadata service | 15:35 |
| efried | Sounds like it's a 169. address, so not a true networky thing | 15:36 |
| edmondsw | that doesn't sound right | 15:36 |
| edmondsw | I don't know if anyone ever considered using it for slot info | 15:37 |
| edmondsw | we always use config drive | 15:37 |
| edmondsw | I think there's even been talk of the metadata service going away at some points, but it never has | 15:37 |
| efried | We don't use config drive to save slot info. | 15:37 |
| *** apearson has quit IRC | 15:37 | |
| efried | Let's keep an eye open. It sounds like there may be an idea to provide a virt driver-accessible storage space where we could conceivably save the slot info. And not need Swift. | 15:37 |
| edmondsw | right... we use it for things that the metadata service could otherwise be used for | 15:37 |
| edmondsw | efried that would certainly be worth keeping an eye open for | 15:38 |
| edmondsw | but today I'm pretty sure the metadata service is read-only | 15:38 |
| efried | Somebody must be able to write it :) | 15:39 |
| *** apearson has joined #openstack-powervm | 15:40 | |
| edmondsw | sure, nova can, but not guests | 15:40 |
| efried | Guests wouldn't need to, for slot info. | 15:42 |
| edmondsw | true... nova-cpu would need to, though | 15:43 |
| edmondsw | I don't know that it can today, but maybe that's solvable | 15:43 |
| edmondsw | I'm sure it _doesn't_ today | 15:43 |
| edmondsw | efried https://docs.openstack.org/nova/latest/user/vendordata.html | 15:44 |
| efried | edmondsw: https://notartom.net/2017/06/20/virtual-device-role-tagging-better-explained/ | 15:45 |
| edmondsw | efried tagging is something that we should look into | 15:46 |
| edmondsw | just hasn't been a priority yet | 15:47 |
| efried | k. I don't know from priorities. I just tell you people things I hear that sound interesting. | 15:49 |
| edmondsw | yep, that's definitely interesting, and will definitely be needed at some point | 15:51 |
| edmondsw | efried interestingly, this explains the metadata service better than the nova docs :) https://docs.openstack.org/dragonflow/latest/specs/metadata_service.html | 15:56 |
| edmondsw | including why you saw 169 | 15:56 |
| efried | That's what artom said when he sent me the link :) | 15:56 |
| efried | oh, your link | 15:57 |
| *** apearson has quit IRC | 15:57 | |
| edmondsw | honestly, I haven't looked at the metadata service in years... I asked mikal about it at the last PTG because I was starting to think maybe it deserved a look, and he scared me off | 15:58 |
| efried | heh | 15:58 |
| edmondsw | as I remember it, he basically said "you don't want to do that" | 15:58 |
| edmondsw | my cache has purged the why | 15:59 |
| *** efried has quit IRC | 16:04 | |
| *** AlexeyAbashkin has quit IRC | 16:19 | |
| *** apearson has joined #openstack-powervm | 16:37 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 17:42 | |
| *** AlexeyAbashkin has quit IRC | 17:46 | |
| edmondsw | esberglu do you think we can remove WIP from https://review.openstack.org/#/c/546813/ soon? | 17:51 |
| esberglu | edmondsw: I'm trying to get neo40 installed so I can test the OVS side of things | 17:51 |
| edmondsw | cool | 17:51 |
| esberglu | As far as SEA detach seems to be working | 17:52 |
| esberglu | And I can see the interface added when I do attach | 17:52 |
| esberglu | But idk the root password for the image I'm using to update the network conf and make sure I actually can connect to the instance | 17:53 |
| edmondsw | where'd you get the image? | 17:53 |
| esberglu | edmondsw: I was using the one from CI | 17:53 |
| edmondsw | adreznec would you know the password for that? | 17:54 |
| edmondsw | don't type it here, of course, but if you know it (or who would) please ping esberglu | 17:54 |
| adreznec | esberglu: This is the Ubuntu 16.04 image? | 17:57 |
| esberglu | adreznec: Yep | 17:57 |
| *** AlexeyAbashkin has joined #openstack-powervm | 17:58 | |
| *** AlexeyAbashkin has quit IRC | 18:03 | |
| esberglu | edmondsw: I put up a localdisk WIP change. It's not actually ready for review, but I just got an IT localdisk spawn/delete to work | 18:47 |
| *** zerick_ has quit IRC | 18:48 | |
| edmondsw | esberglu saw that. You're really knocking these out. I love it | 18:49 |
| *** apearson has quit IRC | 18:52 | |
| *** zerick has joined #openstack-powervm | 18:53 | |
| *** apearson has joined #openstack-powervm | 18:54 | |
| esberglu | edmondsw: I'm really hoping we can get everything up to migration in really early before the cores get busy | 18:57 |
| edmondsw | esberglu totally | 18:57 |
| edmondsw | or at least space things out better than we did in Queens | 18:58 |
| esberglu | edmondsw: You OK to to W+1 these? | 18:59 |
| esberglu | https://review.openstack.org/#/c/548410/ | 18:59 |
| esberglu | https://review.openstack.org/#/c/548412/ | 18:59 |
| esberglu | Since efried is gonna be out next week | 18:59 |
| edmondsw | esberglu is there a reason we can't wait? | 19:01 |
| edmondsw | if they're not bothering anything... | 19:01 |
| esberglu | edmondsw: Doesn't matter to me, thought they were safe enough you wouldn't mind but no hurry | 19:03 |
| edmondsw | I'd be fine pushing them if we needed to, but I'm inclined to wait if there's nothing pushing us to merge them | 19:04 |
| esberglu | cool sounds good | 19:04 |
| esberglu | edmondsw: Regarding https://review.openstack.org/#/c/548721/ | 19:05 |
| esberglu | The previous test (test_remove_block_dev) was mocking sleep | 19:05 |
| esberglu | But it was still taking 10s vs 1s locally | 19:05 |
| esberglu | Not sure what I'm missing there | 19:06 |
| edmondsw | we're doing this to save 10s? | 19:06 |
| edmondsw | 9s? | 19:06 |
| esberglu | edmondsw: yeah | 19:07 |
| edmondsw | ok... :) | 19:07 |
| edmondsw | that doesn't seem to match the commit saying it's a "very long running" test | 19:08 |
| edmondsw | I'll be interested in efried's thoughts on that one | 19:08 |
| esberglu | I mean the next longest test takes like 1.5s, so long running relative to the rest | 19:08 |
| edmondsw | ic | 19:08 |
| esberglu | I'll just sit on that for now, see what efried says | 19:10 |
| edmondsw | yep | 19:12 |
| edmondsw | I threw a comment in there for him | 19:12 |
| *** AlexeyAbashkin has joined #openstack-powervm | 20:58 | |
| *** AlexeyAbashkin has quit IRC | 21:02 | |
| *** esberglu has quit IRC | 21:23 | |
| *** esberglu has joined #openstack-powervm | 21:24 | |
| *** esberglu has quit IRC | 21:28 | |
| *** esberglu has joined #openstack-powervm | 21:35 | |
| *** esberglu has quit IRC | 22:44 | |
| *** esberglu has joined #openstack-powervm | 22:45 | |
| *** esberglu has quit IRC | 22:50 | |
| *** edmondsw has quit IRC | 23:15 | |
| *** edmondsw has joined #openstack-powervm | 23:15 | |
| *** edmondsw has quit IRC | 23:20 | |
| *** apearson has quit IRC | 23:30 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!