| *** tetsuro has joined #openstack-placement | 05:56 | |
| *** ttsiouts has joined #openstack-placement | 06:23 | |
| openstackgerrit | Merged openstack/placement master: Remove unused import statement. https://review.opendev.org/688026 | 07:53 |
|---|---|---|
| *** e0ne has joined #openstack-placement | 08:00 | |
| *** tetsuro has quit IRC | 09:08 | |
| *** cdent has joined #openstack-placement | 09:29 | |
| *** openstackgerrit has quit IRC | 09:37 | |
| *** ttsiouts has quit IRC | 10:20 | |
| *** ttsiouts has joined #openstack-placement | 10:20 | |
| *** ttsiouts has quit IRC | 10:24 | |
| *** ttsiouts has joined #openstack-placement | 10:50 | |
| *** tssurya has joined #openstack-placement | 10:51 | |
| *** ttsiouts has quit IRC | 10:55 | |
| *** ttsiouts has joined #openstack-placement | 11:13 | |
| *** ttsiouts has quit IRC | 11:28 | |
| *** ttsiouts has joined #openstack-placement | 11:29 | |
| *** ttsiouts has quit IRC | 11:33 | |
| *** ttsiouts has joined #openstack-placement | 11:55 | |
| efried | cdent: Based on a real quick skim: | 13:06 |
| efried | microversion_parse has middleware to check for microversion headers | 13:06 |
| efried | nova doesn't use it | 13:06 |
| efried | nova does its own microversion header parsing & validation | 13:06 |
| efried | nova could cut over to microversion_parse | 13:06 |
| efried | ? | 13:06 |
| cdent | efried: I tried to cut over to microversion parse at some point and was unable to get support for doing so from the conservatives, so gave up | 13:07 |
| cdent | times have since changed, so I'd say it ought to be possible | 13:07 |
| efried | okay, the problem was political, not technical. | 13:08 |
| cdent | as I recall, yes | 13:08 |
| efried | Did you get as far as proposing code? | 13:08 |
| cdent | I think there was a change that added it for the simple parts, but not the middleware: I9098ab8f0800fb0550887f5eff219b18addde709 | 13:09 |
| cdent | https://review.opendev.org/#/c/300076/ | 13:10 |
| cdent | that commit mentions further patches to do the rest, but ... | 13:12 |
| cdent | I can't find an abandoned anything | 13:14 |
| efried | Okay. This kind of tech debt reduction doesn't get a huge amount of traction, but IMO it's worth doing. So if we move forward with the adoption of mvp by oslo, I think it would be worth resurrecting the effort. | 13:14 |
| * cdent nods | 13:15 | |
| gibi | if this means we delete code from nova then count me in | 13:46 |
| efried | cdent: how does the placement upgrade path work for sucking in new traits? | 13:56 |
| efried | I have placement and os-traits at train-ish versions... | 13:56 |
| efried | and I want to start using the new COMPUTE_NODE trait in my deployment | 13:56 |
| efried | how do I make placement know about COMPUTE_NODE? | 13:57 |
| efried | upgrade os-traits to 1.1.0 and then...? | 13:57 |
| gibi | does restarting placment service syncs the traits from os-traits? | 13:57 |
| efried | ...yup, that does it, thanks gibi. | 13:58 |
| efried | next q, did we ever merge the thing that lets you add a trait without specifying all the traits? | 13:59 |
| efried | ahem, why don't I see resource provider stuff on https://docs.openstack.org/python-openstackclient/latest/cli/command-list.html ? | 14:01 |
| *** mriedem has joined #openstack-placement | 14:20 | |
| gibi | efried: I guess because placement cli is defined in osc-placement | 14:21 |
| efried | Okay, I found it now. That's kind of esoteric. If I'm trying to find the `openstack` command to do a thing, I would expect to find it in the same reference as the other `openstack` commands. I'm making excuses for my own ignorance here, but I doubt I'm the only one who's been confused by this. | 14:26 |
| efried | so, ugh, it looks like we really don't have a way to add a trait?? | 14:27 |
| *** ttsiouts has quit IRC | 14:35 | |
| *** ttsiouts has joined #openstack-placement | 14:35 | |
| *** ttsiouts has quit IRC | 14:40 | |
| efried | mriedem, cdent: Would you please weigh in on backward-compatible suggestions for osc-placement ways to modify the trait list https://storyboard.openstack.org//#!/story/2005258 | 14:42 |
| mriedem | i feel like i already opened a bug about this once | 14:49 |
| mriedem | checking | 14:49 |
| mriedem | oh heh that's mine | 14:49 |
| mriedem | efried: fwiw i think this is supposed to be the main osc docs page for plugin commands and placement isn't listed there, not sure if that's a bug or what https://docs.openstack.org/python-openstackclient/latest/cli/plugin-commands.html | 14:50 |
| *** dklyle has joined #openstack-placement | 14:50 | |
| efried | mriedem: ack. IMO it's a bug from a ux perspective that I can't find placement commands easily from that ref | 14:51 |
| mriedem | commented in storyboard but i'd follow what melwitt did in train with adding --amend to openstack resource provider inventory set: https://docs.openstack.org/osc-placement/train/cli/index.html#resource-provider-inventory-set | 14:53 |
| mriedem | that would at least be consistent | 14:54 |
| efried | mriedem: ack. But we need to provide a way to remove a trait, which we don't have for inventories. So --amend --[no-]trait seems reasonable. | 14:55 |
| efried | I'm PoCing it. | 14:55 |
| mriedem | i think that's pretty gross, | 14:56 |
| mriedem | just add an "openstack resource provider trait unset" command | 14:56 |
| mriedem | that's closer to like how you can remove specific metadata items from a server in osc: | 14:56 |
| mriedem | https://docs.openstack.org/python-openstackclient/latest/cli/command-objects/server.html#server-unset | 14:56 |
| mriedem | or, "openstack resource provider unset --trait x --trait y" | 14:57 |
| efried | but then set and unset are asymmetrical, which is more grosserer. | 14:58 |
| mriedem | so do trait set and trait unset i guess | 14:58 |
| mriedem | but --no-trait makes me want to explode | 14:58 |
| mriedem | i was just thinking about how "openstack resource provider trait set --trait" is redundant, and we should have just done "openstack resource provider set --trait" | 14:59 |
| efried | that was what dtroyer suggested. | 14:59 |
| mriedem | but the CLIs have mostly followed the API resource methods | 15:00 |
| mriedem | ok i guess openstack port set has things like --tag and --no-tag, and also has a port unset --tag option, so there are multiple ways to skin that cat in osc it seems, so whatever dean wants i guess | 15:01 |
| *** ttsiouts has joined #openstack-placement | 15:04 | |
| cdent | mriedem, efried: As I've tried to disclaim many times: Asking me for osc naming suggestions is fruitless. I simply don't grok the original rules that we are supposed to be trying to follow. | 15:14 |
| cdent | So in the end finding something similar in one of the others services and using that as a precedent seems the best bet | 15:14 |
| efried | ack | 15:14 |
| *** artom has joined #openstack-placement | 15:19 | |
| efried | oic, --no-tag doesn't remove a tag, it removes all tags :( | 15:19 |
| efried | another option is --trait !TRAIT_TO_REMOVE | 15:19 |
| efried | but that sucks | 15:20 |
| efried | esp because don't you have to escape a bang on a bash cli? | 15:20 |
| * cdent nos | 15:20 | |
| cdent | nods | 15:20 |
| * cdent wants a new keyboard | 15:20 | |
| cdent | and hands | 15:20 |
| efried | well f, I guess I'll throw this up to show that it works, and then we can bikeshed over the syntax | 15:22 |
| *** openstackgerrit has joined #openstack-placement | 15:30 | |
| openstackgerrit | Eric Fried proposed openstack/osc-placement master: WIP: resource provider trait set --amend --[no-]trait https://review.opendev.org/689823 | 15:30 |
| efried | ^ tested locally, works fine. Added some comments to start the discussion on the syntax. | 15:44 |
| *** ttsiouts has quit IRC | 15:48 | |
| *** ttsiouts has joined #openstack-placement | 15:48 | |
| *** tssurya has quit IRC | 15:52 | |
| *** ttsiouts has quit IRC | 15:53 | |
| *** e0ne has quit IRC | 15:58 | |
| *** cdent has quit IRC | 16:03 | |
| *** N3l1x has joined #openstack-placement | 18:11 | |
| *** e0ne has joined #openstack-placement | 18:15 | |
| *** e0ne has quit IRC | 18:38 | |
| *** e0ne has joined #openstack-placement | 18:45 | |
| *** artom has quit IRC | 19:54 | |
| *** mriedem is now known as mriedem_afk | 20:05 | |
| *** spatel has joined #openstack-placement | 20:08 | |
| *** spatel has quit IRC | 20:51 | |
| efried | cdent: Eric Fried proposed openstack/python-openstackclient master: Add placement to known plugins https://review.opendev.org/689874 | 20:52 |
| efried | Note that this doesn't help the fact that https://docs.openstack.org/python-openstackclient/latest/cli/plugin-commands.html is still worthless -- https://storyboard.openstack.org/#!/story/1735016 exists for that. | 20:53 |
| *** mriedem_afk is now known as mriedem | 20:54 | |
| *** e0ne has quit IRC | 21:13 | |
| *** e0ne has joined #openstack-placement | 22:01 | |
| *** openstackgerrit has quit IRC | 22:07 | |
| *** e0ne has quit IRC | 22:16 | |
| *** e0ne has joined #openstack-placement | 22:18 | |
| *** dklyle has quit IRC | 22:42 | |
| *** david-lyle has joined #openstack-placement | 22:42 | |
| *** artom has joined #openstack-placement | 23:08 | |
| *** e0ne has quit IRC | 23:58 | |
| *** e0ne has joined #openstack-placement | 23:58 | |
| *** e0ne has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!