| opendevreview | Takashi Kajinami proposed openstack/ironic-inspector master: [coordination] backend_url should be secret https://review.opendev.org/c/openstack/ironic-inspector/+/903638 | 06:42 |
|---|---|---|
| opendevreview | Takashi Kajinami proposed openstack/networking-generic-switch master: [coordination] backend_url should be secret https://review.opendev.org/c/openstack/networking-generic-switch/+/903639 | 06:44 |
| opendevreview | Takashi Kajinami proposed openstack/networking-generic-switch master: [ngs_coordination] backend_url should be secret https://review.opendev.org/c/openstack/networking-generic-switch/+/903639 | 06:45 |
| opendevreview | Merged openstack/ironic master: Revise ramdisk boot interface for clarity https://review.opendev.org/c/openstack/ironic/+/903621 | 07:39 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: CI: use Swift for configdrive when available https://review.opendev.org/c/openstack/ironic/+/903642 | 08:04 |
| rpittau | good morning ironic! o/ | 08:10 |
| opendevreview | Damien RANNOU proposed openstack/ironic-python-agent master: [WIP] USB device discovery https://review.opendev.org/c/openstack/ironic-python-agent/+/902769 | 08:19 |
| dtantsur | iurygregory: do you remember why we don't cache BIOS settings and firmware versions on the transition to manageable (like we do for vendor and boot mode)? | 08:39 |
| dtantsur | Cleaning may be disabled, and anyway it's a bit late at that point.. | 08:40 |
| dtantsur | I also wonder if we should add caching these to the inspection code path.. It is, in a sense, inspection. | 08:40 |
| dtantsur | My bad, misread the code. We do cache BIOS settings on verify. Not caching firmware settings is probably an omission. | 08:42 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Cache firwmare components on the transition to "manageable" https://review.opendev.org/c/openstack/ironic/+/903646 | 08:52 |
| dtantsur | iurygregory: ^^^ | 08:52 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [WIP] Test redfish with reduced sushy-tools feature set https://review.opendev.org/c/openstack/ironic/+/903361 | 08:52 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [DNM] Break redfish baseline https://review.opendev.org/c/openstack/ironic/+/903588 | 08:52 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [DNM] Break redfish baseline https://review.opendev.org/c/openstack/ironic/+/903588 | 08:53 |
| drannou | dtantsur: I get your point about my code about USB, just moving the caller from hardware.py to inspcetor would be enough on your POV ? | 09:15 |
| dtantsur | drannou: check how it's done for PCI devices. It's a plugin, it has its own entry point in setup.cfg. | 09:51 |
| dtantsur | mraineri: hi! is this change correct from your perspective (the Drives one)? https://review.opendev.org/c/openstack/ironic/+/902744/5/ironic/drivers/modules/redfish/raid.py#619 | 09:55 |
| dtantsur | I wonder why it was done without Links initially. | 09:55 |
| dtantsur | I see that someone has an example without Links: https://jonamiki.com/2019/04/14/example-redfish-rest-calls-create-raid-volume/ | 09:57 |
| opendevreview | Merged openstack/networking-generic-switch master: [ngs_coordination] backend_url should be secret https://review.opendev.org/c/openstack/networking-generic-switch/+/903639 | 10:04 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Configure the new inspection implementation https://review.opendev.org/c/openstack/bifrost/+/896925 | 10:48 |
| opendevreview | Verification of a change to openstack/ironic-inspector master failed: [coordination] backend_url should be secret https://review.opendev.org/c/openstack/ironic-inspector/+/903638 | 10:49 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Configure the new inspection implementation https://review.opendev.org/c/openstack/bifrost/+/896925 | 10:55 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Configure the new inspection implementation https://review.opendev.org/c/openstack/bifrost/+/896925 | 11:06 |
| iurygregory | good morning Ironic | 11:08 |
| iurygregory | dtantsur, in the transition to manageable you mean from enroll to it right? | 11:11 |
| dtantsur | yep | 11:11 |
| iurygregory | I probably didn't add in verifying because "oh this is just cleaning step/ deploy step" so I shouldn't touch things in verify | 11:15 |
| iurygregory | I agree it makes sense to check the information and cache it | 11:15 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Drop support for rootwrap https://review.opendev.org/c/openstack/bifrost/+/903671 | 11:57 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: DNM test the CI https://review.opendev.org/c/openstack/bifrost/+/903677 | 13:10 |
| opendevreview | Merged openstack/ironic-inspector master: [coordination] backend_url should be secret https://review.opendev.org/c/openstack/ironic-inspector/+/903638 | 14:07 |
| opendevreview | Damien RANNOU proposed openstack/ironic-python-agent master: [WIP] USB device discovery https://review.opendev.org/c/openstack/ironic-python-agent/+/902769 | 14:08 |
| opendevreview | Damien RANNOU proposed openstack/ironic-python-agent master: [WIP] USB device discovery https://review.opendev.org/c/openstack/ironic-python-agent/+/902769 | 14:21 |
| opendevreview | Merged openstack/ironic master: Cache firwmare components on the transition to "manageable" https://review.opendev.org/c/openstack/ironic/+/903646 | 14:43 |
| JayF | \o/ the entire openstacksdk migration chain in ironic<>nova driver finished landing this morning | 15:08 |
| iurygregory | \o/ | 15:20 |
| jamesdenton_ | Good morning, all. Question about bifrost: is the schema for node_network_data defined anywhere? | 15:32 |
| JayF | Hey, it takes the standard openstack network data that cloud-init/nova/etc all interchanges | 15:36 |
| JayF | I believe the best doc we have for it is the original nova spec | 15:36 |
| jamesdenton_ | got it, thank you | 15:37 |
| *** jamesdenton_ is now known as jamesdenton | 15:37 | |
| JayF | did you find the url? | 15:37 |
| JayF | I'm looking for it :d | 15:37 |
| JayF | https://github.com/openstack/nova-specs/blob/master/specs/kilo/approved/metadata-service-network-info.rst is essentially it | 15:37 |
| JayF | jamesdenton: I'll note that feature generally has some rough edges | 15:37 |
| jamesdenton | i found this: https://docs.openstack.org/nova/latest/_downloads/9119ca7ac90aa2990e762c08baea3a36/network_data.json | 15:38 |
| JayF | jamesdenton: mainly around having to build your ramdisks to explicitly support network data | 15:38 |
| jamesdenton | gotcha | 15:38 |
| JayF | which insures for provisioning actions we use the node.network_data and not, for instance, the leftover configdrive from the last tenant instance :) | 15:39 |
| TheJulia | glean also has some awesome examples | 15:39 |
| jamesdenton | i'm interested in cloud-init generating the netplan configuration, including bonds. Not sure what that looks like just yet | 15:39 |
| TheJulia | of the network metadata payloads | 15:39 |
| JayF | Julia is the expert in this, she's been playing exterminator on this feature for the last month :) | 15:40 |
| jamesdenton | :) | 15:40 |
| JayF | s/month/year/ probably is more accurate lol :) | 15:40 |
| TheJulia | I'm not sure how well it knows to transform that into, at least depending on the network metadata | 15:40 |
| TheJulia | cloud-init has some funky logic around interface launch because... being a cloud tool you don't need a bond so it just looks to fire up dhcp by default on interfaces | 15:40 |
| jamesdenton | that makes sense | 15:41 |
| TheJulia | ... there is a flag for that though! | 15:41 |
| jamesdenton | of course there is :D | 15:41 |
| TheJulia | give me a few minutes and I'll see if I can hunt it down for you | 15:41 |
| TheJulia | I need to finish another thought first | 15:41 |
| jamesdenton | no rush at all | 15:41 |
| jamesdenton | as an alternative, is it possible to just shove user_data_content in there (which could include a cloud-init friendly network block) or is there a better alternative to user-servicable user-data? | 16:23 |
| JayF | So node.network_data is for *provisioning actions* -- e.g. "I need this node to have a strange/unique network config during deployment/rescue/cleaning/etc" | 16:29 |
| JayF | network/user/metadata for instances is done via configdrive | 16:30 |
| JayF | https://docs.openstack.org/ironic/latest/user/deploy.html#deploy-configdrive | 16:30 |
| * JayF not sure if he understood your statement | 16:31 | |
| jamesdenton | nope, this is perfect, thanks. I had attempted to use --configdrive but hadn't constructed the json properly | 16:31 |
| jamesdenton | btw - not sure if this has been fixed, but it seems like my bifrost-cli wants --configdrive not --config-drive | 16:32 |
| JayF | So those are ironic docs | 16:32 |
| JayF | bifrost is ansible for driving ironic | 16:32 |
| JayF | so the concepts should map but they won't be exact 1:1 wordings | 16:32 |
| jamesdenton | ahh, right, thanks | 16:32 |
| JayF | Oooh that means configdrive support might be different. hmm | 16:32 |
| JayF | https://github.com/openstack/bifrost/blob/master/playbooks/roles/bifrost-configdrives-dynamic/README.md | 16:33 |
| JayF | hey, that looks relevant :D | 16:33 |
| JayF | https://docs.openstack.org/bifrost/latest/user/howto.html#deployment-and-configuration-of-operating-systems | 16:33 |
| JayF | Looks like you get a really simple configdrive by default | 16:33 |
| TheJulia | jamesdenton: https://review.opendev.org/c/openstack/diskimage-builder/+/899886/1/diskimage_builder/elements/cloud-init-datasources/install.d/05-set-cloud-init-sources#41 what you need to know if you don't have DHCP in place | 16:33 |
| JayF | should be able to use that as a template to get a more advanced configdrive going | 16:34 |
| jamesdenton | i've got some reading to do. thank you both | 16:35 |
| JayF | No problem :) Happy to help, there are a lot of knobs to turn | 16:35 |
| TheJulia | tons | 16:36 |
| TheJulia | in various ways | 16:36 |
| TheJulia | and some knobs don't agree with other knobs | 16:36 |
| jamesdenton | i've got a pretty stable, simple deployment method going on here with bifrost at the moment, with some post-deploy ansible that does the netplan and other cloud-inity things, but would rather it all be done at the time of deploy | 16:40 |
| jamesdenton | and it's not easy to walk away from this for a few weeks/months and expect to retain the nuances | 16:40 |
| JayF | I usually recommend a pattern of using cloud-init/configdrive to bootstrap into a higher level config management | 16:43 |
| jamesdenton | not a bad approach | 16:43 |
| JayF | but I'll note in this case, you could potentially approach it like ... never leaving ansible in the first place. I've never done it; but I suspect it's possible to append your own playbooks to the end of bifrosts' work | 16:44 |
| JayF | so when it sees the node go active, it can execute a playbook to try to ssh and do the config | 16:44 |
| TheJulia | JayF: designed so you can do that :) | 16:44 |
| JayF | so you can oneshot it all in ansible | 16:44 |
| TheJulia | and just let ansible do the needful, it is why the examples have login/actions locally | 16:44 |
| jamesdenton | yep, today it's two steps but ideally one | 16:44 |
| jamesdenton | i just need to clean up my inventory generation processes | 16:44 |
| TheJulia | And that can also be done with a plugin too | 16:45 |
| TheJulia | or from an existing ironic :) | 16:45 |
| rpittau | good night! o/ | 17:05 |
| iurygregory | wow, OpenInfra Summit Asia O.o | 19:33 |
| samcat116 | FWIW I was never able to successfully use the network_data and virtual media to deploy a node with a bond | 19:33 |
| samcat116 | But I didn't realize the prebuilt ipa images don't support it | 19:34 |
| dtantsur | iurygregory: and another one in CERN, we must organize a team meeting next to it! | 19:34 |
| TheJulia | you and some board members want to visit there. I will note, no shutdown this year :) | 19:35 |
| iurygregory | dtantsur, ++! | 19:35 |
| dtantsur | I should be able to convince my wife that we need a short vacation in Geneva | 19:36 |
| dtantsur | especially if arne_wiebalck organizes a new visit to the experiments | 19:36 |
| iurygregory | for me is more complicated lol | 19:37 |
| dtantsur | yeah, I can imagine | 19:37 |
| iurygregory | but maybe with a talk a I have a chance to go to the Summit lol | 19:37 |
| dtantsur | South Korea? Could be cool, budget... sigh. | 19:39 |
| iurygregory | yup! | 19:40 |
| iurygregory | budget is the key | 19:40 |
| iurygregory | but, on the bright side, it's in Sept... so maybe things are cheaper lol | 19:41 |
| dtantsur | Possibly? I'd expect September to be nice time there | 19:41 |
| iurygregory | I should try to do some Open Infra Day here in Brazil lol | 19:46 |
| arne_wiebalck | we plan to do tours, absolutely :) | 20:31 |
| TheJulia | iurygregory: oh very yes | 20:39 |
| TheJulia | iurygregory: ferment/grow a community | 20:39 |
| TheJulia | Ever watch Stargate? | 20:39 |
| JayF | Can someone spare 5-10 minutes to quack at a unit test failure with me real quick? | 20:41 |
| iurygregory | re Stargate - no .-. | 20:42 |
| JayF | I'm thinking it's gotta be something silly I'm missing | 20:42 |
| JayF | rebasing https://review.opendev.org/c/openstack/nova/+/900831/ against the openstacksdk migration stuff, and just can't get a test to pass | 20:42 |
| * iurygregory looks | 20:42 | |
| JayF | looks like _refresh_cache is swallowing an exception, but I can't tell how/where | 20:42 |
| JayF | iurygregory: it's local | 20:42 |
| JayF | we'll have to zoom if you wanna see | 20:42 |
| TheJulia | iurygregory: we need https://stargate.fandom.com/wiki/Replicator to build a community :) | 20:42 |
| JayF | I'm mid-broken-rebase :) | 20:42 |
| iurygregory | JayF, so you are rebasing 900831 on top of master and trying to solve the conflict that shows during rebase right? | 20:43 |
| JayF | well the conflicts are solved | 20:44 |
| JayF | but the tests are failing | 20:44 |
| JayF | and I can't get it figured out how | 20:44 |
| JayF | essentially, I have a mock setup to make _refresh_cache raise an exception | 20:44 |
| JayF | but it's not causing list_instances to throw the exception | 20:44 |
| JayF | so it's getting swallowed *somewhere* | 20:44 |
| JayF | and I can't find where | 20:44 |
| iurygregory | TheJulia, awesome! In the past we used to have OpenStack Days Sao Paulo in Brazil, so I'm in touch with ppl from other companies to see if we can organize Open Infra Days =) | 20:44 |
| iurygregory | JayF, hummm | 20:45 |
| JayF | https://us06web.zoom.us/j/89282955673?pwd=KU5NPqW8BqIa1xbSNWnlSFAmhs97M6.1 | 20:45 |
| JayF | if you wanna see | 20:45 |
| iurygregory | going | 20:45 |
| iurygregory | let me grab my headphone | 20:46 |
| JayF | iurygregory++ I had swapped argument order of a mock in the test | 21:03 |
| iurygregory | \o/ | 21:04 |
| *** dmellado2 is now known as dmelladoo | 21:55 | |
| *** dmelladoo is now known as dmellado | 21:58 | |
| dmellado | iurygregory not watching stargate should be a crime xD | 22:01 |
| TheJulia | samcat116: yeah, I think I saw some of why you had issues too even beyond that, with how cloud-init and simple-init can end up in some images together if you build them | 22:02 |
| TheJulia | samcat116: and while yes, you can use either, cloud-init has some hard expectations out of the box about dhcp being present on the first ethernet interface as if it is a VM which breaks things unless you tell it not to in the image deployed | 22:02 |
| iurygregory | dmellado, lol | 22:04 |
| iurygregory | I will try to find time during my PTO in Jan =P | 22:04 |
| TheJulia | Also, all star trek mirror universe episodes | 22:04 |
| TheJulia | That, too, is crime to have not watched. :) | 22:04 |
| iurygregory | Star Trek I've watched all of them I think | 22:04 |
| dmellado | that said, I'd avoid the tv series with McGyver xD | 22:07 |
| dmellado | but that's just my .02 | 22:07 |
| dmellado | pretty much prefer the film | 22:07 |
| JayF | iurygregory: we can form a no-stargate-club | 22:10 |
| * JayF has not seen the films/tvs either | 22:11 | |
| iurygregory | JayF, but according to TheJulia we need stargate to build a community, so better avoid the no-stargate-club lol | 22:13 |
| JayF | for every culture there is a counterculture, definitionally! | 22:13 |
| JayF | our mere existence helps proves her point ;) | 22:13 |
| TheJulia | iurygregory: we just need the replicators | 22:13 |
| TheJulia | And counter culture ++ | 22:14 |
| iurygregory | agree | 22:14 |
| dmellado | xDDD | 22:26 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!