| zigo | It's nice, but I think I'd prefer one directive per file type. | 00:00 |
|---|---|---|
| zigo | [pxe] | 00:01 |
| zigo | ipxe_efi_path= | 00:01 |
| zigo | bootx86_efi_path= | 00:01 |
| zigo | grubx86_efi_path= | 00:01 |
| zigo | The reason is: tools like puppet and so on can more granularly edit these. | 00:01 |
| zigo | Then at the distro level, I'd just add this etc/ironic/ironic-dist-deb-family.conf in the namespace list, rather than a specific .conf file to load... | 00:02 |
| zigo | As in, I'd be happy to have Debian specific nice defaults, rather than just something wrong. | 00:02 |
| zigo | But I don't think adding another custom config file to the --config-file list is a good idea (this would mean modifying the startup init / systemd script...). | 00:03 |
| zigo | Just my 2 cents, hoping this helps ... :) | 00:04 |
| JayF | Those loader paths are already in code, and have been. Changing them would require deprecation period, etc. I think that stevebaker[m]'s change is more about documenting known-good-values. | 00:15 |
| JayF | stevebaker[m]: That is a /super sweet change/ and I love it in theory. Will review it now. | 00:15 |
| stevebaker[m] | zigo: the config interface is already defined, so it will be loader_file_paths. If a deployer wants to have a custom value they'll need to duplicate all of loader_file_paths with their changes. Its not perfect but I think it strikes a balance of just-working for defaults, but still customizable | 00:19 |
| zigo | ok | 00:20 |
| zigo | I very much appreciate the effort. | 00:21 |
| stevebaker[m] | zigo: what does it mean to add to a namespace list? It sounds better than adding a new --config-file argument | 00:21 |
| stevebaker[m] | JayF: thank you! | 00:21 |
| zigo | stevebaker[m]: I meant adding it here: https://salsa.debian.org/openstack-team/services/ironic/-/blob/debian/xena/debian/rules#L42 | 00:22 |
| zigo | The idea was: depending on which distro, one would load different sets of default for the same directive ... | 00:22 |
| zigo | (so different distro could generate different ironic.conf ...) | 00:23 |
| zigo | That's IMO nicer than a distro-specific config file to load. | 00:23 |
| JayF | zigo: stevebaker[m]: My comment was trying to get at the idea of something like could we make oslo-config-generator output the right default given an argument? | 00:23 |
| zigo | JayF: Yeah, that's my idea ... | 00:24 |
| JayF | Admittedly, that'd be a lot more useful if Ironic still shipped a conf.sample in-tree. It's not super discoverable for someone, on the internet, if it's only output from oslo-config-generator | 00:24 |
| JayF | just an unfortunate side effect of how configs have been generated for a while | 00:25 |
| zigo | Something like: --namespace ironic.conf.pxe.debian vs --namespace ironic.conf.pxe.fedora | 00:25 |
| JayF | but tbh, this is one of those places where any idea is better than not doing it at all | 00:25 |
| JayF | so take my feedback, at least, as just very advisory | 00:25 |
| zigo | JayF: I'm very much OK with OpenStack services not having samples by default, considering it's bound to their dependencies. | 00:26 |
| stevebaker[m] | oh so these dist.conf files could be represented in python in ironic, now that would be interesting | 00:26 |
| JayF | zigo: gah, you're right | 00:26 |
| zigo | We had that discussion in the OpenStack community for a long time ... | 00:26 |
| zigo | (and a long time ago) | 00:26 |
| JayF | zigo: I was here for those discussions, and still felt like operators got a little boned in the end there, but I didn't have better ideas or time to argue about it at the time :D | 00:26 |
| zigo | :) | 00:27 |
| zigo | It's the jobs of distros to generate the .conf files, so operators have them anyways... | 00:27 |
| zigo | It's just probably a mater of documentation, and that can be addressed ... | 00:28 |
| JayF | Everywhere I've ever had Ironic deployed, we deployed either from git or from pypi. | 00:28 |
| zigo | Ah... | 00:28 |
| JayF | But I know my experience is at a different scale than the average user :) | 00:28 |
| zigo | There's a reason why distros have invented dependency and maintainer scripts, do you know ? :) | 00:29 |
| stevebaker[m] | I've called that "python packaged" in the commit message | 00:29 |
| zigo | Anyway, I should be sleeping at this time... | 00:30 |
| zigo | 2:30 am, and I get up at 8... | 00:30 |
| zigo | BFN | 00:30 |
| stevebaker[m] | \o | 00:31 |
| JayF | o/ | 00:32 |
| *** thelounge70 is now known as janders | 01:56 | |
| opendevreview | Nisha Agarwal proposed openstack/ironic master: Add nvme as interface_type for 'agent' raid interface https://review.opendev.org/c/openstack/ironic/+/815110 | 06:18 |
| iurygregory | good morning Ironic o/ | 06:23 |
| janders | hey iurygregory o/ | 06:24 |
| iurygregory | janders, o/ | 06:24 |
| arne_wiebalck | Good morning iurygregory janders and Ironic! | 06:38 |
| janders | hey arne_wiebalck o/ | 06:39 |
| iurygregory | arne_wiebalck, o/ | 06:47 |
| rpittau | good morning ironic! o/ | 07:23 |
| iurygregory | morning rpittau o/ | 07:23 |
| rpittau | hey iurygregory :) | 07:24 |
| janders | hey rpittau o/ | 07:46 |
| rpittau | hey janders :) | 07:46 |
| dtantsur | morning folks | 07:49 |
| janders | hey dtantsur | 07:51 |
| janders | I'm having a hard time trying to install bifrost on my new lab server. Hitting this: https://paste.openstack.org/show/810207/ | 07:58 |
| dtantsur | janders: there should be some cause before this | 07:59 |
| janders | Full story: the install went smooth the first time, but I was hitting some weird ipv6 issues (IPA was booting in VMs but failing with log errors complaining about name resolution). After disabling v6 and reboot it never quite started working again. I tried clearing the install and reinstalling and I hit this. | 08:01 |
| janders | I wonder if it may be some leftovers | 08:01 |
| opendevreview | Arun S A G proposed openstack/ironic master: Fix various issues in the anaconda deploy interface https://review.opendev.org/c/openstack/ironic/+/814087 | 08:22 |
| janders | dtantsur you are right: https://paste.openstack.org/show/810208/ | 08:34 |
| janders | I'm not sure what's happening there but it does feel like some fallout from IPv6 un-configuration | 08:34 |
| dtantsur | yep | 08:35 |
| dtantsur | it cannot bind to :: you need something else | 08:35 |
| janders | I'm not sure where it's getting that from though (I think I cleared the config) | 08:35 |
| dtantsur | :: may be the default? | 08:35 |
| janders | maybe I will kill this install, re-start it and see if there may be an earlier failure that prevents config from being written or something | 08:37 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Trivial: do not stop None rpcserver https://review.opendev.org/c/openstack/ironic/+/815422 | 08:39 |
| dtantsur | janders: ^^ | 08:39 |
| dtantsur | the default host_ip seems to be 0.0.0.0 | 08:42 |
| dtantsur | and bifrost is not IPv6 aware, so it shouldn't change anything | 08:42 |
| janders | dtantsur looks like I'm hitting this https://bugzilla.redhat.com/show_bug.cgi?id=1829071 | 09:08 |
| janders | (or something similar) | 09:10 |
| janders | sounds like using a less "drastic" way of getting rid of IPv6 addresses may be the go in this case | 09:13 |
| opendevreview | Bernd Mueller proposed openstack/ironic-python-agent-builder master: add stress-ng package to extra-hardware https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/815453 | 09:22 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent master: Fix error messages in burnin code https://review.opendev.org/c/openstack/ironic-python-agent/+/815455 | 09:33 |
| dtantsur | arne_wiebalck: FYI ^^^ | 09:33 |
| arne_wiebalck | dtantsur: +2, thanks | 09:36 |
| arne_wiebalck | dtantsur: what is our policy re s/w packages needed for clean steps? this is re Bernd's patch, do we need to ship all used packages all the time or do we leave this to the image builder to make sure the required packages are included? | 09:37 |
| dtantsur | arne_wiebalck: I'd vote for a new optional element | 09:38 |
| dtantsur | (I left a comment to this extent) | 09:38 |
| arne_wiebalck | oh, ok ... yeah, that sounds like a good option | 09:38 |
| arne_wiebalck | so, we do not include it by default, but make it easy to include if needed | 09:38 |
| dtantsur | yep | 09:38 |
| arne_wiebalck | +1 (great to see burn-in gets some usage :) | 09:40 |
| dtantsur | yeah, and so quickly after the introduction! | 09:40 |
| arne_wiebalck | we are working on the dynamic matching of nodes for the network burn-in, so pairing nodes explicitly will become optional (nodes will look for a benchmarking partner) | 09:49 |
| arne_wiebalck | maybe sth for one of the SIG talks | 09:50 |
| dtantsur | ++ | 09:51 |
| opendevreview | Bernd Mueller proposed openstack/ironic-python-agent-builder master: new element burn-in for package stress-ng https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/815463 | 10:16 |
| opendevreview | Bernd Mueller proposed openstack/ironic-python-agent-builder master: new element burn-in for package stress-ng https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/815463 | 10:17 |
| opendevreview | Dmitry Tantsur proposed openstack/sushy master: Migrate common constants to enums https://review.opendev.org/c/openstack/sushy/+/815107 | 11:04 |
| dtantsur | this ^^^ looks scary but actually pretty straightforward | 11:07 |
| *** dviroel|rover|out is now known as dviroel|rover | 11:11 | |
| *** tosky_ is now known as tosky | 11:25 | |
| arne_wiebalck | dtantsur: any particular reason why fio is in extra-hardware? this should probably move to the new burn-in element as well, no? | 11:41 |
| dtantsur | arne_wiebalck: I *think* it can also be used by python-hardware | 11:41 |
| opendevreview | Dmitry Tantsur proposed openstack/sushy master: Migrate Protocol constants to enums https://review.opendev.org/c/openstack/sushy/+/815475 | 11:41 |
| arne_wiebalck | dtantsur: oh, ok | 11:42 |
| jeeva | can bifrost related questions be asked here ? | 11:43 |
| dtantsur | jeeva: yes | 11:43 |
| jeeva | https://docs.openstack.org/kolla-ansible/wallaby/reference/deployment-and-bootstrapping/bifrost.html | 11:43 |
| jeeva | where does one get the UUID as per that doc ? | 11:43 |
| jeeva | "create a bifrost inventory" | 11:43 |
| dtantsur | jeeva: you can make it up or skip completely | 11:44 |
| jeeva | gave up trying to get kolla ironic going, trying the bifrost route now | 11:44 |
| dtantsur | (it's your future node UUID) | 11:44 |
| jeeva | ok thanks | 11:44 |
| dtantsur | jeeva: a small request: if you make it work, could you post any changes you needed to their docs? | 11:45 |
| opendevreview | Dmitry Tantsur proposed openstack/sushy master: Add some VirtualMedia fields from 1.3.0 and 1.4.0 https://review.opendev.org/c/openstack/sushy/+/814191 | 11:49 |
| opendevreview | Dmitry Tantsur proposed openstack/sushy master: WIP certificate collection https://review.opendev.org/c/openstack/sushy/+/815479 | 12:04 |
| opendevreview | Bernd Mueller proposed openstack/ironic-python-agent-builder master: new element burn-in for package stress-ng https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/815453 | 12:17 |
| TheJulia | good morning | 12:29 |
| rpittau | good morning TheJulia :) | 12:34 |
| opendevreview | Merged x/sushy-oem-idrac master: Update unit test folder structure https://review.opendev.org/c/x/sushy-oem-idrac/+/814919 | 12:44 |
| *** Guest3656 is now known as redrobot | 13:03 | |
| arne_wiebalck | Hey TheJulia o/ | 13:23 |
| arne_wiebalck | dtantsur: is extra-hardware guaranteed to be there? and: how does DIB handle double requests for the same package? | 13:23 |
| arne_wiebalck | dtantsur: what I am getting at is whether fio should be added to the element for burn-in in addition | 13:24 |
| dtantsur | arne_wiebalck: DIB should be fine with duplicated packages | 13:24 |
| arne_wiebalck | dtantsur: in that case I would suggest to add it twice | 13:25 |
| opendevreview | Bernd Mueller proposed openstack/ironic-python-agent-builder master: new element burn-in for package stress-ng, added fio https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/815453 | 13:38 |
| arne_wiebalck | it is the time of the year again where nodes do not boot since the IPA image is getting too large (centos stream 8) | 13:40 |
| arne_wiebalck | I checked the largest packages and have removed some big ones from a test image: geolite2-city libicu perl-interpreter python3-babel perl-Encode | 13:41 |
| arne_wiebalck | let me know if you think removing any of these is a bad idea | 13:41 |
| rpittau | python3-babel mmmm | 13:41 |
| arne_wiebalck | heh | 13:41 |
| arne_wiebalck | this shrunk the compressed image by 30MB which made it work again | 13:42 |
| rpittau | just wondering what's installing python3-babel, it's ok to remove, at least for ipa or python-hardware | 13:44 |
| arne_wiebalck | rpittau: mind you, this is not a 100% upstream image | 13:44 |
| rpittau | arne_wiebalck: gotcha | 13:44 |
| arne_wiebalck | rpittau: could be we somehow have sth in there that is not there usually | 13:45 |
| arne_wiebalck | the total size is now 390MB | 13:45 |
| arne_wiebalck | 420MB does not work | 13:45 |
| arne_wiebalck | I guess perl we do not need either | 13:46 |
| rpittau | yeah, I'm trying to consider the minimum, so ipa and python-hardware | 13:46 |
| rpittau | arne_wiebalck: yeah, you can safely remove perl | 13:46 |
| rpittau | I don't see issues with the packages you mentioned | 13:46 |
| arne_wiebalck | rpittau: these packages are among the 10 or so largest | 13:46 |
| rpittau | I'm more interested to understand what is bringing them in in the first place :) | 13:47 |
| arne_wiebalck | 20 largest, 10MB and above | 13:47 |
| rpittau | starting to have a look at centos-9 these days and it looks like we'll have to increase memory :/ | 13:47 |
| arne_wiebalck | maybe good to check on a plain cs8 machine if you have one handy | 13:47 |
| rpittau | yeah | 13:48 |
| arne_wiebalck | if they are in a vanilla one, we may want to add more packages to the removal in the IPA dib element | 13:50 |
| dtantsur | I suspect some oslo stuff may still pull in babel | 13:58 |
| dtantsur | perl may be pulled in by git or anything else like that | 13:59 |
| *** sshnaidm_ is now known as sshnaidm | 14:03 | |
| arne_wiebalck | bye everyone, see you tomorrow o/ | 16:01 |
| rpittau | good night! o/ | 16:02 |
| opendevreview | Merged openstack/ironic bugfix/18.1: Add iDRAC configuration mold docs https://review.opendev.org/c/openstack/ironic/+/812209 | 16:16 |
| opendevreview | Ruby Loo proposed openstack/ironic master: anaconda interface, clean/rescue use agent ramdisk https://review.opendev.org/c/openstack/ironic/+/815528 | 16:55 |
| opendevreview | Verification of a change to openstack/ironic master failed: There is no aim, we do deploy/manage baremetal. https://review.opendev.org/c/openstack/ironic/+/814239 | 17:02 |
| opendevreview | Merged openstack/ironic-python-agent master: Respect global parameters when downloading a configdrive https://review.opendev.org/c/openstack/ironic-python-agent/+/814771 | 17:41 |
| opendevreview | Leo McGann proposed openstack/ironic master: Adds Attestation Interface to Node https://review.opendev.org/c/openstack/ironic/+/755836 | 17:59 |
| opendevreview | Leo McGann proposed openstack/ironic master: Keylime attestation interface https://review.opendev.org/c/openstack/ironic/+/806751 | 17:59 |
| opendevreview | Merged openstack/ironic-python-agent master: Fix error messages in burnin code https://review.opendev.org/c/openstack/ironic-python-agent/+/815455 | 19:50 |
| opendevreview | Piotr Parczewski proposed openstack/ironic master: Docs: Remove broken anchor URL https://review.opendev.org/c/openstack/ironic/+/815555 | 20:29 |
| *** dviroel|rover is now known as dviroel|rover|afk | 20:42 | |
| opendevreview | Leo McGann proposed openstack/python-ironicclient master: Adds Attestation Interface to Node and Driver https://review.opendev.org/c/openstack/python-ironicclient/+/755837 | 20:44 |
| opendevreview | Verification of a change to openstack/metalsmith master failed: Update pep8 test requirements https://review.opendev.org/c/openstack/metalsmith/+/814543 | 21:41 |
| cvstealth | Hi folks, I had a question around the physical network that I couldn't easily determine from the docs. Is the cleaning/provisioning network configured as untagged traffic and the end user/composable networks setup as tagged traffic? Wasn't sure if it should be in this configuration or another interface was needed for the composable network traffic. | 22:28 |
| opendevreview | Arun S A G proposed openstack/ironic master: Fix various issues in the anaconda deploy interface https://review.opendev.org/c/openstack/ironic/+/814087 | 23:49 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!