| mmalchuk | cores, please review https://review.opendev.org/c/openstack/bifrost/+/896168 | 00:08 |
|---|---|---|
| opendevreview | Adam Rozman proposed openstack/ironic master: implement basic-auth support for user image download process https://review.opendev.org/c/openstack/ironic/+/890411 | 07:39 |
| TheJulia | good morning | 10:59 |
| TheJulia | ... says the one with insomnia | 10:59 |
| dtantsur | oh, that's early. Good morning TheJulia | 11:14 |
| TheJulia | so Httpboot? thoughts on seeing if we can pull off in this next cycle? | 11:22 |
| TheJulia | for simulating in ci, I guess it could get translated out the same way we handle vmedia, since vendor to vendor it is likely to differ | 11:29 |
| dtantsur | TheJulia, assuming you're talking about UEFI httpboot, which one do you have in mind: DHCP-driven or Redfish-driven? | 11:32 |
| TheJulia | downside of being up hours before the wifey.... none of the lights are on and I can't find my glasses | 11:33 |
| TheJulia | I think redfish driven would be ideal for a driver interface. I think dhcp driven is a nice stretch goal afterwards if it turns out to be relatively easy | 11:33 |
| dtantsur | I wouldn't expect anything with "DHCP" in it to be easy :) | 11:34 |
| TheJulia | well, so *technically* we already kind of do it for PXE | 11:34 |
| TheJulia | *kind of * | 11:34 |
| TheJulia | the .... ipv6 pattern is a URL | 11:34 |
| dtantsur | I'm also more interested in the Redfish approach, which seem to be a lightweight alternative for virtual media, and is indeed mostly the same thing code-wise | 11:34 |
| TheJulia | just we load ipxe/grub instead | 11:34 |
| dtantsur | Can be quite crazy for non-neutron cases | 11:35 |
| TheJulia | eh, relatively simple | 11:36 |
| TheJulia | relatively | 11:36 |
| dtantsur | How do you distinguish iPXE and UEFI boot in a static DHCP config? | 11:36 |
| TheJulia | UEFIHTTP Boot, using the vendor-class identifier | 11:37 |
| dtantsur | aha, not bad | 11:37 |
| TheJulia | ... I kind of *feel* like on a pxe/ipxe route, maybe we just wire to the existing bootloaders since there is tons of duplicate logic there | 11:37 |
| dtantsur | This requires using a different boot device, right? | 11:37 |
| TheJulia | which over laps to | 11:37 |
| TheJulia | too | 11:37 |
| TheJulia | for uefihttp boot? | 11:38 |
| dtantsur | yep (I know for Redfish the answer is yes) | 11:38 |
| TheJulia | for our testing, yeah, we would likely attach it as a CD device like the vmedia code | 11:38 |
| TheJulia | tianocore does seem to have logic/understanding, but there doesn't appear to be a way to hint it from libvirt | 11:39 |
| dtantsur | I mean... how does it decide to send the UEFIHTTP Boot vendor class? | 11:39 |
| TheJulia | I bet what the bmcs do is they update the nvram with new config for it to load | 11:39 |
| TheJulia | for physical machines, default config, tianocore e2dk in jammy and beyond try both pxe and httpboot | 11:40 |
| TheJulia | they alternate the vendor class | 11:40 |
| TheJulia | ... it is rather annoying. | 11:41 |
| dtantsur | got it, makes sense | 11:41 |
| dtantsur | so from Ironic's perspective, it's the same network boot device? | 11:41 |
| TheJulia | hell, it just makes sense to send alternate vendor class config to dnsmasq with the pxe stuffs | 11:41 |
| TheJulia | yeah | 11:41 |
| dtantsur | okay, I guess it depends on whether someone can dedicate cycles to working on it. Our team might.. or might not depending on $priorities. | 11:43 |
| TheJulia | Happens to be the same reason I'm brining it up, trying to set my priorities for next quarter | 11:45 |
| iurygregory | good morning Ironic | 11:51 |
| TheJulia | good morning iurygregory! | 11:52 |
| * TheJulia sips the coffee and acts animated | 11:53 | |
| iurygregory | good morning TheJulia :D | 11:54 |
| * iurygregory just finished reading the conversation about UEFIHTTP Boot | 11:54 | |
| iurygregory | I just hope this friday won't be like the one from last week | 11:55 |
| TheJulia | Yeah, no. That sounded rough :( | 11:55 |
| iurygregory | yup! | 11:56 |
| TheJulia | Although I've been up since 2 AM, there is no guarentee today is not a half day at this point or a "oh, it is 9 AM, nap time!" | 11:56 |
| iurygregory | ++ | 11:57 |
| iurygregory | you should probably go back to sleep =) | 11:57 |
| TheJulia | nah! | 11:59 |
| iurygregory | one more coffee and I will go back to finish the presentation I've started on Wed \o/ | 12:01 |
| iurygregory | and hopefully I will finish today | 12:01 |
| TheJulia | I need to try and wrap up slides for my talk in 1.5 weeks sometime today or this weekend | 12:08 |
| TheJulia | 40 minutes \o/ | 12:08 |
| TheJulia | if Anyone has any good metal3 slides, I'd <3 if you'd share because that will save me some time | 12:08 |
| iurygregory | my talk is next Tuesday :D | 12:08 |
| iurygregory | depends on what you need about metal3 | 12:09 |
| TheJulia | if you've got an example CRD update slide or slides on it's structure, that would be epic | 12:10 |
| TheJulia | or even just a example slide of a populated CRD | 12:10 |
| TheJulia | I'm going to briefly talk about metal3/bifrost/ironic all together | 12:10 |
| TheJulia | it should be *fun* | 12:10 |
| iurygregory | ok, we probably have something in our downstream presentations... | 12:10 |
| * iurygregory checks | 12:10 | |
| TheJulia | thanks! | 12:13 |
| iurygregory | yw =) | 12:14 |
| opendevreview | Merged openstack/ironic-python-agent master: Extend the lookup timeout to 600 seconds https://review.opendev.org/c/openstack/ironic-python-agent/+/892686 | 12:35 |
| iurygregory | backports? :D | 12:38 |
| TheJulia | could be good ? :) | 12:49 |
| * dtantsur holds fingers crossed for the tinyipa build job | 12:50 | |
| TheJulia | the prior timeout expects things are in a great state | 12:50 |
| TheJulia | oh! that reminds me | 12:50 |
| TheJulia | did it do the needful going to Jammy | 12:50 |
| dtantsur | it looks so: https://zuul.opendev.org/t/openstack/builds?job_name=ironic-python-agent-build-image-tinyipa&skip=0 | 12:50 |
| dtantsur | hence my hopes are quite high :) | 12:51 |
| TheJulia | \o/ | 12:51 |
| TheJulia | yeah, that was bizzar | 12:51 |
| TheJulia | but I guess it was something with the chroot | 12:51 |
| TheJulia | ... those bugs are super rare too | 12:51 |
| dtantsur | Thanks for handling it! | 12:52 |
| TheJulia | no problem! | 12:55 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent stable/2023.2: Extend the lookup timeout to 600 seconds https://review.opendev.org/c/openstack/ironic-python-agent/+/896206 | 13:04 |
| dtantsur | okay, if we merge ^^, we'll finally get tinyipa for 2023.2 | 13:04 |
| TheJulia | Easy enough :) | 13:04 |
| iurygregory | but wait, don't we need https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/896146 backported to 2023.2 ? | 13:09 |
| dtantsur | fair | 13:09 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent-builder stable/2023.2: CI: Change image-build to use Ubuntu Jammy https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/896207 | 13:09 |
| dtantsur | okay, so this goes first ^^ | 13:09 |
| iurygregory | yeah | 13:09 |
| iurygregory | ++ | 13:09 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent stable/2023.2: Extend the lookup timeout to 600 seconds https://review.opendev.org/c/openstack/ironic-python-agent/+/896206 | 13:10 |
| dtantsur | added depends-on to be sure | 13:10 |
| iurygregory | perfect | 13:11 |
| iurygregory | dtantsur++ | 13:11 |
| opendevreview | Kaifeng Wang proposed openstack/python-ironicclient master: Client support port name https://review.opendev.org/c/openstack/python-ironicclient/+/896067 | 13:42 |
| TheJulia | https://redfish.dmtf.org/redfish/mockups/v1/1354#Systems--Card1-SoC Interesting! | 13:45 |
| opendevreview | OpenStack Release Bot proposed openstack/ironic stable/2023.2: Update .gitreview for stable/2023.2 https://review.opendev.org/c/openstack/ironic/+/896241 | 13:48 |
| opendevreview | OpenStack Release Bot proposed openstack/ironic stable/2023.2: Update TOX_CONSTRAINTS_FILE for stable/2023.2 https://review.opendev.org/c/openstack/ironic/+/896242 | 13:48 |
| opendevreview | OpenStack Release Bot proposed openstack/ironic master: Update master for stable/2023.2 https://review.opendev.org/c/openstack/ironic/+/896243 | 13:48 |
| opendevreview | Merged openstack/ironic-python-agent-builder stable/2023.2: CI: Change image-build to use Ubuntu Jammy https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/896207 | 13:56 |
| opendevreview | Harald Jensås proposed openstack/ironic stable/zed: redfish_address - wrap_ipv6 address https://review.opendev.org/c/openstack/ironic/+/896208 | 13:58 |
| opendevreview | Harald Jensås proposed openstack/ironic stable/yoga: redfish_address - wrap_ipv6 address https://review.opendev.org/c/openstack/ironic/+/896209 | 13:58 |
| iurygregory | waittt now smart nic is recognized as a System? | 13:59 |
| iurygregory | in the redfish world... | 13:59 |
| opendevreview | Harald Jensås proposed openstack/ironic stable/xena: redfish_address - wrap_ipv6 address https://review.opendev.org/c/openstack/ironic/+/896210 | 13:59 |
| TheJulia | it can be, if nc-si is supported on both sides and the bmc is aware | 14:00 |
| TheJulia | but not *always* the case | 14:00 |
| TheJulia | since some are doing their own BMCs too | 14:00 |
| iurygregory | yeah | 14:01 |
| iurygregory | but sounds interesting indeed! | 14:01 |
| TheJulia | somewhat yeah | 14:09 |
| dtantsur | and my pet peeve: which system does the virtualmedia belong to? :) | 14:20 |
| ashinclouds[m] | Exactly! | 14:28 |
| opendevreview | Merged openstack/ironic-python-agent stable/2023.2: Extend the lookup timeout to 600 seconds https://review.opendev.org/c/openstack/ironic-python-agent/+/896206 | 14:39 |
| dtantsur | woohoo, we have tinyipa 2023.2 built | 15:07 |
| clarkb | dtantsur: does that mean the newer build host just worked? | 15:48 |
| dtantsur | clarkb, you mean jammy? yep. | 15:48 |
| clarkb | yup. Fun my guess then is something in the build causes the focal kernel to panic | 15:48 |
| clarkb | because the host was going away and having all sorts of problems | 15:48 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: Add inspection hooks https://review.opendev.org/c/openstack/ironic/+/890817 | 21:50 |
| shermanm | hey, I was curious if it was possible to set ironic's "root device hints" via nova? I've got some users who'd like to pick which disk their instance boots from | 22:35 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!