| opendevreview | Jakub Jelinek proposed openstack/ironic-python-agent master: Log efibootmgr output as part of the system logs https://review.opendev.org/c/openstack/ironic-python-agent/+/954563 | 06:59 |
|---|---|---|
| rpittau | good morning ironic! o/ | 07:26 |
| queensly[m] | Good morning | 07:36 |
| FreemanBoss[m] | Good morning | 07:37 |
| opendevreview | Merged openstack/ironic-inspector master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic-inspector/+/954403 | 08:13 |
| abongale | gm ironic! | 08:36 |
| opendevreview | Jakub Jelinek proposed openstack/ironic-python-agent master: Log efibootmgr output as part of the system logs https://review.opendev.org/c/openstack/ironic-python-agent/+/954563 | 09:02 |
| opendevreview | Verification of a change to openstack/ironic-python-agent master failed: Fix software RAID creation on different physical devices https://review.opendev.org/c/openstack/ironic-python-agent/+/953122 | 09:31 |
| opendevreview | Merged openstack/ironic master: update Jinja2 to address CVE-2024-2383 https://review.opendev.org/c/openstack/ironic/+/953902 | 09:33 |
| Sandzwerg[m] | <TheJulia> "Sandzwerg: Yeah, they can be..." <- I installed the version ~1 week ago directly via pip so that should be the most recent release | 11:16 |
| Sandzwerg[m] | <TheJulia> "but still worthwhile double..." <- I'll see if I can get someone to look into that for me, in doubt I have to live with that till we upgrade ironic next and can switch to https boot (which apparently performs much faster and can use the 10G or faster NICs of the node itself instead of the 1G BMC) | 11:17 |
| TheJulia | hmmm | 12:58 |
| TheJulia | Sandzwerg[m]: this would do it: https://review.opendev.org/c/openstack/ironic/+/881576/1/ironic/conf/redfish.py it was Ironic :) | 13:05 |
| TheJulia | good morning abongale | 13:05 |
| TheJulia | cardoe: ahh, that is totally a behavior we could partially detect/fix, at leas the unplugged ports. :\ | 13:07 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: WIP: Deferred Tasks https://review.opendev.org/c/openstack/ironic-specs/+/954612 | 13:35 |
| TheJulia | dtantsur: cid: ^ | 13:35 |
| TheJulia | adamcarthur5: sorry! I've been crazy busy | 13:35 |
| cid | \o/ | 13:43 |
| Sandzwerg[m] | <TheJulia> "Sandzwerg: this would do it..." <- Ah nice. Will take a while till we'll see that, just upgraded to 2023.1 but hopefully I have time for the next upgrade in the second half of the year. Thansk! | 13:45 |
| adamcarthur5 | TheJulia of course, no worries :) | 13:53 |
| opendevreview | Merged openstack/ironic-python-agent-builder master: Allow dib build to remove firmware in a structure. https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/954524 | 14:02 |
| TheJulia | NobodyCam: https://review.opendev.org/c/openstack/ironic-specs/+/952533 may be of interest to you | 14:23 |
| iurygregory | cid, hey o/ by any chance you have the link of the runbook on youtube handy? | 14:23 |
| TheJulia | adamcarthur5: so, I guess what I'm missing is what does the scatter/gather result look like in terms of telling a conductor something needs to happen and how the conductor picks up the work. I guess the confusion is more like 1) issue request 2) ?????? 3) conductors start doing $things. 4) ????? (do conductor elves appear here?) 5) profit! I mean, shutdown! | 14:26 |
| TheJulia | warning: I've been up since 4-ish AM, so I'm a bit awake. | 14:27 |
| TheJulia | and slightly manic. | 14:27 |
| TheJulia | adamcarthur5: under the standard rpc model, work would be addressed to a conductor, and it would launch a task, but then how do other conductors know what to do. How do they intertwine/take action | 14:28 |
| TheJulia | and how do we get consolidated status updates back. I've floated an idea of maybe each conductor injects and manages it's own record in that table | 14:29 |
| TheJulia | also, so your distributing updates/states across the fleet as applicable. The other aspect is if you have an island which is isolated to a conductor or conductors, for example, the BMC switch is lost on one fleet, then if you see a single conductor's actions are not moving forward or failing, then its at least able to be identified | 14:30 |
| TheJulia | Happy to chat about it, that being said I also have ~4 hours of meetings later today. | 14:33 |
| TheJulia | (so might be best to discuss today) | 14:33 |
| TheJulia | or not | 14:33 |
| TheJulia | :) | 14:33 |
| cid | iurygregory, let me quickly fetch that. | 14:35 |
| iurygregory | cid, tks! | 14:35 |
| cid | https://youtu.be/00PJS4SXFYQ | 14:36 |
| cid | Hope that's ^^ what you meant :D, iurygregory. | 14:37 |
| iurygregory | cid, yes it was! tks o/ | 14:38 |
| adamcarthur5 | TheJulia I'm currently deep in PhD work, but would definitely like to meet | 14:43 |
| TheJulia | adamcarthur5: when would be good? hopefully my points are making sense | 14:44 |
| opendevreview | Merged openstack/ironic master: feat: add verify ca conf support for drivers https://review.opendev.org/c/openstack/ironic/+/947544 | 14:52 |
| kubajj | TheJulia: btw, I did some digging into why we have so many efibootmgr entries and I think we suffer from https://bugs.launchpad.net/ironic/+bug/2072336 (at least on the couple of nodes that I tested it on). I might make my intern have a look into fixing the bug if nobody picks it up by then | 14:53 |
| TheJulia | oh wow | 14:54 |
| TheJulia | yeah | 14:54 |
| kubajj | TheJulia: would you say a sensible solution to this problem would be to check if at least one entry was matched (there should be PXE or something that the IPA booted through) and if not, try to decode them with utf-8? | 14:54 |
| TheJulia | kubajj: yeah, I think so | 14:55 |
| TheJulia | We've sort of had this back and forth on the reality of records and encoding, so that seems like the cleanest path, really | 14:55 |
| kubajj | ok, I will have the intern implement it (I know I could do it quicker, but would be nice for him to get familiar with ironic :D ) | 14:56 |
| TheJulia | ++ | 14:56 |
| dtantsur | TheJulia: nice! a whole bunch of comments posted | 15:43 |
| TheJulia | I figured, its rough, but seems "do-able" | 15:43 |
| dtantsur | (I haven't read the email exchange, so I hope I don't miss anything critical) | 15:43 |
| TheJulia | I think it kind of sums up the high level | 15:44 |
| TheJulia | but, by all means do try to revisit when you get a chance | 15:44 |
| TheJulia | There is a question in my last email which I think it help to level set further | 15:44 |
| dtantsur | Yeah, I'll check it now, just have been doing something vital for higher-ups | 15:45 |
| TheJulia | no worries | 15:45 |
| TheJulia | adamcarthur5: you got a review from Arne, well done! | 15:45 |
| opendevreview | Merged openstack/ironic master: ci: stabilize ironic-standalone-redfish https://review.opendev.org/c/openstack/ironic/+/954303 | 15:54 |
| adamcarthur5 | Hey TheJulia, I can meet for 30mins-an hour in 20 minutes if that works for you? | 16:10 |
| adamcarthur5 | Finished earlier than expected | 16:10 |
| TheJulia | That should work | 16:11 |
| adamcarthur5 | Okay sweet, see you soon :) | 16:12 |
| TheJulia | dtantsur: I think your thining a bit more complexity on deferred tasks than I was thinking, but you've got a point on the contet, I was more thinking the split points are largely around periodics like "we did a firmware update" "oh, we need to wait for a while" and the code pattern on the next invocation could be aware the generalized context. I twould require more specific and intentional splitting of flows which would | 17:11 |
| TheJulia | maintain that continuity. I guess either way. Anyway, comments posted. Thanks! | 17:11 |
| dtantsur | responding already, thank you | 17:13 |
| dtantsur | okay, this should clear up some confusion I created :) | 17:18 |
| dtantsur | will continue tomorrow, see ya | 17:18 |
| TheJulia | cool cool, awesome | 17:19 |
| TheJulia | I just workflowed your local rpc patch | 17:20 |
| dtantsur | \o/ | 17:20 |
| TheJulia | Worst comes to worst, we can revert it or revise it more as we go | 17:20 |
| dtantsur | It's not the end of it, just the first step forward | 17:20 |
| TheJulia | Once it has merged, I'll rebase the patch sequence and see if that cleans things up | 17:20 |
| TheJulia | oh, okay | 17:20 |
| dtantsur | yeah, unfortunately | 17:20 |
| TheJulia | I'm likely not going to be able to work on anything load testing/measurement related until next week | 17:20 |
| dtantsur | it will work for cases where RPC is present in the all-in-one executable | 17:21 |
| * TheJulia lives at a service center presently | 17:21 | |
| dtantsur | it won't work for rpc_transport=none just yet | 17:21 |
| dtantsur | ouch | 17:21 |
| * TheJulia hopes to go home this weekend | 17:21 | |
| dtantsur | I hope they have treats at least | 17:21 |
| TheJulia | eh... not really. :( | 17:21 |
| TheJulia | and I have 4 more meetings to go today | 17:21 |
| dtantsur | damn | 17:21 |
| dtantsur | my condolences | 17:21 |
| dtantsur | I won't distract you until tomorrow then :) | 17:21 |
| TheJulia | the way I was reading it, it would look like it *might* kind of work, but I didn't put tooooo much thought into it | 17:22 |
| TheJulia | thanks, have a great evening | 17:22 |
| opendevreview | Merged openstack/ironic master: Stop short-cutting to local manager with all-in-one processes https://review.opendev.org/c/openstack/ironic/+/954363 | 18:13 |
| * TheJulia goes to push a rebase button a few times | 18:28 | |
| TheJulia | :) | 18:28 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!