| opendevreview | Verification of a change to openstack/ironic-tempest-plugin master failed: Disable use of the SSH agent for the rescue test https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/872567 | 00:15 |
|---|---|---|
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: WIP: Framework for DPU management/orchustration https://review.opendev.org/c/openstack/ironic-specs/+/874189 | 03:17 |
| TheJulia | jrosser: idea ^^ thoughts? I could see that plus a distinct "bluefield" deploy interface being a path to iterate and deploy to multiple DPU cards. | 03:18 |
| TheJulia | NobodyCam: also ^^^ | 03:19 |
| janders | TheJulia interesting BF discussion. Will try invite Moshe to the PTG sessions :D | 04:01 |
| rpittau | good morning ironic! o/ | 08:45 |
| kubajj | Good morning o/ | 08:49 |
| mnasiadka | good morning, seems bifrost broke on Ubuntu Jammy (https://742be02cc882e0872e77-1e03f6a0971577a3603e6088d0a330da.ssl.cf1.rackcdn.com/874055/1/gate/kolla-build-ubuntu/66fbe2c/kolla/build/000_FAILED_bifrost-base.log) - anybody has an idea what is going on? | 09:53 |
| mnasiadka | hmm, they renamed shimx64.efi.signed to shimx64.efi.signed.latest? | 09:54 |
| opendevreview | Verification of a change to openstack/ironic master failed: Indicate maintenance mode https://review.opendev.org/c/openstack/ironic/+/874085 | 10:34 |
| iurygregory | morning ironic, happy friday | 11:16 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: Disable use of the SSH agent for the rescue test https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/872567 | 11:18 |
| dtantsur | TheJulia: I've been wondering at which point we end up with children-parent relationship between nodes :D | 12:06 |
| dtantsur | TheJulia: but I've been also worried for a long time that we're overbloating the node object | 12:08 |
| dtantsur | (this is one of the reasons why my Deployment API proposal came to life) | 12:08 |
| dtantsur | TheJulia: I guess the question will be: how many of the normal server notions apply to subnodes? Provision and power states? Consoles? Boot interface? | 12:09 |
| dtantsur | I wish we had a Machine object that would point to a more lightweight Node object and to a bunch of children, also Nodes | 12:09 |
| dtantsur | so a Machine is something that has a full lifecycle, while a Node is something that we can manage | 12:09 |
| dtantsur | yes, I'm probably inventing API v2. I'll show myself the door. | 12:10 |
| iurygregory | children-parent reminds me of the time when people wanted this concept in projects to handle quotas | 12:10 |
| iurygregory | oh wow, ironic API v2 =) | 12:10 |
| dtantsur | API v2 has been a running semi-joke for half a decade already :) | 12:11 |
| dtantsur | I should probably take a list of things I would change in Ironic and write a blog post about it (purely for entertainment) | 12:11 |
| kubajj | dtantsur: I had some problems with my testing environment, but managed to solve it now and https://review.opendev.org/c/openstack/ironic/+/874074 is ready for re-review | 13:19 |
| dtantsur | Thanks! | 13:23 |
| opendevreview | Merged openstack/ironic master: Indicate maintenance mode https://review.opendev.org/c/openstack/ironic/+/874085 | 13:53 |
| TheJulia | dtantsur: so, I think whole server power on/off would apply to the child nodes. The cards have their own consoles, their own power states, and realistically boot interfaces. Think independent computer on some level which could need entirely independent lifecycle management | 14:36 |
| TheJulia | from an overload standpoint, I do sort of agree with you in terms of when a detailed list is queried. I've proposed we do something with chassis as well... but the gravity is all around the nodes table, all we can *really* do is make sure we do things quick/effectively/efficiently at the lower layers (which we've gotten better at as we've had to do some of that refactoring as time has passed. | 14:41 |
| TheJulia | mnasiadka: I'm not really surprised. They have changed the filename before :( | 14:42 |
| kubajj | TheJulia: if you had a minute, could I ask for a review of https://review.opendev.org/c/openstack/ironic/+/874074 ? | 14:50 |
| kubajj | TheJulia: Thanks | 14:55 |
| * JayF emerges from quarantine | 15:03 | |
| JayF | What year is it? | 15:03 |
| JayF | :D | 15:03 |
| TheJulia | 2023, you doing okay? | 15:03 |
| JayF | I woke up this morning feeling really good | 15:04 |
| JayF | so time to see how long it can hold lol | 15:04 |
| TheJulia | don't push yourself too much, covid is no joke | 15:05 |
| JayF | Basically, Weds I was on the couch all day, 102F fever, got some new covid pill (not paxlovid; some other one)... by midday thursday I was feeling human again | 15:08 |
| JayF | and literally I feel better this morning than I've felt in weeks | 15:09 |
| JayF | (so far) | 15:09 |
| JayF | like the fog has risen on my brain | 15:09 |
| TheJulia | that is good | 15:09 |
| TheJulia | anti-virals for the win | 15:09 |
| *** dking is now known as Guest5153 | 15:31 | |
| *** Guest5153 is now known as dking | 15:31 | |
| *** tzumainn_ is now known as tzumainn | 16:01 | |
| JayF | rpittau: iurygregory: dtantsur: Can one of you own making sure we get library releases out before the freeze? | 16:22 |
| JayF | (please) | 16:22 |
| iurygregory | JayF, sure | 16:23 |
| JayF | thank you \o/ | 16:23 |
| dtantsur | I'll leave it up to Iury, I'm doing too many things on the metal3 side nowadays.. | 16:38 |
| rpittau | iurygregory: let me know if you need help or want to split the work | 16:41 |
| iurygregory | sure =) | 16:41 |
| rpittau | on Monday though, flying away now! :) | 16:41 |
| rpittau | bye everyone, have a great weekend! o/ | 16:41 |
| iurygregory | yeah | 16:42 |
| iurygregory | bye rpittau o/ | 16:42 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix Inventory DB https://review.opendev.org/c/openstack/ironic/+/874074 | 16:49 |
| JayF | is the "why" for https://review.opendev.org/c/openstack/ironic/+/864781/ performance, sqla2 compat, or else()? | 16:49 |
| JayF | hjensas: ^ | 16:49 |
| JayF | TheJulia: you might know too if you're around ^ you have a +2 on it | 16:51 |
| TheJulia | speaking of shards, did a python-ironicclient change get proposed yet? | 16:52 |
| mnasiadka | TheJulia: fantastic, at least they have a track record of being successful at this :) | 16:53 |
| TheJulia | JayF: performance, otherwise the query pattern is get list, later on reconcile it for each entry | 16:53 |
| JayF | Is there a benchmark in there anywhere I missed out on? | 16:53 |
| TheJulia | move the logic/data generation closer to the DB | 16:54 |
| JayF | TheJulia: no, it's on my list | 16:54 |
| JayF | Or was one run, you know it's better, and I can trust it :) | 16:54 |
| JayF | TheJulia: honestly today is going to be shard release notes + cycle highlights + trying to write that PR | 16:54 |
| TheJulia | JayF: harald added ports/portgroups to the benchmark script, so it should be able to compare that change and any change of master branch at this point | 16:54 |
| JayF | bluntly, I don't want to do that work in my already-compressed week. I was hoping someone already had and could attest it was faster | 16:55 |
| TheJulia | He did post results at some point... | 16:55 |
| JayF | oh, good | 16:55 |
| JayF | you are confident it has been tested and found faster? | 16:55 |
| TheJulia | but that was at least 3 months ago | 16:55 |
| JayF | that's the real question I'm trying to answer | 16:55 |
| JayF | "was a benchmark run on this? was it better?" | 16:55 |
| JayF | not trying to create a paper trail lol | 16:55 |
| TheJulia | yes, the entire chain makes each problem area faster as it weeds out our inefficient patterns | 16:55 |
| TheJulia | yes, I'm fairly certain it makes list ops faster :) | 16:56 |
| JayF | TheJulia: hash ring rest Pr; did you want Iury to review it b/c I was out, or b/c he should review before it lands | 17:00 |
| * JayF wants to land it | 17:00 | |
| TheJulia | just another set of eyes | 17:01 |
| TheJulia | also since you were out | 17:01 |
| JayF | ack | 17:01 |
| JayF | that's a great change | 17:01 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix Inventory DB https://review.opendev.org/c/openstack/ironic/+/874074 | 17:25 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Add release note for node sharding https://review.opendev.org/c/openstack/ironic/+/873652 | 17:38 |
| JayF | TheJulia: ^ I must have had early-covid-brain on that first patch, yikes. | 17:38 |
| JayF | sorry for pushing something so ugly | 17:38 |
| TheJulia | c'est la vie | 17:42 |
| opendevreview | Merged openstack/sushy-tools master: Save previous hd boot device when setting Pxe https://review.opendev.org/c/openstack/sushy-tools/+/869379 | 17:45 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: WIP: Framework for DPU management/orchustration https://review.opendev.org/c/openstack/ironic-specs/+/874189 | 18:16 |
| TheJulia | dtantsur: idea refined a little further and tried to answer your overall questions in the text. Also thanks for the alternative :) | 18:16 |
| iurygregory | JayF, regarding library releases we already have for sushy, metalsmith, ironic-lib and for clients the only thing missing would be cut stable/antelope for them | 18:55 |
| iurygregory | we can probably go ahead if we want to create the branch already for them | 18:56 |
| JayF | iurygregory: https://lists.openstack.org/pipermail/openstack-discuss/2023-February/032234.html I mainly want to be sure we've action'd this | 18:56 |
| iurygregory | looking | 18:56 |
| JayF | we have 6 things on that list | 18:56 |
| iurygregory | some of them we do together with ironic when we are creating the stable brnach | 18:57 |
| iurygregory | going to check the open patches in this repos | 18:57 |
| JayF | can you just own making releases team happy w/r/t that email? | 19:02 |
| iurygregory | yup will do | 19:02 |
| opendevreview | Verification of a change to openstack/ironic master failed: Use association_proxy for port groups node_uuid https://review.opendev.org/c/openstack/ironic/+/864781 | 19:04 |
| TheJulia | vanou: Regarding your two backports against Zed, I'd like to discuss sometime next week about them because I think there is some missing context or understanding that we need to clear up to find the path forward. Your Tuesday / my Monday might work best. | 19:23 |
| opendevreview | Verification of a change to openstack/ironic master failed: Respond to rpc requests on stop until hash ring reset https://review.opendev.org/c/openstack/ironic/+/872601 | 19:46 |
| opendevreview | Merged openstack/ironic master: Fix Inventory DB https://review.opendev.org/c/openstack/ironic/+/874074 | 21:14 |
| kubajj | \o/ | 21:18 |
| * JayF working on cycle highlights | 22:24 | |
| JayF | does the node inventory support in API 1.81 do anything before the whole inspector-absorbtion work is done? | 22:24 |
| * JayF planning to not include it in lieu of other information | 22:27 | |
| JayF | I was also really hoping metal3 CI will be done, I want to call that out in cycle highlights as "CI coverage for sqlite usage" with a stick-tap to metal3 | 22:28 |
| JayF | also on the fence about if I should include the utf8{,_mb4} changes | 22:30 |
| JayF | https://review.opendev.org/c/openstack/releases/+/874338 if you wanna just put comments there | 22:32 |
| kubajj | JayF: It's not doing anything so far | 23:43 |
| JayF | ack, ty | 23:44 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!