| opendevreview | yatin proposed openstack/ironic master: [DNM][jammy] Test dnsmasq with fix https://review.opendev.org/c/openstack/ironic/+/897277 | 06:29 |
|---|---|---|
| rpittau | good morning ironic! o/ | 06:39 |
| opendevreview | Riccardo Pittau proposed openstack/ironic bugfix/22.1: Fix the HTTP code for reaching max_concurrent_deploy: 503 instead of 500 https://review.opendev.org/c/openstack/ironic/+/897180 | 07:11 |
| opendevreview | Riccardo Pittau proposed openstack/ironic bugfix/22.0: Fix the HTTP code for reaching max_concurrent_deploy: 503 instead of 500 https://review.opendev.org/c/openstack/ironic/+/897281 | 07:11 |
| opendevreview | Verification of a change to openstack/ironic-python-agent master failed: Conditional creation of RAIDed ESP for UEFI Software RAID https://review.opendev.org/c/openstack/ironic-python-agent/+/891609 | 07:19 |
| opendevreview | Verification of a change to openstack/ironic-python-agent master failed: Conditional creation of RAIDed ESP for UEFI Software RAID https://review.opendev.org/c/openstack/ironic-python-agent/+/891609 | 11:07 |
| iurygregory | good morning Ironic | 11:20 |
| opendevreview | Merged openstack/ironic bugfix/22.1: Fix the HTTP code for reaching max_concurrent_deploy: 503 instead of 500 https://review.opendev.org/c/openstack/ironic/+/897180 | 12:04 |
| dtantsur | JayF, our scaling struggles were indeed because the tester forgot to turn some hardcore network problem emulation off | 12:09 |
| dtantsur | But I do want to pursue optimizing the heartbeat case further | 12:09 |
| iurygregory | dtantsur, really?! | 12:09 |
| iurygregory | =O | 12:09 |
| iurygregory | now I'm wondering if would affect the other bug I was working on | 12:10 |
| * iurygregory checks the internal slack | 12:10 | |
| dtantsur | dunno. ask. | 12:10 |
| dtantsur | but it seems that 500/600 nodes at a time is not a huge deal on a well-behaved network :D | 12:10 |
| dtantsur | iurygregory, "98.2% of the clusters installed successfully" | 12:12 |
| iurygregory | yeah | 12:15 |
| iurygregory | I'm wondering if in the one for the proxy error in inspector they were also using the mixed setup or not | 12:15 |
| opendevreview | Merged openstack/ironic bugfix/22.0: Fix the HTTP code for reaching max_concurrent_deploy: 503 instead of 500 https://review.opendev.org/c/openstack/ironic/+/897281 | 12:20 |
| dtantsur | iurygregory, JayF, there are possibilities for improvements nonetheless, see my notes here https://bugs.launchpad.net/ironic/+bug/2038438 | 13:35 |
| iurygregory | looking | 13:35 |
| dtantsur | not sure it helps with the occasional inspector 502's though | 13:36 |
| iurygregory | got it | 13:38 |
| iurygregory | it does sounds interesting the proposal you have in the bug | 13:38 |
| TheJulia | Good morning | 13:41 |
| TheJulia | dtantsur: over how much time was that deploy for it to be 98.2 percent successful? | 13:42 |
| dtantsur | TheJulia, slightly more than 8 hours | 13:45 |
| dtantsur | In the failed case, it started failing after a couple of hours already | 13:45 |
| TheJulia | how big was the image? | 13:49 |
| TheJulia | well, bytes that get transferred to complete deploy over network | 13:49 |
| TheJulia | still, at 8 hours, not bad | 13:50 |
| dtantsur | TheJulia, typical CoreOS ISO, so 700-800M | 13:54 |
| dtantsur | also, worth noting that their "definition of done" is "OpenShift node installed", not just "OS deployed" | 14:05 |
| TheJulia | yeah, I figured | 14:05 |
| TheJulia | wheeee bobcat is released and blog posts are showing up on openstack.org/blog | 14:21 |
| dtantsur | \o/ | 14:21 |
| dtantsur | now on to the next middle-sized cat! | 14:21 |
| TheJulia | Now can anyone explain the "digital badge trend" ? | 14:25 |
| dtantsur | mmm? | 14:26 |
| iurygregory | I'm also trying to understand the digital badge | 14:27 |
| TheJulia | so there is this trend to seeminly gamify... everything by creating "digital badges" | 14:28 |
| dtantsur | ah, not very new, I would say | 14:29 |
| TheJulia | Just seems... silly... to me. | 14:29 |
| dtantsur | Fedora has been annoying me with that for a long time :D | 14:29 |
| iurygregory | they should show the badge in the profiles :D | 14:34 |
| * dtantsur gives TheJulia badge "Database conqueror" | 14:50 | |
| TheJulia | ... That is actually not a bad idea, "awards for achieving/conquering the needful" | 14:52 |
| clarkb | TheJulia: https://arxiv.org/abs/2006.02371 tldr appears to be that it works. I'm still not convinced it leads to the types of contribution that lead to healthy long lived communities though | 15:41 |
| TheJulia | Interesting | 15:42 |
| TheJulia | I think I'll join you in the not convinced boat | 15:42 |
| * TheJulia sures the population of said boat is larger than it seems, but it may in fact just be a Tardis | 15:43 | |
| TheJulia | err, is sure | 15:43 |
| * TheJulia goes back to dry running slide deck | 15:43 | |
| dtantsur | Yeah, I'm not sure how it works long-term. I remember having a lot of fun in Duolingo until I can really sick of its gamification aspects. | 15:50 |
| rpittau | see you tomorrow! o/ | 15:56 |
| TheJulia | Anyone want to update the latest release version on the ironic website? :) | 16:01 |
| iurygregory | I can do after lunch | 16:07 |
| JayF | clarkb: Honestly, in some ways I think you have to get fresh eyes on an OSS project to cultivate those contributors. Maybe out of 100 drive-by contributions to get a badge, one or two people stick around and help out with more bugs around the edges and maybe one outta every thousand goes on to work in openstack? | 16:09 |
| JayF | that'd be better than we're doing now tbh | 16:09 |
| clarkb | ya I'm not against it. I just think it isn't sufficient | 16:10 |
| JayF | I see both sides of that network effect from my work on armada; while we had a lot of drive-by contributions driven by gamification, those contributors went on to be promoters of the project, raising the profile of it and making it a more active topic on socials | 16:10 |
| clarkb | There are also downsides like the flood of bad PRs during hacktober or whatever the digital ocean event is | 16:10 |
| clarkb | have to find a balance between too many bad side effects and good ones | 16:10 |
| JayF | While that's a good observation; I think the default endpoint for too little new blood is staleness and eventual spin down; if you have too much attention and are overwhelmed, at least the longer term outcomes are more managable in some ways. | 16:11 |
| JayF | People being interested in pushing bad PRs are still people being interested :D | 16:12 |
| JayF | and I can work to make a bad change into a good one if that interested person wants to learn | 16:12 |
| clarkb | eh I'm not sure I agree. It is trivial to write trivial typo fix PRs against a random assortment of participating repos. You write a tool to meet the game goal and don't care anything about the projects you are pushing code against | 16:12 |
| clarkb | Basically PRs against projects of that sort don't indicate any interest in the project | 16:13 |
| clarkb | they may but it isn't implied | 16:13 |
| JayF | Interesting because that case really does take the interest part away | 16:13 |
| JayF | that's not something I've run up against much personally, but I know the original hacktoberfest had a lot of that crap (and the program was since revised to disincentivize it) | 16:14 |
| clarkb | this is the first year they are running with the revisions. We'll see in a few weeks if they have helped | 16:14 |
| TheJulia | iurygregory: thanks | 16:15 |
| clarkb | JayF: another fun thing that happens on Github is people make those types of PRs to build rep so the accounts can be used for star selling :) | 16:16 |
| clarkb | again they don't care anything about your project. They just need projects who can be used as a means to the end of abusing social media for money | 16:17 |
| dtantsur | People are terrible. Fortunately, we'll all be replaced by AI soon \o/ | 16:17 |
| * dtantsur ducks | 16:17 | |
| TheJulia | dtantsur: AI work product cannot be copyrighted in the US | 16:17 |
| dtantsur | Ooops | 16:18 |
| JayF | clarkb: I think there's a philosophical difference in my view; I worry we have a huge false positive rate in keeping out abusive changes -- our bar is so high that it's difficult for even an informed human who is unfamiliar to push a change... I'd rather have to fight off github spam, and have 1% of those PRs be a real person who wants to learn about clouds than to send that | 16:19 |
| JayF | person away | 16:19 |
| clarkb | JayF: if you haven't yet I suggest you subscribe to ironic project events in github. Its not that I don't want to help I do. I'm subscribed and do my best to bring up PRs that should probably be fostered with the projects. Just did that with swift the other day even. unfortunately I'm not sure the swift team ahas done anything yet | 16:21 |
| clarkb | but ya if you feel that way you should subscribe to the project events in github. its a useful too | 16:21 |
| clarkb | *tool | 16:21 |
| JayF | heh, I've done that before as a poll; but unfortunately my github notifications are ... the most horrible place in the universe | 16:21 |
| JayF | that may be a valid reason to go clean up like, 15 years of old subscriptions on github or whatever it is lol | 16:22 |
| clarkb | just filter out the ironic ones then | 16:22 |
| clarkb | there is probably an event once a year | 16:22 |
| JayF | if I did it I'd probably do it for openstack/* not just Ironic | 16:22 |
| clarkb | maybe a little more than that but its easy to filter out of the mess | 16:22 |
| * JayF makes a note | 16:22 | |
| clarkb | https://github.com/openstack/liberasurecode/pull/5 | 16:23 |
| clarkb | is the one I brought up in the swift irc channel the other day | 16:23 |
| JayF | clarkb: ah, I remember what it is now; I saw a github PR link in a downstream pbr patch in gentoo; I upstreamed that ... I do ned to get the notification firehose going | 16:23 |
| clarkb | I'm in all the places and contribute to all the random things. I think where I get frustrated is where people haev decided there is one true tool and act like anyone doing anything else is the problem. Meanwhile I'm on mailman mailing lists, in google groups, discord servers, matrix, irc, github, gitlab, more than one gerrit, and so on. Thats just reality | 16:26 |
| clarkb | last week I was even writing java :) | 16:27 |
| JayF | I agree 100%; but I have also interacted with enough new-to-programming folks who really are accellarated by the syntax sugar of github prs and such | 16:28 |
| JayF | and it makes me wish our technology was more welcoming -- I think we are open but not always welcoming because of the number of barriers to jump over | 16:28 |
| JayF | I don't know how to solve this problem but that's the corner of it, in context of OpenStack, that I get frustrated with. Seeing operators or young devs just basically punt on pushing something upstream because they don't have an hour+ to dedicate to learning a new way | 16:29 |
| JayF | and it's reasonable of us to require it; and it's reasonable of them to eschew it... and I don't know where the better way is | 16:29 |
| clarkb | it goes both ways though and I Think that is what people fail to acknowledge https://mitchellh.com/writing/github-changesets | 16:30 |
| clarkb | even mitchellh is basically arguing we should all be using something better than github which surprised me | 16:30 |
| JayF | The folks I'm talking about, frankly, don't care about those details. Github or gerrit or "is it open to use a SCM that might scan and AIify al lthe tings" are not what those folks want; they are trying to take the shortest path to get a thing done in a world where there is not enough time and too many things to do | 16:31 |
| JayF | and I don't blame them for that; many times the choices to use or even to be taught github workflows were made for them by schools or early employers | 16:32 |
| JayF | I want to find a way to bridge both types of folks together, and I don't know how, yet | 16:33 |
| JayF | but this is the community problem that proverbially 'keeps me up at night' | 16:33 |
| clarkb | ultimately I think it comes down to people needing to accept that different groups have different priorities and meeting them halfway. I don't personally see "this is a different process" as "unwlecoming" and I think that attitude is part of the problem. The reality of this sort of job is that you'll intercat with lots of different systems, tools and practices due to | 16:35 |
| clarkb | differing priorities. Accepting that and meeting people where they are rather than asking them to reinvent their world is important | 16:35 |
| clarkb | doesn't make me happy I'd love to avoid all the proprietary tooling out there butI'd never get anything done if I did | 16:36 |
| clarkb | you're always going to be more comfortable and efficient working in the system you are familiar with. We need to detach that from "can I get work done anyway" | 16:37 |
| JayF | to be clear; I didn't mean "we are unwelcoming by doing X" I mean more "being different can be percieved as unwelcoming" :) | 16:39 |
| iurygregory | TheJulia, https://github.com/OpenStackweb/ironic-website/pull/60 | 16:44 |
| JayF | looking | 16:44 |
| JayF | I think I can land things in that repo; lets see once CI passes | 16:45 |
| JayF | iurygregory: your timing is super hilarious given the chat Clark and I were just having | 16:45 |
| JayF | even OpenStack-proper you gotta know >1 change-proposing-webapp to particiapte in all corners | 16:45 |
| iurygregory | LOL | 16:46 |
| TheJulia | well, the website is less openstack proper as much as "the provider toolchain for the website dictates github" | 16:46 |
| clarkb | and we (opendev) provide a process that doesn't require that which is used by other projects just fine | 16:47 |
| clarkb | you've chosen a toolchain for your website that dictates github | 16:47 |
| clarkb | not really openstack's fault | 16:47 |
| JayF | you, aiui, in this case is oif to be clear | 16:48 |
| clarkb | JayF: that part is unclear to me. Zuul doesn't do that for example | 16:48 |
| clarkb | I think the others do and the foundation may offer additionjal support if you go down that path | 16:48 |
| clarkb | but it isn't a requirement as far as I can tell | 16:48 |
| JayF | I'm saying that as far as I understand; ironicbaremetal.org is an oif-funded and supported thing; not neccessarily led from the technical side of openstack contribution | 16:49 |
| TheJulia | clarkb: well, the foundation chose for us, really | 16:49 |
| JayF | yes, some of us have contributed content | 16:49 |
| clarkb | I wasn't involved so I can't say one way or another. I do know that the content management system is a popular one that people enjoy using though (which goes back to differing priorities) | 16:51 |
| * TheJulia resets slides and re-dry runs presentation | 16:51 | |
| fungi | wow, lots of scrollback about this... just catching up, but the ironic website was requested by ironic project leadership in order to provide a distinct identity separate from and unhindered by public perception that it's use needs openstack. if memory serves, it was hoped this would grease the wheels for metal3 adoption so that kubernetes users wouldn't get the perception they needed | 17:53 |
| fungi | to install all of openstack with it | 17:53 |
| fungi | the foundation agreed to pay contractors to put that site together and maintain it | 17:53 |
| fungi | but the ironic team could have chosen to do so themselves with different tools and workflows | 17:53 |
| fungi | the contract website management firm the foundation hires for that stuff has a workflow and platform/framework they prefer, so that's what gets used | 17:54 |
| JayF | thanks for the historical context around that | 17:55 |
| JayF | BTW; I think the website has been pretty successful. | 17:55 |
| JayF | And honestly; even though in some ways it was created and percieved as an effort for Ironic standalone; I find it as a useful tool to introduce people into the ecosystem as a whole | 17:56 |
| fungi | sure, i just didn't want people to think the foundation pushed that site on anyone. they were asked to provide it, they paid someone to do so, and that's what's there | 17:56 |
| JayF | who would've thought; people who want bare metal server provisioning also have all the other infra stuff to manage, too :D | 17:56 |
| fungi | while i'm no fan of github pull request workflow, it's honestly 10x better than the process for managing the openstack.org website, which is mostly in a database-backed cms edited through a clicky webui | 18:00 |
| fungi | basically no audit trail much less revision control | 18:00 |
| JayF | Yeah; I think that's actually an interesting corrolary to the topic from earlier. | 18:05 |
| JayF | Where I think some folks don't want an "SCM", they want more like a "CMS" except for code/data files/etc (and github panders to this desire) | 18:06 |
| fungi | if it's netlify based like some of their other sites, then it is basically both | 18:08 |
| fungi | there's a cms interface where users can create content updates which the cms pushes as pull requests that can then be approved and get deployed | 18:09 |
| fungi | they built a netlify backend for gerrit as well, but we haven't yet hit the confluence of a project wanting to review changes in gerrit which have been created by a cms front-end | 18:09 |
| fungi | (i think you can also give netlify credentials to approve pull requests or changes, for people who want to review things purely through the cms interface, but i don't recall the details now because it was all created years ago) | 18:11 |
| JayF | Yeah, this apparently is netlify. I think mostly it's updated at this point by Julia/Iury/Me/Dmitry via PRs though. | 18:11 |
| JayF | I actually have a blog post to write on my todo list which I intend to hopefully publish both on ironicbaremetal.org and gr's blog | 18:12 |
| fungi | the only straight-up project website that's being managed through gerrit and deployed to an opendev-hosted webserver is zuul-ci.org, and i think the hope was to try switching it to the netlify+gerrit workflow, but that stalled behind converting it from hand-written html/css to gatsby templating | 18:14 |
| JayF | I'm partial to hugo myself for such things; https://github.com/jayofdoom/jay.jvf.cc | 18:15 |
| JayF | but generally static site generation is massively better than dynamic stuff | 18:15 |
| JayF | (both oldos.org and jay.jvf.cc are hugo; jay.jvf.cc uses a mostly off the shelf template but I obviously wrote my own for oldos) | 18:16 |
| fungi | though we definitely have sites like docs.openstack.org, governance.openstack.org, releases.openstack.org, security.openstack.org, docs.starlingx.io, docs.airshipit.org, and so on which are done with gerrit/zuul and hosted in opendev, the openstack/starlingx/airship/kata project sites are contracted out by the foundation | 18:17 |
| JayF | It is easy to forget the massive amount of infrastructure there is around all these projects; at openstack and oif level | 18:17 |
| TheJulia | fungi: DIY was not really presented as an option and being with-in the overall framework of a foundation asset. Or if it was, it was lost in translation. Regardless, substantial time has past along with the fine details of such. | 18:32 |
| JayF | (and to be clear; no complaints from me, it works fine, it's just interesting that even we have assets in github) | 18:33 |
| TheJulia | JayF: ditto, definitely not complaining | 18:33 |
| fungi | TheJulia: yeah, it was probably more like "if the foundation is going to manage the website we contract it out to people who use this stuff" | 18:39 |
| fungi | as i said, we definitely have at least one project whose community maintains their own primary site and hosts it in opendev | 18:40 |
| fungi | including dns hosting in opendev's nameservers | 18:40 |
| TheJulia | Yeah, zuul if memory serves | 18:41 |
| fungi | as well as numerous sub-sites for different projects' docs and stuff | 18:45 |
| stevebaker[m] | good morning | 19:51 |
| TheJulia | Good morning | 19:53 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!