| *** diablo_rojo is now known as Guest3323 | 02:26 | |
| *** pojadhav is now known as pojadhav|afk | 08:49 | |
| *** pojadhav|afk is now known as pojadhav | 10:27 | |
| *** whoami-rajat__ is now known as whoami-rajat | 11:15 | |
| *** pojadhav is now known as pojadhav|afk | 13:41 | |
| opendevreview | Slawek Kaplonski proposed openstack/governance master: Define Technical preview framework https://review.opendev.org/c/openstack/governance/+/839880 | 14:22 |
|---|---|---|
| spotz | gmann quick recap about e-r. It is currently hosted on a small private tenant at Vexxhost. The TripleO team is willing to expand the software for use by other teams but it would need infrastructure and someone to help maintain it. We don't have a fulltime person for that | 14:22 |
| fungi | note that what openstack is looking for is just an instance of the elastic-recheck daemon which periodically runs queries against the log database and generates a web page with graphs of match frequencies for the set of tracked queries. all of the backend infrastructure which handles log collection and indexing is already taken care of, just looking for an instance of the | 14:58 |
| fungi | https://opendev.org/opendev/elastic-recheck daemon pointing to the existing log database | 14:58 |
| fungi | the question was whether the tripleo/rdo elastic-recheck daemon could be pointed at https://opensearch.logs.openstack.org/ | 15:00 |
| gmann | spotz: thanks, dpawlik so if I am understanding correctly, their current hosted service cannot be used? and they need infra(server to host) and someone to maintain which is nothing but opendev situation now. | 15:52 |
| gmann | spotz: dpawlik I think key question stay same that we need server to host it. | 15:59 |
| fungi | well, finding a server isn't really a challenge. opendev has lots of quota, and we provided dpawlik with one to run the ci-logscraper system. it may have sufficient capacity to also run the periodic elastic-recheck query graph regeneration and an apache vhost to serve those to web browsers, or it could go on a new server | 16:03 |
| gmann | fungi: ok, that is good to know. let's get more clarity from tripleo team otherwise we can discuss the new server path. | 16:05 |
| fungi | the reason tripleo/rdo's elastic-recheck service was suggested is that it already has people running it. a separate server for openstack's e-r graphs would need people to run it | 16:05 |
| gmann | yeah, that can be best and win-win for all but let's see | 16:06 |
| gmann | I am not sure how a disagreement can be reflected using unappropriated word in ML :) not sure what value it give to them, may be they think scolding or being very aggressive is one way to get things in their way instead of telling us their actual requirement ? | 16:19 |
| dansmith | gmann: yeah, the current discussion is pretty frustrating | 16:49 |
| dansmith | I'm not overly offended by the use of "bullshit" in this way, even though it's pretty silly to imply that people *can't* tell what year a release is | 16:50 |
| dansmith | but it's clearly not very "nice" especially when you're trying to persuade someone to do more work for you | 16:51 |
| fungi | on a related note, i'm amused that people are so insistent on dropping release names now, but three years ago the tc thought the proposal to stop using release names after zed and switch to release numbers was an unnecessarily rash reaction | 16:51 |
| dansmith | fungi: aren't there more data points since then that change the equation? | 16:52 |
| fungi | yes, i just find it interesting how opinions can swing so wide in three short years | 16:52 |
| gmann | dansmith: exactly, we as community is working together and for everyone. but honestly saying 'bullshit' is not appropriate word if anyone use on anyone work or so but may be its just me | 16:53 |
| gmann | fungi: three years are not short :) | 16:53 |
| gmann | and yes there are more data points | 16:53 |
| fungi | the ussuri naming round was the most contentious we've had to date, and it was after that when the tc rejected the idea of switching from names to just numbers once we finished the latin alphabet | 16:56 |
| gmann | in ussuri, issue was 'why this name was removed' not 'selected name is not good'. in zed its latter one | 16:57 |
| fungi | got it, so you're saying that contention over name selection is not the driving factor, what really made the process hard was people objecting to zed after it was selected | 16:59 |
| dansmith | I'm not saying that it is "(too) hard" I'm saying that I think that we're on track to have complaints about names every cycle and it just doesn't seem worth our time to try to filter, justify, and potentially re-choose if it's going to be that frequent | 17:00 |
| gmann | and its not about who and how we are doing it. Anyone do the name or with any process such objection on selected name will stay same | 17:01 |
| fungi | makes sense. i do think debian has an easier time of it because they don't ask for community input, the release managers just pick upcomig development cycle names (many) years in advance and the community deals with the result | 17:01 |
| dansmith | if the foundation thinks it's important for marketing, then whatever, let them do it | 17:01 |
| gmann | I hope it is just OpenStack marketing they need not all openinfra projects as other projects do not do 'name' | 17:02 |
| dansmith | fungi: to be honest, debian's naming is the least useful to me of the various examples, because I'm not as plugged into it.. I can never remember what stretch, buster, etc mean without looking it up | 17:02 |
| dansmith | at least ubuntu has the date-based sequence number and I know the 04 will be an LTS | 17:03 |
| fungi | yep. and the debian community mostly doesn't object to name choices because they have no reason to believe that objecting to it would have any chance of changing the choice | 17:03 |
| dansmith | fungi: it's funny how just choosing a name that the community must "deal with" is more community-focused than evil TC members deciding that fighting over a name is less community-ish than just not having a name | 17:04 |
| dansmith | doesn't make sense to me | 17:04 |
| gmann | I heard many people in past especially during ussuri time that *we do not care much about names* and not sure why dropping them is issue now. | 17:06 |
| clarkb | gmann: I think there may be two different groups | 17:07 |
| clarkb | the people writing the software seem to not care much but the people packaging, deploying, running, using, marketing the software do | 17:07 |
| gmann | yeah, we understand the use but point is they are more painful now than its worthiness. as slaweq mentioned in review there are lot of very popular software without name so we cannot same name is must to have better marketting | 17:10 |
| dansmith | clarkb: one packager does, but the package names are done with the version not the name, am I wrong? | 17:10 |
| gmann | s/same/say | 17:10 |
| fungi | ubuntu and rdo both use the cycle names as keys for their release-specific package repositories | 17:11 |
| fungi | (that's not to say that they care about it, or would be inconvenienced by switching to numbers for that) | 17:11 |
| gmann | and also marketing depends on what name. Yoga might be good but ZED can be issue by seeing that war things community brought. | 17:11 |
| dansmith | gmann: especially in all CAPS :P | 17:12 |
| gmann | :) | 17:12 |
| clarkb | gmann: I understand the concerns. But I think people who are interested in keeping it are offering to do the work? | 17:12 |
| clarkb | Its not my decision, but thats what we've been trying to encourage people to do in opendev with the lack of help. Basically if you can own it we will do our best to support you | 17:13 |
| fungi | i think the counterargument though is that community divisiveness over a name is work for the tc and other community members even if they're not handling the name selection process itself | 17:13 |
| gmann | yeah, that is why I keep mentioning. issue is not *who does* issue is *name satisfying everything especially when it comes to cultural/historical way* | 17:15 |
| gmann | but we will be open to hear if foundation want to take it is ok but we should discuss how we will solve the existing issues we have with 'name' | 17:16 |
| dansmith | I totally don't think the arguing is going to go away if the foundation does the polling, or the picking with no polling, but I'm also willing to let those people who desperately want it and want to own the process be responsible for that | 17:17 |
| *** diablo_rojo__ is now known as diablo_rojo | 17:18 | |
| gmann | which is key point, it will end up like 'there are issues on this name' - 'TC will say ask foundation or foundation join community to solve/answer this' | 17:18 |
| fungi | yes, i agree that's the implied commitment for taking on name selection | 17:19 |
| gmann | we should not just pass the ball with issue from one place to other instead we should be clear how we handle the issues if they comes up again | 17:20 |
| gmann | asking release team is also one option but it is same thing just passing the issue | 17:21 |
| fungi | well, also any sort of decision-making, whether it's technical decisions over features in the software or social decisions like naming, requires that the people taking on the responsibility of choice are willing to upset some people (you can please some people some of the time...) | 17:22 |
| gmann | dansmith: off course, it should not be like TC asking foundation to choose name and rest process/discussion TC continue to handle. it should be complete process transfer. | 17:22 |
| dansmith | gmann: yup | 17:23 |
| dansmith | "this is yours now and if you break it, you own both pieces" | 17:23 |
| fungi | i appreciate that the tc is willing to draw a line as to what sorts of decisions it's willing to upset some people in order to make, and which sorts of decisions are just not worth that | 17:23 |
| gmann | yeah | 17:23 |
| dansmith | fungi: it's also frustrating to say "we don't think making this decision each cycle is worth the hassle" and that decision not to decide ... is a big hassle | 17:24 |
| dansmith | there's no easy way out of the box, but at least making the decision to exit hopefully ends up with less hassle long-term | 17:24 |
| * dansmith consults Rush on this topic | 17:24 | |
| fungi | the stones said "you can't always get what you want, but if you try, sometimes, you might just get what you need" ;) | 17:25 |
| dansmith | not sure that's going to happen in this case, but.. yeah :) | 17:27 |
| *** timburke__ is now known as timburke | 19:24 | |
| *** diablo_rojo is now known as Guest3381 | 19:44 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!