*** jamesmcarthur has joined #openstack-tc | 02:24 | |
*** jamesmcarthur has quit IRC | 02:50 | |
*** whoami-rajat has joined #openstack-tc | 02:52 | |
*** jamesmcarthur has joined #openstack-tc | 03:51 | |
*** jamesmcarthur has quit IRC | 04:18 | |
*** ricolin has joined #openstack-tc | 04:27 | |
*** jamesmcarthur has joined #openstack-tc | 05:19 | |
*** jamesmcarthur has quit IRC | 05:23 | |
openstackgerrit | zhurong proposed openstack/governance master: Retire murano-deployment https://review.openstack.org/628860 | 05:27 |
---|---|---|
*** diablo_rojo has joined #openstack-tc | 05:50 | |
*** Luzi has joined #openstack-tc | 06:57 | |
*** tosky has joined #openstack-tc | 07:45 | |
*** diablo_rojo has quit IRC | 08:02 | |
*** jpich has joined #openstack-tc | 08:57 | |
*** ssbarnea|bkp2 has joined #openstack-tc | 09:37 | |
*** ssbarnea has quit IRC | 09:39 | |
ttx | bsilverman: regarding KubeCon attendees not "knowing" OpenStack: that conference is really for cloud-native application developers/deployers, who all assume a cloud (infrastructure) is present. They may consume it from AWS, GCE, Azure or OpenStack, they don't really care that much about that layer. It's basically not their job to care about that layer. So I'm not surprised. | 09:42 |
ttx | That is also why an open infrastructure summit makes sense. To talk to the infrastructure providers that want to use open source solutions | 09:44 |
*** jaosorior has joined #openstack-tc | 10:03 | |
*** dtantsur|afk is now known as dtantsur | 10:08 | |
*** cdent has joined #openstack-tc | 11:10 | |
*** dangtrinhnt has joined #openstack-tc | 11:28 | |
*** jaypipes has joined #openstack-tc | 12:51 | |
*** evrardjp has joined #openstack-tc | 12:54 | |
*** evrardjp has quit IRC | 12:59 | |
*** evrardjp has joined #openstack-tc | 12:59 | |
*** cdent has quit IRC | 13:20 | |
*** cdent has joined #openstack-tc | 13:34 | |
*** jamesmcarthur has joined #openstack-tc | 13:41 | |
fungi | granted, in some cases the people deploying kubernetes may want to run it on hardware under their own control, and at that point openstack becomes an attractive option compared to commercial/proprietary software or third-party-managed on-premises black boxes | 13:44 |
evrardjp | o/ | 13:57 |
*** whoami-rajat has quit IRC | 14:01 | |
*** cdent has quit IRC | 14:11 | |
*** e0ne has joined #openstack-tc | 14:12 | |
*** irclogbot_3 has quit IRC | 14:14 | |
*** jamesmcarthur has quit IRC | 14:18 | |
*** e0ne has quit IRC | 14:19 | |
dims | o/ | 14:29 |
*** needsleep is now known as TheJulia | 14:36 | |
*** irclogbot_3 has joined #openstack-tc | 14:38 | |
openstackgerrit | Merged openstack/governance master: Technical vision: hide implementation details https://review.openstack.org/628181 | 14:41 |
*** jamesmcarthur has joined #openstack-tc | 14:42 | |
*** e0ne has joined #openstack-tc | 14:46 | |
*** e0ne has quit IRC | 14:47 | |
*** irclogbot_3 has quit IRC | 15:06 | |
*** cdent has joined #openstack-tc | 15:07 | |
*** irclogbot_3 has joined #openstack-tc | 15:12 | |
*** openstackgerrit has quit IRC | 15:22 | |
*** Luzi has quit IRC | 15:24 | |
*** e0ne has joined #openstack-tc | 15:42 | |
*** e0ne has quit IRC | 15:50 | |
*** openstackgerrit has joined #openstack-tc | 15:53 | |
openstackgerrit | Merged openstack/governance master: New Sahara repositories for split plugins https://review.openstack.org/628210 | 15:53 |
*** jamesmcarthur has quit IRC | 16:17 | |
*** jamesmcarthur_ has joined #openstack-tc | 16:17 | |
*** jamesmcarthur_ has quit IRC | 16:22 | |
*** jamesmcarthur has joined #openstack-tc | 16:22 | |
*** whoami-rajat has joined #openstack-tc | 16:28 | |
bsilverman | ttx: I don't agree it's not their job to care. People keep saying that the developers don't care about infrastructure and that's simply not true. Even if they don't care, they need it. They cared plenty when they spoke about whether to use AWS EKS or GKE, so they do care about where it's running and what it's running on. My concern was that by not being more visible during KubeCon is squandering an opportunity to | 16:47 |
bsilverman | reach out more to a community that would benefit from OpenStack if they knew it existed, or, could take a more active role in recommending it. | 16:47 |
bsilverman | I wish I would have made a video, I had whole tables of attendees talking about Open Infrastructure and OpenStack without more than a little nudge in that direction. | 16:49 |
bsilverman | Even the use case sessions were lacking maturity in the infrastructure. Chic-fil-a still deploys individual bare metal clusters by hand to host their Kubernetes workloads and operations, scaling and HA are all done as a truck role, and/or remote into the servers (so very 1990's :) ) Their talk was all about how they use GitOps to solve CI/CD issues but all I was thinking is that they had bigger issues with infra. | 16:54 |
bsilverman | Other companies were all-in AWS but were lamenting the fact that they had no infrastructure control. | 16:54 |
*** notmyname has quit IRC | 16:55 | |
*** notmyname has joined #openstack-tc | 16:57 | |
* bsilverman is away. | 17:02 | |
TheJulia | bsilverman: I think the break in agreement is their job is to make x thing work for their employer. To change context to another platform... to want or to make it work on another platform is either driven by a business decision or an architectural need that can be solved via the move in which they can sell it to their team and management. I do like the idea of a video or something where it is talked about and stressed, | 17:11 |
TheJulia | and I suspect the appeal is to the infrastructure operators, to the techops teams going into data centers. Which makes me wonder... did you observe any commonalities from those lamenting lack of infrastructure control? Anything that stood out? | 17:11 |
ttx | bsilverman: you're right, especially wrt. KubeCon (Kubernetes is really the interface between both worlds). We have tried to be more apparent with OpenStack presence at KubeCon in the past, but that was generally declined as off-topic | 17:22 |
ttx | K8s/OpenStack SIG was really the only way to cover OpenStack at CNCFCon | 17:23 |
mnaser | i think having a presence for openstack in the context of showing off how it's the type of technology that allows you do to k8s *better* | 17:23 |
mnaser | the faces i saw when i was telling people about what the cloud provider can do were all those of excitement | 17:24 |
mnaser | maybe it's just a positioning thing | 17:24 |
ttx | I think casual mentions to OpenStack (like the CERN keynote in Copenhagen) is one of the best way to appear on the map there | 17:25 |
*** jpich has quit IRC | 17:31 | |
bsilverman | I had a good discussion with Kelsey Hightower about how he felt about OpenStack and K8s and he's actually a supporter of the platform. It made me realize that outreach is really important and that I shouldn't assume someone would be resistent to a good idea simply because of who they work for. | 17:35 |
bsilverman | ttx: I can understand why they felt it was off-topic, but that's like saying that talks about CPU pinning and line rate network hardware is off topic for the Open Infrastructure Summit because it happens at a layer below OpenStack. I think there's an argument there for inclusion beyond the SIGs, but it's their party, they make the rules :) | 17:39 |
bsilverman | I did appreciate everyone from the Foundation who was there, it was great seeing you after I had to miss Berlin. | 17:42 |
scas | my casual observation is that when people make a concerted effort to abstract the platform away from the developer in a typical company, be it deploying openstack or k8s, or something else entirely, it's often times at a detriment due to lack of proper insight and breadcrumbs to gain context. often times the right people can have the wrong context | 17:42 |
*** dtantsur is now known as dtantsur|afk | 17:45 | |
scas | end users may not care a lick what the magic internet box does behind the scenes, but anyone deploying any technology should be painfully aware of what they're working with | 17:46 |
*** diablo_rojo has joined #openstack-tc | 18:00 | |
*** diablo_rojo has quit IRC | 18:05 | |
cdent | TheJulia: should we make some kind of plan on how/when to give the cloud vision thread a bump? I'm thinking about that now because I'm starting to consider how to do the what is the tc evaluation thread. | 18:13 |
cdent | definitely too soon at this point to be concerned, just thinking about it | 18:13 |
TheJulia | Well, I think so yes, and I suspect maybe tomorrow might be a good idea. | 18:14 |
TheJulia | I need to make a "help wanted" list for ironic, it wouldn't be too hard for me to write a short bit of prose as an example to what we're hoping for | 18:15 |
cdent | ah, an example could be very helpful | 18:17 |
TheJulia | I'll put it on my after lunch todo list | 18:17 |
*** jamesmcarthur_ has joined #openstack-tc | 18:19 | |
*** jamesmcarthur has quit IRC | 18:19 | |
*** jamesmcarthur_ has quit IRC | 18:23 | |
*** jamesmcarthur has joined #openstack-tc | 18:23 | |
*** jamesmcarthur has quit IRC | 18:24 | |
*** jamesmcarthur has joined #openstack-tc | 18:24 | |
*** jamesmcarthur has quit IRC | 18:24 | |
*** jamesmcarthur has joined #openstack-tc | 18:25 | |
*** jamesmcarthur has quit IRC | 18:26 | |
*** jamesmcarthur_ has joined #openstack-tc | 18:26 | |
*** jamesmcarthur_ has quit IRC | 18:27 | |
*** jamesmcarthur has joined #openstack-tc | 18:32 | |
*** ricolin has quit IRC | 18:33 | |
*** jamesmcarthur_ has joined #openstack-tc | 18:34 | |
*** jamesmcarthur has quit IRC | 18:34 | |
*** jamesmcarthur_ has quit IRC | 18:40 | |
*** tosky has quit IRC | 19:05 | |
*** jamesmcarthur has joined #openstack-tc | 19:11 | |
*** jamesmcarthur has quit IRC | 19:18 | |
*** cdent has quit IRC | 19:25 | |
*** cdent has joined #openstack-tc | 19:34 | |
*** jamesmcarthur has joined #openstack-tc | 19:47 | |
*** whoami-rajat has quit IRC | 19:58 | |
dhellmann | tc-members: here's something to think about this week and for the future: at what point do we consider projects without python 3 support to be unmaintained enough that we might not want them on the official project list? not until after we drop 2 support in U, for sure. Maybe only after Jan 2020? Maybe never? | 20:17 |
dhellmann | apropos of http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001465.html | 20:18 |
cdent | good question | 20:18 |
cdent | presumably that's orthogonal from other indicators like "doing releases"? | 20:18 |
smcginnis | I would say when we drop 2 support, if a project cannot operate after that then they are unmaintained and no longer should be on the official list. | 20:19 |
smcginnis | So I think if they have not updated by U, that's kind of a hard deadline. | 20:19 |
dhellmann | I'm not sure that doing releases on unsupported platforms counts enough for me to support them. I honestly don't know how far to let things drag on. | 20:19 |
dhellmann | *if* we're going to set a deadline, we should be planning to announce that well in advance | 20:20 |
dhellmann | and we should coordinate with the trademark program wg | 20:21 |
fungi | i agree with smcginnis. whenever we decide to drop python 2 testing for master branches of official projects, that's the hard deadline | 20:21 |
dhellmann | we have not so far set a date for that. we have set a date *before which* we would not do it. | 20:21 |
TheJulia | I don't think it is a matter of a project not being able to operator, but do they do they work or not. I think the bottom line is that if they don't do the work, they are inherently un-maintained and should be dropped. | 20:21 |
fungi | but yeah, knowing what date that is far in advance would be good for communication purposes | 20:22 |
TheJulia | ++++ | 20:22 |
dhellmann | anyway, I want us to think about the answer to the question carefully, since it's likely to have a big impact on the community | 20:22 |
dhellmann | we should, for example, define "support" | 20:22 |
dhellmann | some of the projects have unit tests but not functional tests, for example | 20:23 |
TheJulia | I suspect "support" as a word should be avoided due to implications of use and meaning, where as we're really talking about currently supported python version compatibility | 20:24 |
dhellmann | yes, perhaps we should describe it in terms of the level of testing expected | 20:24 |
dhellmann | that's consistent with what we do with the PTI elsewhere | 20:24 |
TheJulia | I kind of thought we already did, tbh. I think the missing point is the "or else" | 20:25 |
dhellmann | but I think we need to be specific | 20:25 |
TheJulia | I concur | 20:25 |
clarkb | would python2 via pypy support be valid? | 20:25 |
dhellmann | we say for stein projects must target and test against 2.7 and 3.6, but we don't define "test against" to include functional tests | 20:26 |
clarkb | (I don't think anyone is doing that today, but is a supported python2 interpreter) | 20:26 |
dhellmann | clarkb : do you mean 3 there? | 20:26 |
clarkb | dhellmann: no, pypy has committed to indefinite python2 support | 20:26 |
dhellmann | I didn't realize that | 20:27 |
clarkb | so if you really want to write python2 compatible code for some reason you can use pypy | 20:27 |
cdent | I think we should not muddy the waters by providing an escape clause | 20:27 |
smcginnis | I seem to remember other issues with running with pypy, but things may have changed since then. | 20:27 |
cdent | setting a boundary on when python3 must the default is a good and worthwhile gatekeeper | 20:27 |
smcginnis | cdent: ++ | 20:27 |
cdent | s/the/be | 20:27 |
dhellmann | yeah, I'm not sure why we would want to worry about that | 20:27 |
clarkb | smcginnis: ya it would be effort, probably just as much or more than using python3 | 20:28 |
TheJulia | We want a community that is on some level moving forward, not treading water, so I totally agree with cdent's statement about not providing an escape clause. | 20:28 |
TheJulia | Which also raises something I've been wondering for the last half hour. At what point do we begin to consider re-evaluating projects that are doing minimal levels of maintenance to their code bases? | 20:30 |
dhellmann | that's another good thing for us to consider | 20:33 |
TheJulia | I'm not trying or wanting to seem mean about it, but there is a logisitcal nature of "well, it is only minor activity... and openstack's release model kind of requires a particular pattern which doesn't represent the current state of the project" and we effectively get software being released with only boilerplate activity that potentially at worst is taking room away from other projects on press releases and | 20:35 |
TheJulia | documentation. | 20:35 |
* TheJulia wonders what planet she has come from and what happened to her brain | 20:35 | |
dhellmann | well, I'd like us to eliminate any work that would lead to releases just for the sake of having a release | 20:36 |
dhellmann | I think we're mostly there, now that we don't sync requirements | 20:36 |
dhellmann | and I think it's fine to keep stable projects on the list | 20:37 |
smcginnis | That kind of goes back to some earlier discussions we've had too. What categorizes a project as "code complete" vs projects that are no longer active and maintained. | 20:37 |
dhellmann | we have a release model set up to support that | 20:37 |
dhellmann | right | 20:37 |
TheJulia | Yup | 20:37 |
smcginnis | Sometimes it's very hard to tell the difference from the outside. | 20:37 |
dhellmann | but I do think we should look at projects that are inactive through lack of attention | 20:37 |
clarkb | maybe lean on test results and skew from requirements updates to distinguish? | 20:40 |
clarkb | a code complete project in maintanance is still going to need the occasional change to keep tests working or update deps due to a security vulnerability | 20:40 |
clarkb | an abandoned project won't be doing these taks | 20:40 |
*** e0ne has joined #openstack-tc | 20:40 | |
*** cdent has quit IRC | 20:49 | |
*** jamesmcarthur has quit IRC | 20:54 | |
TheJulia | That is a really good point | 20:55 |
TheJulia | I feel like the time scale of that might be a bit long though | 20:56 |
*** jamesmcarthur has joined #openstack-tc | 21:39 | |
*** e0ne has quit IRC | 21:44 | |
*** e0ne has joined #openstack-tc | 22:02 | |
*** ianw_pto is now known as ianw | 22:26 | |
*** e0ne has quit IRC | 22:32 | |
*** diablo_rojo has joined #openstack-tc | 22:43 | |
*** tosky has joined #openstack-tc | 22:52 | |
*** diablo_rojo has quit IRC | 23:14 | |
*** ianychoi has quit IRC | 23:21 | |
*** tosky has quit IRC | 23:48 | |
-openstackstatus- NOTICE: The Etherpad service at https://etherpad.openstack.org/ has been offline since 23:22 UTC due to a hypervisor issue in our service provider, but should hopefully return to service shortly. | 23:50 | |
*** jamesmcarthur has quit IRC | 23:56 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!