| *** dangtrinhnt has joined #openstack-tc | 00:11 | |
| *** jamesmcarthur has joined #openstack-tc | 00:15 | |
| *** jamesmcarthur has quit IRC | 00:31 | |
| *** jamesmcarthur has joined #openstack-tc | 00:31 | |
| *** jamesmcarthur has quit IRC | 00:35 | |
| *** diablo_rojo has quit IRC | 00:38 | |
| *** mriedem has quit IRC | 00:38 | |
| *** diablo_rojo has joined #openstack-tc | 00:43 | |
| *** ricolin has joined #openstack-tc | 00:56 | |
| ricolin | o/ | 00:57 |
|---|---|---|
| fungi | and now it's office hour | 01:00 |
| gmann | o/ | 01:00 |
| diablo_rojo | o/ | 01:09 |
| *** whoami-rajat has joined #openstack-tc | 01:31 | |
| fungi | since there's not much else going on, worth mentioning i put together a draft list of project moves/renames for friday based on the rules jroll enumerated: http://paste.openstack.org/show/749402/ | 01:35 |
| fungi | the opendev infra team, after great bikeshedding, decided that unofficial projects which don't request their own namespaces are moving to a namespace called "x" | 01:36 |
| fungi | also, if there are any official teams asking to add deliverables in openstack, we should try to get their pending governance changes flushed so they don't get renamed out of the openstack namespace on friday | 01:38 |
| gmann | fungi: it will be 'x' or 'unknown' ? as mentioned in https://governance.openstack.org/tc/resolutions/20190322-namespace-unofficial-projects.html | 01:41 |
| fungi | it will be x. the openstack project doesn't actually have jurisdiction to decide what's happening to non-openstack projects other than to declare they're not able to share the openstack namespace | 01:43 |
| fungi | so "unknown" was a placeholder | 01:43 |
| fungi | as well as maybe a suggestion | 01:43 |
| fungi | the opendev infra team ended up wanting to go with something shorter and perhaps less objectionable | 01:44 |
| gmann | k | 01:45 |
| clarkb | ya I wanted to go with something that would avoid implying anything | 01:50 |
| clarkb | 'unknowm' could be seen as a negtive | 01:51 |
| fungi | as opposed to "x" which is how you mark buried treasure on a map ;) | 01:51 |
| gmann | i am just concern about that resolution. nothing else :) | 01:52 |
| clarkb | fungi: makes the spot | 01:53 |
| clarkb | er marks | 01:53 |
| fungi | the resolution is fine. any parts which were overreaching can just be considered suggestions | 01:53 |
| gmann | there are lot of x/*. it might be good to put that list on ML if anyone care about their repo/project | 01:54 |
| clarkb | I sent the list out earlier today | 01:54 |
| fungi | clarkb linked it from the updates he just sent | 01:54 |
| gmann | thanks | 01:54 |
| fungi | it's a bit long to embed directly in an ml post (825 entries) | 01:54 |
| gmann | for example this one might be needed under openstack somewhere - openstack/upstream-institute-virtual-environment -> x/upstream-institute-virtual-environment | 01:55 |
| fungi | yep, that draft list was intended to solicit feedback | 01:57 |
| *** Sundar has quit IRC | 01:58 | |
| *** diablo_rojo has quit IRC | 02:05 | |
| gmann | 'gerrit-dash-creator' was not under infra ? | 02:06 |
| *** mriedem has joined #openstack-tc | 02:11 | |
| fungi | i think that's something sdague wrote years ago, wasn't an official infra deliverable | 02:23 |
| gmann | yeah, i just remember it because recently i fixed +A thing there :) | 02:25 |
| *** lbragstad has quit IRC | 02:30 | |
| *** diablo_rojo has joined #openstack-tc | 02:46 | |
| *** dangtrinhnt has quit IRC | 02:59 | |
| *** dangtrinhnt has joined #openstack-tc | 02:59 | |
| *** diablo_rojo has quit IRC | 03:44 | |
| *** diablo_rojo has joined #openstack-tc | 03:44 | |
| *** mriedem has quit IRC | 04:08 | |
| *** david-lyle has joined #openstack-tc | 04:13 | |
| *** dklyle has quit IRC | 04:16 | |
| *** e0ne has joined #openstack-tc | 04:35 | |
| *** e0ne has quit IRC | 04:42 | |
| *** e0ne has joined #openstack-tc | 04:44 | |
| *** sapd1 has quit IRC | 04:50 | |
| *** zhurong has quit IRC | 05:00 | |
| *** zhurong has joined #openstack-tc | 05:07 | |
| *** e0ne has quit IRC | 05:16 | |
| *** e0ne has joined #openstack-tc | 05:40 | |
| *** e0ne has quit IRC | 05:59 | |
| *** diablo_rojo has quit IRC | 06:45 | |
| *** tosky has joined #openstack-tc | 07:29 | |
| *** dtantsur|afk is now known as dtantsur | 07:33 | |
| *** e0ne has joined #openstack-tc | 07:51 | |
| *** jpich has joined #openstack-tc | 08:01 | |
| *** e0ne has quit IRC | 08:38 | |
| *** e0ne has joined #openstack-tc | 08:42 | |
| *** e0ne has quit IRC | 08:43 | |
| *** e0ne has joined #openstack-tc | 08:43 | |
| openstackgerrit | Ildiko Vancsa proposed openstack/governance master: Add Upstream Institute training image repo under Docs https://review.openstack.org/653360 | 08:58 |
| *** e0ne has quit IRC | 08:58 | |
| *** e0ne has joined #openstack-tc | 08:59 | |
| *** e0ne has quit IRC | 09:02 | |
| asettle | Morning o/ | 09:09 |
| *** e0ne has joined #openstack-tc | 09:09 | |
| *** zhipeng has joined #openstack-tc | 09:10 | |
| *** e0ne has quit IRC | 09:37 | |
| *** e0ne has joined #openstack-tc | 09:40 | |
| *** e0ne has quit IRC | 09:46 | |
| *** e0ne has joined #openstack-tc | 09:53 | |
| *** e0ne has quit IRC | 10:27 | |
| *** e0ne has joined #openstack-tc | 10:31 | |
| *** e0ne has quit IRC | 12:00 | |
| *** dtantsur is now known as dtantsur|brb | 12:07 | |
| *** e0ne has joined #openstack-tc | 12:10 | |
| openstackgerrit | Alexandra Settle proposed openstack/governance master: Add PDF doc generation for project docs goal for Train https://review.openstack.org/647712 | 12:12 |
| *** zhurong has quit IRC | 12:20 | |
| *** e0ne has quit IRC | 12:34 | |
| *** ijolliffe has joined #openstack-tc | 12:40 | |
| *** lbragstad has joined #openstack-tc | 12:56 | |
| *** mriedem has joined #openstack-tc | 12:57 | |
| *** e0ne has joined #openstack-tc | 12:58 | |
| *** dtantsur|brb is now known as dtantsur | 13:07 | |
| *** e0ne has quit IRC | 13:30 | |
| TheJulia | Good mroning | 13:50 |
| TheJulia | I'm looking through my email and curious if there is a known location for the joint leadership meeting on the 28th | 13:50 |
| smcginnis | Hmm, not listed on https://wiki.openstack.org/wiki/Governance/Foundation/28April2019BoardMeeting | 13:52 |
| TheJulia | Yeah :\ | 13:52 |
| smcginnis | I would assume it will be in one of the rooms at the conference center, but maybe ttx or one of the other foundation folks know specifics. | 13:54 |
| ttx | Asking around | 13:59 |
| jroll | tc-members: per fungi's message last night, this is the only change which appears to make a deliverable official, we should get that through today. docs ptl is +1 on it: https://review.openstack.org/#/c/653360/ | 14:06 |
| jroll | this would also be nice to get done but the dependencies aren't ready yet: https://review.openstack.org/#/c/649071/ | 14:07 |
| fungi | for any we can't get in but want to keep in openstack, we can put them into the ethercalc with a "rename" of foo/openstack to foo/openstack | 14:11 |
| fungi | like, for example, if we are unable to get the governance approvals fast-tracked | 14:12 |
| fungi | just list jroll as the contact for those | 14:12 |
| jroll | looking again, networking-l2gw-tempest-plugin seems like the only miss (should be in projects.yaml). upstream-institute-virtual-environment has a governance change up to make it official, and the rest appear that they should remain independent (in x/ or otherwise) | 14:13 |
| fungi | longer term, we probably should track things which we want to authorize for inclusion in the openstack namespace, because the opendev team is going to be looking to us for an okay any time someone wants to add or move a project into that namespace in the future too | 14:13 |
| fungi | someone suggested that moving the openstack/openstack superrepo to x/openstack seemed weird | 14:14 |
| jroll | it seems like our current tools for tracking that are fairly sufficient | 14:14 |
| fungi | frickler: ^ maybe that was you | 14:14 |
| jroll | ah, right | 14:14 |
| ttx | fungi: yeah there are a couple of repos we should save so that they don't need to move | 14:14 |
| ttx | I did post to -infra | 14:14 |
| ttx | If there is agreement there, I can propose the relevant patches | 14:15 |
| jroll | ttx: I was about to say I could do that, but go for it :) | 14:15 |
| fungi | ildikov also asked that we not move the edge group's repo out of openstack for now (and later we can talk about moving foundation/board repos into a dedicated osf namespace or something) | 14:15 |
| ttx | ok | 14:15 |
| jroll | hm, I missed that one | 14:15 |
| jroll | is the edge group's repo not part of a sig or something? | 14:16 |
| ttx | The edge group one can be moved under FEMDC SIG alright | 14:16 |
| fungi | i'll look into why networking-l2gw-tempest-plugin got missed | 14:16 |
| ttx | TheJulia: Board meeting will be at the Hyatt | 14:16 |
| smcginnis | Thanks ttx | 14:16 |
| ttx | TheJulia: I'll have more precision once we actually have a room name for it | 14:16 |
| jroll | well, this seems handled. fungi, ttx, let me know if I can help anywhere here :) | 14:16 |
| ildikov | o/ | 14:18 |
| ildikov | yes, I did asked fungi's opinion as that repo is a bit of a special flower and that's what we came up with | 14:18 |
| ildikov | happy to explore further options if people disagree | 14:18 |
| fungi | for the record, 649071 won't impact the move whether we merge it or not, since we're expressly not moving things listed as legacy openstack projects unless they're already assigned to another confirmed or pilot open infrastructure project | 14:18 |
| fungi | ildikov: ttx is suggesting moving it under the fedmc wg | 14:18 |
| ttx | SIG | 14:19 |
| ttx | not wg | 14:19 |
| fungi | er, sig, right | 14:19 |
| ildikov | that doesn't exist anymore officially | 14:19 |
| ildikov | that's my only concern with that move | 14:19 |
| ttx | Well, officially, it does | 14:19 |
| ttx | https://governance.openstack.org/sigs/ | 14:19 |
| ttx | it might be dead, but it still exists "officially" | 14:20 |
| TheJulia | ttx: awesome, thanks | 14:20 |
| ildikov | it may have not be removed from there but I remeber we kind of announced in the annual report or somewhere that it got merged into the Edge WG | 14:20 |
| ildikov | might've missed some administration steps which is unfortunate | 14:20 |
| ildikov | can look into that | 14:20 |
| ildikov | it's also an OpenStack SIG while the Edge WG is more of an OSF WG but not governed by the board | 14:21 |
| fungi | jroll: openstack/networking-l2gw-tempest-plugin isn't in reference/projects.yaml, that's why it's getting evicted in the draft list. do we have a change proposed to add it? | 14:21 |
| jroll | fungi: we don't, I'm happy to make one | 14:22 |
| ttx | we could move it to osf/ I guess. We already have a couple of things I'd like to move there | 14:22 |
| fungi | jroll: thanks! sounds good | 14:22 |
| ildikov | ttx: +1 | 14:22 |
| csatari | FEMDC exists, but suspended its operations to focus more on ECG. | 14:22 |
| csatari | A bit of circular dependency :) | 14:22 |
| fungi | ttx: we can add osf things to the ethercalc if they're not tracked in openstack/governance and the script will just pick them up | 14:22 |
| fungi | also the ethercalc overrides anything listed in reference/legacy.yaml to cover cases where repos moved or are moving to dedicated namespaces, we just include the legacy list so that anything there doesn't get evicted to the x namespace | 14:24 |
| jroll | fungi: strangely enough, networking-l2gw is retired :) | 14:24 |
| ttx | fungi: ok, if you can remind me the link for the ethercalc I'll do that now | 14:24 |
| fungi | jroll: oh! so maybe we simply missed adding it to legacy.yaml when retiring it | 14:24 |
| fungi | ttx: https://ethercalc.openstack.org/opendev-transition | 14:24 |
| fungi | i guess list yourself as the contact for any osf namespace entries | 14:25 |
| jroll | fungi: or we found a corner case - these deliverables were retired as *official* projects from neutron when they evicted the plugins, but are still alive and well | 14:25 |
| jroll | or at least active. | 14:25 |
| ildikov | csatari: I'll find to sync up with Adrien and crew to see how to administer what's up with it | 14:25 |
| ildikov | *find time :) | 14:25 |
| csatari | ok, ildikov. Thanks. | 14:26 |
| fungi | jroll: yeah, any repos we take out of reference/projects.yaml need to go in reference/legacy.yaml (that way we also count contributions for them as valid electors for those teams or the tc for a while after an abrupt removal) | 14:26 |
| openstackgerrit | Thierry Carrez proposed openstack/governance master: Adopt openstack/governance-website as TC repo https://review.openstack.org/653446 | 14:26 |
| openstackgerrit | Thierry Carrez proposed openstack/governance master: Adopt openstack/arch-wg as TC repo https://review.openstack.org/653447 | 14:26 |
| jroll | fungi: right, but this makes me realize there are projects in legacy.yaml which are unofficial and active. should those be in openstack/ or x/ ? | 14:27 |
| * ttx adds openstack/openstack while at it | 14:27 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/governance master: Add openstack/networking-l2gw-tempest-plugin to legacy.yaml https://review.openstack.org/653449 | 14:27 |
| jroll | at any rate, this change makes sense ^ | 14:28 |
| jroll | but the question is, do those now-unofficial projects belong in the openstack namespace? | 14:28 |
| fungi | jroll: i think it's a bigger discussion, likely related to my ml post about forcing projects which want to continue outside openstack governance to fork so we can retire the copies in our namespace | 14:28 |
| jroll | fungi: probably. skip it for this week, I guess? | 14:28 |
| fungi | i think we can solve it after migration | 14:29 |
| openstackgerrit | Thierry Carrez proposed openstack/governance master: Adopt openstack/openstack as TC repository https://review.openstack.org/653450 | 14:29 |
| jroll | ok | 14:29 |
| ttx | ildikov: so should I suggest openstack/edge-computing-group -> osf/edge-computing-group with you as contact? | 14:32 |
| ildikov | ttx: yes, sounds good | 14:33 |
| ildikov | ttx: thank you!! | 14:33 |
| ttx | done | 14:34 |
| ildikov | \o/ :) | 14:35 |
| *** david-lyle is now known as dklyle | 14:35 | |
| ttx | ildikov: the list also contains openstack/upstream-institute-virtual-environment | 14:38 |
| ildikov | ttx: proposed a patch today to move it under the docs umbrella | 14:39 |
| ttx | ok | 14:39 |
| jroll | that one is here: https://review.openstack.org/#/c/653360/ | 14:39 |
| ildikov | https://review.openstack.org/#/c/653366/ | 14:39 |
| ildikov | so it lives at the same place as the training-guides repo | 14:40 |
| ildikov | jroll: correct, thanks, I grabbed the wrong link | 14:41 |
| jroll | no worries | 14:41 |
| gmann | 'microversion-parse' is the another i feel should be added somewhere but do know where | 14:42 |
| fungi | as a fallback, if we want to avoid something moving out of openstack but don't know where we want to track it officially yet, we *can* stick it in the ethercalc "renamed" to itself and that should cause it to not get moved | 14:44 |
| mnaser | tc-members: I'd like folks to please pay attention to this review -- https://review.openstack.org/#/c/645361/4 | 14:45 |
| fungi | i just want to make sure, if we do that, we also circle back around after the migration and decide what to do with it | 14:45 |
| mnaser | it has minimal number of votes, but it's pretty big, so I'd be more comfortable with more eyes on it. | 14:45 |
| * lbragstad is around for questioning if needed | 14:49 | |
| *** AJaeger has joined #openstack-tc | 14:50 | |
| AJaeger | jroll, I commented on https://review.openstack.org/#/c/653449 - that plugin repo was never under governance, it was created after the neutron-stadium changes and retirement of the other repo. So, a nice corner case | 14:52 |
| *** e0ne has joined #openstack-tc | 14:53 | |
| openstackgerrit | Merged openstack/governance master: Rename Chef OpenStack project and update it https://review.openstack.org/648356 | 14:55 |
| *** weshay is now known as weshay|rover | 14:57 | |
| gmann | fungi: ok but at the end (after migration), it cannot stay as openstack/ because that is not official project right ? so now or later they need to be moved. | 14:59 |
| fungi | gmann: well, we need to discuss how we handle projects if they were official repos at any point in the past | 15:00 |
| fungi | since we've had users get confused after we removed active projects from openstack and assume we're still in control of them | 15:01 |
| mnaser | that's a good point | 15:02 |
| mnaser | we don't want things to just start randomly 404-ing either :\ | 15:02 |
| mnaser | unless we can do a redirect when we remove something | 15:02 |
| fungi | well, the redirect is part of the problem, in my opinion. i've been leaning toward requiring them to fork, and retiring the copy in our namespace | 15:02 |
| fungi | the readme in the retired repo can then clearly state that the effort is being continued by some people outside of openstack, and where to find them | 15:03 |
| fungi | for example, if we'd retired all the fuel repos and forced them to fork copies into a fuel namespace | 15:04 |
| fungi | i think there'd have been far fewer users assuming it was still an openstack project when the copies in the openstack namespace are empty except for a readme | 15:05 |
| fungi | whereas moving them and leaving a redirect behind means folks pulling from the old git remote never notice, unless they pay attention to the redirect mentioned on stderr by the git utility | 15:06 |
| fungi | so they're still consuming openstack/fuel-whatever as far as they know | 15:06 |
| fungi | and in this new world where we're going to start equating the openstack namespace with official status of some kind, that can be significantly misleading | 15:07 |
| fungi | (i think it's less of a concern for stuff which moved before or during this week, when the meaning of the openstack namespace is still very muddy) | 15:07 |
| gmann | yeah, i agree on that point. there are lot people (specially not involved in community ) needs to know that change. | 15:08 |
| gmann | i mainly concern about 'microversion-parse' which is very much close and only(till now) for openstack microvesion only but not official. let me propose to keep it under openstack/ in ethercal and later we decide if we can add it somewhere like in olso etc | 15:11 |
| clarkb | gmann: cdent repsonded to that and didn't seem to want to move it | 15:12 |
| clarkb | er didn't want to move it under openstack/ | 15:12 |
| gmann | clarkb: yeah. I will put my suggestion to his reply if he like. there are more classes i feel we can put under this lib from nova repo if that lib stay somewhere in openstack | 15:15 |
| jroll | AJaeger: bah, that's fun. I'm not sure what to do with it then. | 15:44 |
| jroll | seems like either l2gw should be in x/ or l2gw-tempest-plugin should be in openstack/, I don't know the best way to accomplish (and document) that cleanly | 15:45 |
| *** adriant has quit IRC | 15:51 | |
| *** adriant has joined #openstack-tc | 15:53 | |
| fungi | i think we could keep it in openstack for the migration, but then have a bigger discussion about what to do with stuff in legacy.yaml which isn't retired but is still being actively developed while no longer under our governance | 15:57 |
| fungi | (and i know that repo specifically isn't in legacy.yaml, but the same team seems to be caring for it as another one which is) | 15:58 |
| fungi | friday is far from being our only opportunity to move things to different namespaces, just a good opportunity to do any which are easy decisions | 15:59 |
| *** diablo_rojo has joined #openstack-tc | 15:59 | |
| fungi | kicking the harder choices down the road a ways so we can focus on other priorities seems a reasonable approach | 15:59 |
| *** dims has quit IRC | 16:07 | |
| jroll | ++ | 16:12 |
| *** jpich has quit IRC | 16:20 | |
| *** dtantsur is now known as dtantsur|afk | 16:24 | |
| *** david-lyle has joined #openstack-tc | 16:24 | |
| *** dangtrinhnt_ has joined #openstack-tc | 16:25 | |
| *** dangtrinhnt has quit IRC | 16:27 | |
| *** dklyle has quit IRC | 16:27 | |
| *** dims has joined #openstack-tc | 16:54 | |
| *** dims has quit IRC | 16:59 | |
| *** e0ne has quit IRC | 17:01 | |
| *** dims has joined #openstack-tc | 17:01 | |
| *** ricolin has quit IRC | 17:04 | |
| mnaser | as an update on tricircle, I've been in touch with them and they're planning to start using their meeting slot once again | 17:07 |
| mnaser | the PTL has been taking initiative and trying to also help out with people to get in the IRC channels | 17:07 |
| mnaser | and I think their plan is to start having meetings from now on there | 17:07 |
| mnaser | so slowly but surely, I think with right guidance (and thanks to help from Horace), we'll be able to help support them within our community | 17:07 |
| gmann | \o/ | 17:11 |
| gmann | thanks mnaser | 17:18 |
| fungi | that's the best possible outcome | 17:45 |
| *** david-lyle is now known as dklyle | 17:55 | |
| *** AJaeger has left #openstack-tc | 18:43 | |
| *** e0ne has joined #openstack-tc | 19:17 | |
| *** e0ne has quit IRC | 19:20 | |
| *** e0ne has joined #openstack-tc | 19:22 | |
| *** tjgresha has quit IRC | 19:33 | |
| *** tjgresha_nope has joined #openstack-tc | 19:33 | |
| *** tjgresha_nope has quit IRC | 19:40 | |
| *** tjgresha has joined #openstack-tc | 19:44 | |
| *** tjgresha has quit IRC | 19:51 | |
| diablo_rojo | +2 Best possible outcome :) | 19:56 |
| diablo_rojo | thanks mnaser! | 19:56 |
| openstackgerrit | Ghanshyam Mann proposed openstack/governance master: Add IPV6 support and testing goal for Train https://review.openstack.org/653545 | 20:01 |
| gmann | tc-members: i have drafted the IPV6 support and testing goal proposal: https://review.openstack.org/653545 | 20:02 |
| gmann | lbragstad: evrardjp: ^^ | 20:02 |
| lbragstad | thanks gmann | 20:03 |
| *** ijolliffe has quit IRC | 20:03 | |
| TheJulia | Is it wrong that I'm excited by just the mention of ipv6 support and testing? | 20:09 |
| gmann | lbragstad: you want to send this to ML for review or you will include in your goal update email | 20:09 |
| gmann | TheJulia: i am just targeting for all services listen on ipv6 and we have testing job to test the same | 20:10 |
| lbragstad | yeah - an update to the mailing list would be good | 20:10 |
| gmann | there are lot more scenario for ipv6 which can be tested one by one | 20:10 |
| gmann | lbragstad: ok. thanks | 20:10 |
| lbragstad | fwiw - i don't think evrardjp or myself were planning on sending another update about goal progress, at least not this late | 20:11 |
| lbragstad | but i'll send something out for this specifically, calling for reviews | 20:11 |
| TheJulia | gmann: Well, I _hope_ it may also help us drive towards ensuring that devstack runs or is executed such that there is always working v6 on the test VM | 20:11 |
| gmann | lbragstad: ok | 20:11 |
| gmann | yeah | 20:12 |
| TheJulia | gmann: It mentally goes back to trying to do it for ironic and hitting roadblocks with test VMs what were v4 only :( | 20:12 |
| mugsie | this will cause some projects to have to test the nova VM IPv6 capabilities as well | 20:12 |
| TheJulia | s/nova/neutron/? | 20:13 |
| mugsie | e.g. octavia ssh's to VMs, and can have VMs call back to the octavia API (which is Ipv6 only will mean the VMs will have to have v6) | 20:13 |
| gmann | you mean connectivity between VMs over ipv6 ? | 20:13 |
| mugsie | yeah | 20:13 |
| mugsie | VM -> control plane over v6 anyway | 20:14 |
| TheJulia | v6 all the things! | 20:14 |
| johnsom | +1 v6 all the things! | 20:14 |
| mugsie | and heat may fall into that bucket with software deployments as well? | 20:14 |
| fungi | a related issue came up ni #openstack-infra earlier today... oslo_service apparently defaults to listening on 0.0.0.0 instead of :: or dynamically identifying available address families | 20:14 |
| TheJulia | do we need v6 all the things stickers? | 20:14 |
| fungi | i think we at least need to make sure our stickers are ipv6-ready ;) | 20:15 |
| mugsie | but it means that the line 34 note may need to be removed to let others test | 20:15 |
| mugsie | (I am ++ on that though :) ) | 20:15 |
| TheJulia | fungi: only if I somehow find the arin stickers a friend gifted me | 20:16 |
| gmann | fungi: that might not be aligned with what that goal propose, we need each service to dynamically identify the listening address | 20:16 |
| fungi | gmann: yeah, i assume that to be the case | 20:17 |
| gmann | mugsie: that is something i was testing on this https://review.openstack.org/#/c/653161/ | 20:17 |
| TheJulia | I suspect that should be a change oslo_service should make to default to making v6 socket binds | 20:18 |
| gmann | but i did not find VM assigned the v6 address by neutron or may be more setting we need | 20:18 |
| clarkb | we actually already do a really good job on this (its definitely not complete but comapred to say docker which completely breaks if you give it an ipv6 address to push to its great) | 20:18 |
| gmann | mugsie: so i mentioned VM connectivity thing out of scope for this goal | 20:19 |
| TheJulia | neutron v6 networking is a little... different and we shouldn't drive chagnes that could negatively impact that experience by creating a focus on v6. The key is going to ensure that the substrate all works completely on v6. I suspect this will align well with many developers. | 20:20 |
| clarkb | there are (or were) tempest tests that conncet to test VMs over ipv6 | 20:21 |
| clarkb | we had to update cirros to get that working iirc | 20:21 |
| mugsie | gmann: but it can't be out of scope, if we want octavia or similar projects to be able to test | 20:21 |
| clarkb | the biggest gap is probably what mugsie points out. It is where our services interact with compute resources | 20:22 |
| clarkb | we also have a job that tests the apis operate over ipv6 too iirc | 20:22 |
| gmann | clarkb: i am trying with IP_VERSION=6 on this - https://review.openstack.org/#/c/653161/ but as you mentioned it might need test image updates | 20:23 |
| clarkb | gmann: there are specific tests that do that already | 20:23 |
| clarkb | or were, maybe we removed them? that would be sad | 20:23 |
| clarkb | they even spun up multiple VMs and checked ssh between the VMs over ipv6 and not just tempest to the VM | 20:24 |
| gmann | there were not removed, let me check | 20:24 |
| TheJulia | maybe instead... the focus should be "I don't want v4 in my openstack" | 20:24 |
| gmann | https://github.com/openstack/tempest/blob/master/tempest/scenario/test_network_v6.py | 20:24 |
| * TheJulia runs to the doctor, bbl | 20:25 | |
| clarkb | gmann: my guess is that your chagne fails because not all VMs are told to boot with the v6 network | 20:25 |
| johnsom | We spin up cirros with IPv6 addresses so we can test IPv6 member servers behind the load balancer. | 20:25 |
| clarkb | but when they are told to do so it works aiui and we test that | 20:25 |
| gmann | yeah | 20:25 |
| johnsom | FYI: https://review.openstack.org/#/c/611405/ | 20:26 |
| gmann | may be i was little over thinking on those- run tempest with ipv4 and ipv6 and both pass. But it may need lot of work. | 20:27 |
| gmann | but not complete tempest run is needed to prove openstack VM working on ipv6 | 20:27 |
| clarkb | no but by testing it as part of the main suite we ensure it is always tested | 20:28 |
| smcginnis | My experience has been like what TheJulia said - tests can pass fine with IPv6 enabled, but the real issues start when you want to ensure there is nothing at all still using IPv4 somewhere in there. | 20:29 |
| clarkb | smcginnis: that is unfortunately incredibly difficult to test because we rely on ipv4 to get to the test VMs | 20:29 |
| gmann | clarkb: yeah, let me update my tempest patch to run ipv6 tests which are marked as 'slow' and being skipped in tempest-full. | 20:30 |
| smcginnis | clarkb: Yeah. The only way we were able to do it in a past life was set up an isolated special purpose lab. | 20:30 |
| clarkb | we could have the test suite disable ipv4 on the test node, disappear for an hour, then turn ipv4 back on so that zuul can ask it if things worked | 20:30 |
| clarkb | but that is likely to be fragile | 20:30 |
| *** e0ne has quit IRC | 20:31 | |
| clarkb | another approach may be to firewall or tcpdump and then check if ipv4 was used at all when done for anything otherthan the zuul communication | 20:32 |
| clarkb | if it were me I'd probably set up a job that defaults to ipv6 for things but doesn't try very hard to disable ipv4. Then check a pcap for content on the ipv4 address | 20:32 |
| clarkb | that will also help you identify what is using ipv4 so that it can be fixed | 20:33 |
| *** tjgresha has joined #openstack-tc | 20:34 | |
| *** tjgresha has quit IRC | 20:39 | |
| *** tjgresha has joined #openstack-tc | 20:54 | |
| *** tjgresha has quit IRC | 20:56 | |
| *** whoami-rajat has quit IRC | 21:01 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/governance master: Add IPV6 support and testing goal for Train https://review.openstack.org/653545 | 21:13 |
| *** logan- has quit IRC | 21:34 | |
| *** logan- has joined #openstack-tc | 21:37 | |
| *** mriedem has quit IRC | 22:10 | |
| *** lbragstad has quit IRC | 22:33 | |
| *** lbragstad has joined #openstack-tc | 22:35 | |
| *** jamesmcarthur has joined #openstack-tc | 23:07 | |
| *** jamesmcarthur has quit IRC | 23:23 | |
| *** tosky has quit IRC | 23:38 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!