14:03:46 <slaweq> #startmeeting neutron_drivers
14:03:46 <opendevmeet> Meeting started Fri Aug  6 14:03:46 2021 UTC and is due to finish in 60 minutes.  The chair is slaweq. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:46 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:03:46 <opendevmeet> The meeting name has been set to 'neutron_drivers'
14:03:48 <mlavalle> o/
14:03:51 <slaweq> hi
14:03:52 <njohnston> o/
14:03:54 <lajoskatona> Hi
14:04:04 <slaweq> sorry for being late, another meeting was longer then I expected :)
14:04:10 <manub> hi
14:04:15 <amotoki> hi
14:04:48 <johnsom> o/
14:05:00 <slaweq> I think we have quorum already so we can start
14:05:29 <slaweq> #topic RFEs
14:05:52 <slaweq> we have one RFE to discuss today
14:05:53 <slaweq> https://bugs.launchpad.net/neutron/+bug/1936408
14:06:27 <slaweq> it was briefly discussed few weeks back but we need to decide something about it
14:06:42 <opendevreview> Lajos Katona proposed openstack/neutron-tempest-plugin master: WIP: Add tap-as-a-service API tests  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/803752
14:07:11 <johnsom> I want to start by highlighting that this statement in the description is inaccurate: "However, other users (Octavia) expect the quota driver to fail when lowering the quota limit under the existing resource usage." Octavia does not expect this nor does the Octavia team want this to change from the current functionality.
14:07:50 <slaweq> johnsom: thx for clarification
14:08:21 <johnsom> I think the reporter got confused by a BZ that mentioned Octavia due to the fact that Octavia and Neutron both handle quotas the same way.
14:08:38 <johnsom> This same RFE was opened against Octavia as well.
14:09:02 <slaweq> TBH I tend to agree with Brian's comment that this new behaviour doesn't makes too much sense
14:09:14 <slaweq> https://bugs.launchpad.net/neutron/+bug/1936408/comments/6
14:10:01 <johnsom> Yes, I think the current behavior is an important feature. It allows you to restrict a project from creating or rebuilding resources without forcing the deletion of currently in use resources.
14:11:08 <slaweq> johnsom: exactly, and IMO if there are other projects which bahaves differently, IMHO such projects should change (fix) it
14:11:20 <slaweq> but that's just my opinion
14:11:26 <johnsom> slaweq +1
14:11:35 <johnsom> My opinion as well
14:12:21 <johnsom> What do others think?
14:12:22 <amotoki> I tend to agree with slaweq and Brian. I wonder why nova (mentioned in the bug description) adopts the behavior to check the current usage when lowering quota
14:13:15 <njohnston> Does refstack reference this behavior at all? Would a cloud with this behavior enabled still be strictly within full cross-compatibility with one that does not?
14:13:49 <slaweq> njohnston: that's good question, I don't know really
14:13:54 <johnsom> I don't think refstack covers quota management. At least I don't remember it being a part of the test suite.
14:14:27 <opendevreview> Mamatisa Nurmatov proposed openstack/neutron master: Use neutron-lib payloads for PORT_FORWARDING  https://review.opendev.org/c/openstack/neutron/+/802990
14:14:34 <njohnston> ok, good to know
14:14:36 <slaweq> njohnston: probably it's like johnsom said, IIRC refstack is not checking "admin" actions
14:14:37 <johnsom> Quotas have historically been a consistency problem across all of the projects in OpenStack.
14:15:22 <amotoki> johnsom: yeah, one of headaches :(
14:15:23 <yamamoto> i agree the current behavior is useful. otoh i don't see a clear motivation to change it.
14:15:34 <mlavalle> gray area where we haven't put effort to achieve consistency
14:15:55 <mlavalle> I don't see the need to change behavior, either
14:17:31 <johnsom> Plus you can still set the quota limit as described in the RFE with the current implementation.
14:18:17 <njohnston> I see no upside to it
14:19:40 <johnsom> So, on the Octavia RFEs, I think I will push back that the current implementation is a feature. Starting the conversation towards an rfe-denied
14:20:05 <njohnston> +1
14:20:06 <slaweq> ok, so I think that we have agreement to deny this RFE, right?
14:20:11 <mlavalle> +1
14:20:34 <johnsom> +1 though my vote in drivers hasn't counted for years... lol
14:20:41 <slaweq> and as a follow up I will send email to the ML to try to start discussion about projects which behaves differently and maybe that could be changed to make it consistent
14:20:46 <amotoki> +1
14:21:09 <mlavalle> johnsom: your opinions are highly appreciated and influential, though
14:21:18 <amotoki> also +1 for slaweq's proposal to raise it to the list
14:21:21 <slaweq> maybe that could be even idea for community goal :)
14:22:06 <yamamoto> +1
14:22:29 <slaweq> ok, so we have agreement on that one
14:22:37 <slaweq> thank You all for Your inputs
14:22:47 <slaweq> and that was only RFE for today
14:23:00 <slaweq> #topic On Demand
14:23:16 <slaweq> do You have any other RFE/topic You want to discuss today?
14:23:54 <mlavalle> not me
14:23:59 <amotoki> nothing from me
14:24:00 <yamamoto> nothing from me
14:24:18 <slaweq> I have one last thing for today then :)
14:24:23 <slaweq> meetings in next 2 weeks
14:24:39 <slaweq> next week there is recharge day in Red Hat so we will be off
14:24:52 <opendevreview> Lucas Alvares Gomes proposed openstack/neutron master: [OVN] Fix Router Availability Zones for segmented networks  https://review.opendev.org/c/openstack/neutron/+/803759
14:25:00 <slaweq> giving that I'm going to cancel next week's meeting as we will probably not have quorum
14:25:17 <mlavalle> ack
14:25:42 <slaweq> and then meeting on 20.08 (in 2 weeks) - I will be offline also as I'm going to Eurodisnayland with my kids :D
14:26:01 <slaweq> so I can cancel it too or maybe mlavalle can chair it
14:26:06 <slaweq> wdyt
14:26:08 <slaweq> ?
14:26:19 <mlavalle> I can host the meeting
14:26:34 <slaweq> thx mlavalle
14:26:50 <slaweq> I will be at work on wednesday 18.08 so I will prepare agenda for the meeting then
14:27:00 <slaweq> and will send You an email with it
14:27:16 <amotoki> perhaps mickey mouse can vote on behalf of slaweq
14:27:26 <slaweq> amotoki: LOL
14:27:27 <mlavalle> and yes, enjoying Disneyland with your kids is a rite of passage for every self respecting parent
14:27:56 <slaweq> I can't wait for that trip :)
14:28:03 <amotoki> :)
14:28:05 <njohnston> have fun!
14:28:19 <slaweq> so that's all for today
14:28:35 <slaweq> thx for attending the meeting, have a great weekend and see You online o/
14:28:51 <johnsom> Thanks everyone
14:28:53 <slaweq> #endmeeting