16:00:01 <rm_work> #startmeeting Octavia
16:00:02 <openstack> Meeting started Wed Jan  6 16:00:01 2021 UTC and is due to finish in 60 minutes.  The chair is rm_work. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:05 <openstack> The meeting name has been set to 'octavia'
16:00:15 <rm_work> Welcome back folks, and welcome to 2021!
16:00:15 <johnsom> o/
16:00:29 <rm_work> #chair johnsom
16:00:30 <openstack> Current chairs: johnsom rm_work
16:00:53 <gthiemonge> hi
16:01:01 <rm_work> oh good, someone else is actually here :P
16:01:22 <rm_work> o/
16:01:29 <rm_work> #topic Announcements
16:01:38 <rm_work> already covered this one a little bit
16:01:42 <haleyb> hi
16:01:54 <rm_work> Happy New Year! Welcome to 2021! We're living in the future! The future is now!
16:02:13 <johnsom> ohhh-Ahhhh-wow
16:02:14 <johnsom> grin
16:02:27 <johnsom> Does that mean I can retire?
16:02:41 <johnsom> Oh, maybe not *that* far into the future
16:02:46 <rm_work> In the words of my wife: no retiring! work until you die! :D
16:03:00 <rm_work> (who wins that argument is still up in the air)
16:03:23 <rm_work> (pretty sure it's gonna be me)
16:03:28 <rm_work> So any other annoucements?
16:03:31 <johnsom> Yeah, sadly it isn't coming any time soon for me
16:03:32 <rm_work> *announcements
16:03:44 <rm_work> just ride the BTC to retirement? :P
16:03:55 <rm_work> anywho...
16:04:06 <rm_work> I don't really have anything else.
16:04:30 <rm_work> How far are we from code freeze now? I feel like we've just started the cycle, so code freeze must be like next week, right? lol
16:05:13 <rm_work> cgoncalves are you around or out on vacation?
16:05:29 <johnsom> Vacation today
16:05:36 <rm_work> cool cool
16:05:40 <rm_work> me too (mentally)
16:05:48 <johnsom> MS2 is January 18t
16:05:51 <rm_work> physically, I'm right here with all of you. metaphorically.
16:06:13 <rm_work> lol ok great so kidding-not-kidding on these milestones sneaking up on us
16:06:19 <johnsom> Last day for libs is first week in March
16:06:27 <johnsom> Yes, VERY fast
16:06:27 <rm_work> anywho, moving on
16:06:30 <rm_work> #topic Brief progress reports / bugs needing review
16:06:45 <johnsom> I have been on vacation, so nothing to report
16:07:26 <gthiemonge> I have a few (lower-constraints fixes on stable branches)
16:07:47 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/768464 https://review.opendev.org/c/openstack/octavia/+/769268 https://review.opendev.org/c/openstack/octavia/+/769348 https://review.opendev.org/c/openstack/octavia/+/765853
16:07:57 <rm_work> My progress: none. I have done essentially nothing this cycle... except one bugfix attempt here: https://review.opendev.org/c/openstack/octavia/+/767648
16:08:16 <gthiemonge> last one is not passing the requirements-check yet, but I'm not losing hope
16:08:18 <rm_work> which I have no idea if it's the right approach or not, i just gave it a quick go, but turns out we don't need that patch internally so I kinda lost interest <_<
16:09:12 <rm_work> I'll try to pick it up again once i've gotten through the majority of the post-long-moratorium fires
16:11:23 <rm_work> so yeah... guess that's it
16:11:31 <rm_work> #topic Open Discussion
16:11:36 <rm_work> so how bout them Mets?
16:11:46 <rm_work> anyone else have anything?
16:12:34 <johnsom> No idea, are they even playing?
16:12:43 <johnsom> I don't have anything else this week
16:12:53 <haleyb> rm_work: have a question for you on https://review.opendev.org/c/openstack/octavia/+/656811/ - there was a comment about using threshold as a %
16:13:05 * haleyb snuck that one in under the wire
16:13:09 <rm_work> I'm going to try to pick up smoking (meats) so I am spending way too much of my free time watching brisket smoking videos.
16:13:21 <rm_work> really embracing the "open" in "open discussion"
16:13:28 <rm_work> that's all I've got
16:13:35 <rm_work> hmmm yeah percentages
16:13:53 <haleyb> rm_work: i could update, think it's pretty straight forward
16:14:03 <rm_work> Percentage is....
16:14:13 <rm_work> I don't think that's what we want honestly? like...
16:14:21 <rm_work> I see why you could think you want that
16:14:39 <rm_work> but I gave it some thought and IMO at any given point there should be essentially zero
16:14:42 <johnsom> One thing on that RFE idea we need to consider is how to override that should the cloud situation require it. I.e. once we hit that threshold, then what.....
16:14:57 <rm_work> in a tiny cloud, i don't think i'd want to limit it to like 10% because that could be ONE and that doesn't tell you anything
16:15:15 <rm_work> in a huge cloud, 10% would be WAY too high, i'd still want it to stop after like 5-10
16:15:37 <johnsom> Make it a float.   0.00000001%
16:15:38 <johnsom> grin
16:15:49 <rm_work> lol i mean, point being the same percentage is wrong for both in opposite ways
16:16:02 <rm_work> 10% too small for a tiny cloud and too big for a large one
16:16:13 <rm_work> I just want to set it to like ... 5
16:16:19 * rm_work shrugs
16:16:21 <haleyb> yeah, it's knowing what to set it at...
16:16:30 <rm_work> 5? or 10 :D
16:16:34 <rm_work> you know your environment
16:16:41 <haleyb> and it's 3 clearly :)
16:16:43 <rm_work> you can change it as the env grows if you think you need to
16:16:52 <rm_work> I SAID FIVE, BRIAN
16:16:59 <johnsom> 42
16:17:02 <rm_work> lol
16:17:09 <haleyb> 5 is right out
16:17:35 <rm_work> anywho, percent is fine? but i think it isn't actually what people want, it's just what they think they want
16:17:37 <johnsom> Yeah, I know we started a discussion over should it be percent or absolute value. I don't remember where that came out.
16:17:40 <rm_work> dunno, could add it, but i'm not going to
16:17:52 <rm_work> someone else can do the work? but it needs to support BOTH
16:18:15 <rm_work> and not be hilariously confusing like "10" and "10%" being radically different numbers of things
16:18:39 <johnsom> Do we have a spec for this?
16:18:46 * rm_work sighs again
16:18:54 <rm_work> don't think so
16:19:00 * johnsom grins with an evil smirk
16:19:16 <haleyb> feel like i'm a stormtrooper getting some jedi-talk, but it was from https://etherpad.opendev.org/p/octavia-train-ptg
16:19:24 <haleyb> ooh, train
16:19:55 <rm_work> ah there is a story
16:20:05 <rm_work> ok so it does say % in there
16:20:49 <haleyb> but it mentions % over a time period
16:21:32 * haleyb could have looked at the story before asking...
16:21:42 <rm_work> HA, quoting michael:
16:21:43 <rm_work> Personally, I am not sure percentage is a good way to measure this as a large deployment may see a single rack outage as a small percentage of their amphora (2,000+ load balancers is not uncommon).
16:21:43 <rm_work> I am thinking the threshold might be best set to a number slightly larger than the maximum number of amphora you would expect on a single host.
16:21:56 <rm_work> --Michael Johnson, 2020-02-21
16:22:28 <haleyb> yeah, that makes more sense, so maybe just need to add to release note or comment in conf file
16:22:34 <johnsom> lol, yeah, I'm open to re-thinking this.
16:22:46 <rm_work> no, no rethinking, i agree with you
16:22:53 <rm_work> your opinion is set in stone right there :P
16:23:07 <johnsom> Oh, I meant the percentage thing in the story
16:23:09 <johnsom> grin
16:23:16 <rm_work> heh
16:26:30 <rm_work> ok soooo
16:26:37 <rm_work> i guess that's it
16:27:02 <haleyb> i guess
16:27:03 <rm_work> unless we want to continue this percentage discussion for another 30m
16:27:08 <rm_work> we CAN do that :P
16:27:09 <johnsom> This is a priority for us to get merged
16:27:23 <rm_work> IMO it is fine as-is
16:27:42 <rm_work> but if you want to add percentage parsing too ... i guess... you could do that if you want
16:28:43 <haleyb> it would have to be a separate option imo
16:29:17 <johnsom> Yeah, I agree
16:29:37 <johnsom> Personally I would implement one way and see if there is a need for the other.
16:31:28 <haleyb> johnsom: ack, maybe we can just add your "threshold slightly larger than # amps on a compute node" to the release note?
16:31:40 <johnsom> Sure
16:32:04 <haleyb> i'll send a quick update since it's needs a rebase anyway most likely
16:32:09 <johnsom> So it looks like an absolute value is what is implemented in the patch, so I think we should just move forward with that and see how it goes
16:33:48 <haleyb> sure
16:36:34 <rm_work> i commented with some examples
16:36:55 <johnsom> +1
16:40:02 <johnsom> Any thing else today?
16:40:34 <rm_work> nope
16:40:36 <rm_work> have fun ya'll
16:40:41 <rm_work> #endmeeting