20:00:05 <johnsom> #startmeeting Octavia
20:00:06 <openstack> Meeting started Wed Jul 25 20:00:05 2018 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:10 <openstack> The meeting name has been set to 'octavia'
20:00:12 <johnsom> Hi folks!
20:00:14 <cgoncalves> o/
20:00:19 <nmagnezi> o/
20:00:47 <xgerman_> o/
20:01:22 <johnsom> #topic Announcements
20:01:38 <johnsom> Lots of things to discuss today, so jumping in
20:01:44 <johnsom> #link https://etherpad.openstack.org/p/octavia-priority-reviews
20:01:48 <johnsom> Priority bug list
20:02:02 <johnsom> We have made great progress!
20:02:22 <johnsom> Below in the agenda I have pulled out "feature" items from the "bug" items.
20:02:34 <johnsom> We have a few more weeks on bugs, but we need to lock in our features this week
20:03:11 <johnsom> FYI Rocky RC1 is August 8th
20:03:37 <johnsom> Also of note, Stein PTL nominations are open
20:03:44 <johnsom> #link https://governance.openstack.org/election/
20:04:01 * johnsom looks around the room....
20:04:03 * nmagnezi points at johnsom
20:04:30 <johnsom> I encourage you all to experience this great opportunity!
20:04:57 <johnsom> Any other announcements today?
20:05:37 <johnsom> #topic Brief progress reports / bugs needing review
20:05:57 <johnsom> I have been busy with internal work, but did get a VRRP failover gate going.
20:06:12 <johnsom> #link https://review.openstack.org/584681
20:06:43 <johnsom> The amp stats tempest service client needs some more work, but is functional  (the VRRP gate uses it).
20:07:05 <johnsom> This was a checkbox I needed to mark as done, so....  It's done!  Grin
20:07:23 <johnsom> Any other status updates?
20:08:28 <xgerman_> in my privsep patch there was some change in iproute2 which broke it… usually I wait some time in the hope it fixes itself
20:09:03 <xgerman_> also threw up a patch earlier today…
20:09:08 <johnsom> Ok, yeah, Stein it is for that
20:09:51 <johnsom> Yeah, that is an interesting case with nova interactions
20:10:18 <johnsom> Ok, I'm guessing everyone is anxious for the next topic....
20:10:25 <johnsom> #topic Decide on Rocky features in flight
20:10:39 <johnsom> On the agenda I have pulled out features we have in flight.
20:10:44 <johnsom> #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Meeting_2018-07-25
20:10:57 <johnsom> We need to decide how we are going to proceed on these.
20:11:02 <johnsom> Options are:
20:11:12 <johnsom> 1. Push the feature out to Stein.
20:11:24 <johnsom> 2. File a FFE and buy a week.
20:11:44 <johnsom> 3. Merge these and fix any issues in the few weeks before RC1.
20:11:44 * xgerman_ gets his ballot ready
20:11:55 <johnsom> 4. Other idea....
20:12:12 <johnsom> Any question about those options?
20:12:45 <cgoncalves> 3.1. if cannot fix issues encountered after merge, revert
20:13:42 <cgoncalves> or, somehow, mark it as experimental (aka non-supported)
20:13:43 <johnsom> Well, yeah, but we hope nothing is in that shape.....
20:14:21 <johnsom> Ok, let's dig in
20:14:29 <johnsom> Neutron-LBaaS to Octavia migration tool
20:14:49 <johnsom> This is in good shape and has a passing gate.  I think we just need to merge it.
20:14:56 <johnsom> #link https://review.openstack.org/578942
20:15:02 <rm_work> sure, it's also just a tool
20:15:28 <rm_work> we can fix it up whenever, if people experience issues
20:15:35 <johnsom> FYI, this is the core review action item list..... grin
20:15:45 <rm_work> :P
20:15:57 <nmagnezi> johnsom, I'll give it my vote on Sunday (going back to office), but yeah if it's in a good shape probably we can fix some issues in followup patches
20:16:27 <johnsom> nmagnezi Sunday is too late. We have to have them merged basically tomorrow to make Rocky.
20:16:30 <xgerman_> FFE
20:17:06 <johnsom> You think we need a FFE for the migration tool?
20:17:23 <xgerman_> you can’t be too careful
20:17:39 <nmagnezi> If other ppl can vote for it, I guess no. But for my vote probably we will
20:17:55 <nmagnezi> I will tested it in both cases, just didn't had the time
20:17:58 <johnsom> So you aren't comfortable with the passing gate enough to merge that tool?
20:18:03 <xgerman_> migration is a “feature"
20:18:11 <johnsom> I'm slightly surprised on this one
20:19:15 <johnsom> xgerman_ What do you see as the risk to the project from that patch that we can't merge it and fix bugs in the next few weeks if someone finds one?
20:20:00 <johnsom> I'm trying to level set our process here, because frankly we can't just FFE all of these.
20:20:25 <xgerman_> I don’t FFE means I see risks - it’s just means a feature lands after the deadline
20:20:43 <johnsom> By exception
20:21:40 <johnsom> Really the only reason we should merge these is if we see it breaking existing functionality or the patch is structurally unsound and needs work beyond just bug fixes.
20:23:36 <xgerman_> ok, I gave you my +2
20:23:38 <johnsom> Do we disagree? Do we have other perspectives on this?
20:24:18 <nmagnezi> I can live with that, if I find anything I'll just file a story, or a patch (or both)
20:24:20 <xgerman_> so I should only merge things which break existing functionality — confused
20:24:52 <johnsom> Ha, sorry, Really the only reason we should NOT merge these is if we see it breaking existing functionality or the patch is structurally unsound and needs work beyond just bug fixes.
20:25:16 * johnsom looks a german for sending me someone asking about quotas during the meeting...  grin
20:25:19 <xgerman_> yeah, a lot of people lately are forgetting NOT
20:25:37 <xgerman_> johnsom: you wrote that code :-)
20:25:57 <johnsom> I did and he was setting neutron-lbaas quotas....
20:26:11 <xgerman_> 🤦
20:27:04 <johnsom> So I only saw one vote, do you all not want to vote on these?
20:27:20 <nmagnezi> I'll vote
20:27:29 <nmagnezi> Will read the gate logs after the meeting
20:27:50 <nmagnezi> If everything makes sense (and it should, since it passed), I'll +2
20:27:51 <nmagnezi> :-)
20:27:52 <johnsom> Ok, I just meant in the meeting, going through the list here
20:28:03 <xgerman_> sounds good
20:28:09 <johnsom> I wanted some community input, but I can make the call on these and ....
20:28:27 <nmagnezi> no worries, I think we all agree here.
20:28:32 <johnsom> Ok, moving on
20:28:33 <cgoncalves> +1 for merge from my side
20:28:41 <johnsom> Fix version discovery for the Octavia API
20:28:47 <johnsom> #link https://review.openstack.org/559460
20:29:18 <johnsom> I know we have had a lot of discussion on this and I have updated to try to address your comments.
20:29:52 <johnsom> My vote is #3 merge it. (I can't as I wrote it)
20:31:08 <johnsom> #link https://review.openstack.org/#/c/559460/9/api-ref/source/examples/versions-get-resp.json
20:31:12 <johnsom> That is the example output
20:32:31 <xgerman_> yeah, I wanted to see more constants used…
20:33:17 <cgoncalves> in https://review.openstack.org/#/c/559460/9/api-ref/source/index.rst you change supported api version to none but the output example shows v2.0
20:34:42 <johnsom> Yeah, that is a bit odd as we don't have a separate doc for 2.1, but we don't really want two versions tagged "CURRENT"
20:35:32 <johnsom> xgerman_ cgoncalves Want to open bugs for those two issues and we will resolve before RC1?
20:35:55 <cgoncalves> ok
20:36:10 <johnsom> by we, I mean I will resolve those
20:36:20 <xgerman_> sure
20:36:24 <johnsom> Unless you hate SB enough to just patch it
20:36:25 <johnsom> grin
20:36:38 <xgerman_> well, then I can’t vote ;-)
20:37:00 <johnsom> Ok, next up
20:37:05 <johnsom> Automatically set Barbican ACLs
20:37:11 <johnsom> #link https://review.openstack.org/552549
20:37:23 <johnsom> I super want to get this in and plan to test this afternoon.
20:37:33 <johnsom> However, one person reported an error with it
20:38:14 <cgoncalves> yeah, I haven't revisited this patch since my last patch set
20:38:26 <xgerman_> probably needs work or better docs ;-)
20:38:36 <cgoncalves> tomorrow I'll be traveling, so Friday the latest
20:38:54 <johnsom> cgoncalves So you think it needs more work and is non-functional?
20:39:03 <cgoncalves> this one could be a RFE as it needs more testing/validation from community
20:39:16 <johnsom> FFE?
20:39:25 <cgoncalves> I ran couple of sanity tests...
20:39:33 <cgoncalves> oops, yeah, FFE
20:39:41 <nmagnezi> +1 in FFE
20:39:43 <nmagnezi> on*
20:39:47 <xgerman_> +1
20:40:10 <johnsom> Ok. I will test this afternoon and give my +2 if it works or we can FFE it.
20:40:23 <johnsom> If we FFE, when can we commit to having it done?
20:40:23 <cgoncalves> nice, thank you!
20:40:33 <johnsom> By next Wednesday meeting?
20:40:54 <cgoncalves> I'd hope so
20:41:31 <johnsom> Ok, that will be the go/no-go date. If it isn't merged by the meeting next week we will hold it until Stein
20:41:49 <johnsom> Provider Driver support
20:42:12 <johnsom> These are a few small patches from the VMware team as they have developed their driver.
20:42:19 <johnsom> Plus two of mine.
20:43:23 <johnsom> Some of these already have one +2
20:44:15 <johnsom> I will volunteer to fix https://review.openstack.org/#/c/572303/ this afternoon if people can vote on it later tonight early tomorrow
20:44:28 <johnsom> It is just some minor docs issues
20:45:07 <johnsom> So we think we can merge the last part of provider drivers?  Any concerns?
20:45:26 * xgerman_ will keep his +2 ready
20:45:38 <johnsom> I get the feeling the vmware team has a working driver that is going into testing now
20:46:39 <johnsom> Darn missed one: Init provider drivers at app setup - https://review.openstack.org/#/c/585193
20:46:48 <johnsom> It's on the priority list
20:47:39 <johnsom> This one is pretty important actually: https://review.openstack.org/575807
20:47:59 <johnsom> As it impacts updates and batch updates, but in a way it's a bug fix too
20:48:33 <johnsom> Ok, so plan there is to merge this stuff?
20:48:35 <cgoncalves> https://review.openstack.org/#/c/585193/ has no unit tests. the change is trivial, though
20:48:56 <nmagnezi> Yup, this is why I voted for it
20:49:04 <johnsom> Yeah, it will give covered by the functional though
20:49:21 <cgoncalves> all right
20:49:30 <johnsom> At least it will get run by functional.....
20:49:45 <johnsom> Ok, moving on to the big one
20:49:51 <johnsom> UDP support
20:50:13 <johnsom> Last time I tested this it worked. it had some issues (patches have been posted).
20:50:42 <johnsom> I used the client too as I was too lazy to do API calls.
20:50:52 <cgoncalves> my vote is -0.5 as it's not passing on centos
20:51:26 <johnsom> cgoncalves Right, you were going to look into that and help Zhao.  Any ideas on it?
20:52:14 <cgoncalves> johnsom, sadly I only had time to run stack.sh so far
20:52:22 <johnsom> Ok
20:52:47 <johnsom> Personally I would like to see this go in as it's a great feature for the release and they have been working hard on it.
20:52:56 <johnsom> Do we think a week FFE will get it there?
20:53:00 <cgoncalves> I looked at the logs. amp agent is erroing for whatever reason. need to prevent HM from touching amps
20:53:58 <cgoncalves> I can't promise I'll have time. I'm okay-ish to +1 if it works on ubuntu
20:54:08 <nmagnezi> cgoncalves, there's an option to prevent reverts ; I also sometimes stop HM when I want it not to kill the amps but that's just hacky
20:54:47 <johnsom> Yeah, I set the revert config and stop the HM typically if I need to catch an amp
20:54:48 <cgoncalves> nmagnezi, yeah, I learned that the other day thanks to our fellas johnsom and rm_work :)
20:55:19 <johnsom> I would like to give this a week FFE and see if we can get it in.  Any objections?
20:55:28 <nmagnezi> Nop
20:55:35 <cgoncalves> +1 for FFE
20:55:41 <johnsom> Ok.
20:55:47 <nmagnezi> cgoncalves, could this be a matter of RPM versions of some sort?
20:55:53 <nmagnezi> In CentOS?
20:56:13 <johnsom> Can we get the client patch in? I can add the release note there.
20:56:34 <cgoncalves> nmagnezi, no idea. from CI logs amp was spitting some 'integer not valid' or whatever which was not enough as hint
20:56:43 <cgoncalves> johnsom, yes
20:56:53 <nmagnezi> cgoncalves, we're known for bleeding edge versioning for other RPMs there.. (a grin by johnsom in 3..2..)
20:57:10 <johnsom> Ok, Early this afternoon I will also fix that release note and vote.  I have been using the client and it seems stable.
20:57:29 <johnsom> lol
20:57:38 <johnsom> Dang almost out of time
20:57:44 <johnsom> Correct naming for quota resources
20:57:50 <johnsom> #link https://review.openstack.org/559672
20:58:01 <johnsom> Anyone feel strongly this needs in Rocky?
20:58:29 <nmagnezi> I tested this, my only comment is that I wanted a client side patch
20:58:32 <nmagnezi> But we can have it in
20:58:48 <nmagnezi> It works as intended IMO
20:58:56 <johnsom> Ok, can you vote on it then?
20:59:02 <nmagnezi> yes
20:59:06 <johnsom> I will review and vote today
20:59:10 <johnsom> Add usage admin resource
20:59:17 <johnsom> #link https://review.openstack.org/#/c/557548/
20:59:31 <johnsom> I think we can defer this to Stein and be ok.
20:59:40 <johnsom> Unless someone has a pressing need for it
21:00:10 <nmagnezi> I didn't get to this one so I'll support that.
21:00:10 <cgoncalves> I don't
21:00:18 <johnsom> I will take that as Stein.
21:00:23 <johnsom> Out of time....
21:00:24 <xgerman_> yeah, I also think usage needs more throught
21:00:30 <johnsom> #endmeeting