00:05:21 <Kylie_> #startmeeting azure_working_group
00:05:21 <zodbot> Meeting started Thu Feb  8 00:05:21 2018 UTC.  The chair is Kylie_. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:05:21 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
00:05:21 <zodbot> The meeting name has been set to 'azure_working_group'
00:05:33 <nitzmahone> @yuwei - the LB changes do not appear to be idempotent at first glance
00:05:38 <Kylie_> let us start.
00:05:40 <yuwei> no
00:05:47 <Kylie_> #chair nitzmahone
00:05:47 <zodbot> Current chairs: Kylie_ nitzmahone
00:05:55 <Kylie_> #chair yuwei
00:05:55 <zodbot> Current chairs: Kylie_ nitzmahone yuwei
00:06:06 <Kylie_> #chair zikalino
00:06:06 <zodbot> Current chairs: Kylie_ nitzmahone yuwei zikalino
00:06:20 <Kylie_> #topic checklist for 2.5
00:06:47 <Kylie_> Any question about LB change?
00:07:43 <yuwei> nitzmahone: i add a test to test the delete idempotent
00:08:14 <nitzmahone> Right, but it looks like it doesn't properly check for update- will always call create_or_update
00:08:19 <yuwei> And the change is to change the parameters can accept list
00:08:27 <nitzmahone> *all* Ansible modules should be idempotent
00:09:24 <yuwei> No, i obey the previous one not support update currently. I think this should be another task to support update
00:09:39 <nitzmahone> We can take it as-is (needs some minor doc/message fixes, but I'll take care of that), but yeah, it's a bug that it's not idempotent.
00:10:03 <nitzmahone> IIRC Thomas was supposed to fix that in 2.4 but it never happened
00:10:18 <zikalino> well, as this particular pr is a bug fix, i think probably we should merge it and create another issue for idempotence
00:10:32 <nitzmahone> yep, that's what I said too :)
00:10:52 <zikalino> if there's an issue than for sure Kylie_ will push as to close it asap ;-)
00:10:55 <yuwei> Yes we are sharing the same point
00:11:21 <nitzmahone> I'll file an issue for that when I merge then
00:11:40 <Kylie_> Great.
00:11:46 <zikalino> and that leads to another thing, we were trying to figure out when we can merge fixes ourselves, and seems like it never works.
00:11:46 <yuwei> thx
00:11:59 <Kylie_> Then last piece for 2.5 is API versioning.
00:12:04 <Kylie_> When is code freeze?
00:12:14 <nitzmahone> Feature freeze is today
00:12:25 <Kylie_> bug fix freeze?
00:12:36 <nitzmahone> There isn't one
00:12:49 <Kylie_> For API versioning, it is bug fix. Right?
00:12:54 <nitzmahone> correct
00:13:12 <nitzmahone> beta 1 is tomorrow, beta 2 is the following week, then we'll cut rc1 the week after
00:13:14 <Kylie_> @Yuwei, we thought we missed listing one task in our scrum. It should be API versioning for you.
00:13:39 <yuwei> oh yes
00:13:48 <nitzmahone> There's still a couple outstanding sanity test failures for me to fix on my versioning PR
00:13:52 <nitzmahone> I'll do that today as well
00:14:16 <yuwei> The PR fails for a sanity test，but I have no idea of that message yesterday
00:14:23 <nitzmahone> I know what needs to happen
00:14:26 <Kylie_> Great. Thank you Matt. And Yuwei, please follow up the rest. Hope customers will be not bothered by API disconnection after 2.5.
00:14:51 <yuwei> great
00:14:59 <nitzmahone> We'll probably need to bump the minimum package versions on all SDK packages for 2.5
00:15:07 <nitzmahone> (maybe just to whatever's current)
00:15:40 <nitzmahone> Laurent also said to request update/release on any SDK package that doesn't include `models` on the operation class, so we should be able to treat them all the same way.
00:15:52 <nitzmahone> He's already done containerinstance
00:16:03 <nitzmahone> (and most of the more commonly used ones)
00:17:18 <nitzmahone> If we want to backport these versioning changes to 2.4, we'll also have to bump the versions there.
00:17:36 <Kylie_> I think 2.4 is stretch goal.
00:17:52 <Kylie_> #topic process for PR merge
00:18:27 <Kylie_> Take LB PR as an example, 2 said "ship it" but "merge" does not happen.
00:18:34 <nitzmahone> Yeah- the trick is that 2.4 is supported longer than our previous releases, so people will be running into more and more problems with it over time as SDKs update. Hopefully 2.5 changes will minimize that issue.
00:19:26 <zikalino> yes, when i do bot_status, it's not quite clear, our shipits are counted as i believe "community_shipit"
00:20:13 <nitzmahone> It should have merged, but I'm not an expert on the bot. I'll mark it bot_broken and see if the bot expert can figure out why.
00:20:56 <nitzmahone> We need to be careful about docs and error messages on things like this.
00:21:04 <zikalino> well it usually is in "waiting_on: ansible" or "waiting_on: something else"
00:21:09 <zikalino> let me find example
00:22:30 <nitzmahone> Yuwei's English is much better than my Chinese, but docs and error message changes should probably be reviewed by a native English speaker before being merged.
00:22:57 <nitzmahone> One of our guys was trying to test this PR last week and was really confused by the deprecation warning.
00:23:19 <nitzmahone> (he was using the old non-list args)
00:24:00 <nitzmahone> I had to look at the code to see what the problem was
00:25:01 <Kylie_> Thank you. If that is bot_broken, it is a bug. If the process does not permit two Ansible/Azure maintainers approve one bug fix, it is the process we need to discuss.
00:25:05 <nitzmahone> Regardless, I don't think that's what kept it from getting merged- will invoke the bot expert
00:25:27 <Kylie_> message review by native speaker is a good suggestion. We also will see whether Harold or Thomas could help.
00:25:42 <Kylie_> Thank you Matt.
00:25:59 <Kylie_> That all topics I have today. Any other open.
00:26:18 <nitzmahone> Nothing from me at the moment
00:26:24 <Kylie_> And Matt, a heads-up that 2/14-2/22 are our Chinese New Year holiday.
00:26:26 <zikalino> i think perhaps we could collect more data, right now when i had problem ans asked somebody, they just merged it, and bot_status disappeared ...
00:26:29 <zikalino> i have one small topic
00:26:53 <nitzmahone> @Kylie_ OK, so probably no meeting the next two weeks?
00:27:13 <nitzmahone> @zikalino go for it
00:27:17 <zikalino> i found this group on linked in: https://www.linkedin.com/groups/4648083/profile
00:27:18 <Kylie_> Yes, people will be on vacation for the following two weeks.
00:27:22 <zikalino> it seems to be abandoned
00:27:44 <zikalino> there's one manager -- Todd Barr from RedHat
00:28:02 <zikalino> nitzmahone could you ping him about it?
00:28:14 <nitzmahone> What do you want to do with it?
00:28:57 <zikalino> just ask him if he could accept new members, or if he would like to add another admin if he doesnt' have time to take care of it.
00:29:14 <zikalino> there's around 3000 people who want to join but are not accepted right now.
00:29:15 <nitzmahone> I don't think we want to use LinkedIn as a support channel
00:29:42 <zikalino> right, so maybe this group should be removed.
00:29:57 <nitzmahone> Michael (Ansible creator) is the only one who can kill it, and I don't have a way to talk to him
00:30:18 <zikalino> you mean Michael DeHaan?
00:30:31 <nitzmahone> Yes- he's been blocking Ansible people on Twitter lately
00:31:05 <zikalino> i actually contacted him, and he said he said he no longer works for Ansible, and I shoudl ask sbd from RedHat
00:31:12 <nitzmahone> (he's the one who created the group, and only the group owner can destroy it)
00:31:30 <zikalino> I didn't realise he was the Ansible creator...
00:31:33 <nitzmahone> Group Managers can let people in and do other admin stuff
00:31:53 <nitzmahone> But I think only the Owner can destroy it
00:32:20 <nitzmahone> So there it sits ;)
00:33:06 <zikalino> ok, seems like a gridlock :-|
00:33:10 <Kylie_> Ok. Seems we cannot have a solution here.
00:33:14 <nitzmahone> unfortunately so
00:33:37 <Kylie_> Ok, then release everyone back to issue/backlog:)
00:33:42 <Kylie_> Thank you all for today meeting.
00:33:45 <Kylie_> #endmeeting