14:04:09 <fifieldt> #startmeeting ops tags
14:04:10 <openstack> Meeting started Thu Mar 31 14:04:09 2016 UTC and is due to finish in 60 minutes.  The chair is fifieldt. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:04:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:04:14 <openstack> The meeting name has been set to 'ops_tags'
14:04:27 <shamail> Awesome, thanks.
14:04:33 <fifieldt> Thanks for arranging the agenda
14:04:37 <fifieldt> looks like 3 items today
14:04:42 <shamail> np
14:04:42 <fifieldt> 1) Review open changes
14:04:42 <fifieldt> 2) Discuss tag update plans for Mitaka
14:04:43 <fifieldt> 3) Planning for ops-tags meeting at the Austin Summit
14:04:49 <shamail> +1
14:05:05 <fifieldt> #topic Review Open Changes
14:05:10 <fifieldt> #link https://review.openstack.org/#/q/status:open+project:openstack/ops-tags-team,n,z
14:05:18 <jproulx> proj_since_descriptor: Thierry says this is already on the navigator (derived from repo info).  Is there a reason to duplicate (or replace) as an ops-tag?
14:05:28 <shamail> Yes I agree.
14:05:50 <jproulx> that there's no reason to duplicate?
14:05:54 <shamail> When I had proposed it, I wasn't aware it was on navigator.  I can abandon the request.
14:06:14 <jproulx> I didn't realize that either, OK one down :)
14:06:18 <fifieldt> cool
14:06:21 <fifieldt> less maint effort is good :)
14:06:29 <shamail> The only question I have is whether it makes to show this for other projects beyond navigator coverage?
14:06:39 <shamail> Or if that data is accessible somewhere too?
14:06:50 <fifieldt> the data is coming from somewhere
14:06:55 <fifieldt> I'm just not sure where
14:07:08 <fifieldt> the navigator is populated from public sources
14:07:12 <shamail> I am fine with abandoning because I don't think we need to maintain, but I would like to find the source eventually :)
14:07:29 <jproulx> Theirry said it was auto derrived but not recorded anywhere execpt the navigator
14:07:42 <fifieldt> ah
14:08:02 <fifieldt> I think it's under the release info
14:08:09 <jproulx> If we find we need ot start duplicating that code, then making it a tag taht used that code makes sense
14:08:14 <fifieldt> http://releases.openstack.org/
14:08:24 <fifieldt> that has the versions of projects released
14:08:32 <fifieldt> so possible to derive age from that
14:09:12 <fifieldt> all good?
14:09:30 <shamail> I can't copy his comment from the review but he said they use git history and git tags.
14:09:48 <fifieldt> aye
14:09:50 <fifieldt> same stuff then
14:10:01 <jproulx> #agreed proj_since_descriptor: not needed at this time as it duplicates info already in teh navigator
14:10:34 <fifieldt> next up, an easy one: Add Ops tag template files
14:10:40 <fifieldt> basically just needs someone to finish it
14:10:49 <jproulx> odd this one has stalled, can we just add in fifieldt's suggestion and publish it?
14:11:20 <shamail> +1
14:11:25 <shamail> I can take the action
14:12:01 <fifieldt> #action shamail to update template patch to incorporate comments and publish
14:12:04 <fifieldt> ok, next up
14:12:05 <shamail> Thanks
14:12:09 <fifieldt> Add ops:ha tag
14:12:26 <fifieldt> This has a few comments that appeared to reach a consensus on an update
14:12:27 <jproulx> that one is probably in need of face to face discussion in Austin
14:12:35 <fifieldt> agreedc
14:12:39 <shamail> jproulx: +1
14:12:46 <fifieldt> would be nice to get those updates made prior to the discussion if possible
14:12:53 <fifieldt> so we present the latest thinking
14:13:40 <fifieldt> any takers?
14:13:43 <shamail> Sorry was reviewing the comments
14:13:56 <shamail> I think it makes sense and I can also check status of HA guide coverage
14:14:03 <shamail> I'll take it as well
14:14:08 <fifieldt> thank you!
14:14:23 <fifieldt> #action shamail to update ops:ha tag with latest comments in prep for austin discussion
14:14:26 <fifieldt> final one
14:14:27 <shamail> Welcome!
14:14:27 <jproulx> shamail needs a prize...
14:14:28 <fifieldt> containerizable
14:14:36 <fifieldt> this one was submitted incomplete
14:14:52 <shamail> jproulx: :-)
14:15:11 <jproulx> criteria is a bit weak "someone somewhere claims to have done", perhaps docs or it didn't happen? This seems to be noted in comments as well
14:15:24 <fifieldt> probably not fixable without further discussion]
14:15:27 <fifieldt> ?
14:15:34 <shamail> We had discussed this one previously and we debated the intent/usefulness
14:15:41 <shamail> fifieldt: +1
14:15:47 <fifieldt> leave it and see how it goes in austin?
14:15:59 <jproulx> fifieldt +1
14:15:59 <shamail> Yes
14:16:09 <fifieldt> #agreed - leave containerizable to austin
14:16:13 <fifieldt> ok, next agenda item
14:16:23 <fifieldt> #topic Discuss tag update plans for Mitaka
14:16:47 <shamail> I wanted to start automating some of the tags but as I looked around, I couldn't identify sources for some.
14:16:51 <fifieldt> so, we have 4 tags to update
14:16:59 <shamail> Has install guide is self explanatory
14:17:09 <fifieldt> ops-docs-install-guide, ops-packaged, ops-production-use, ops-sdk-support
14:17:14 <shamail> But is the SDK data published somewhere or the packaging data?
14:17:31 <fifieldt> SDK support comes from https://docs.google.com/spreadsheets/u/1/d/1nRSzWo4I-isHGL67ROuqdj5My9rrSsMhlf5eSqSntn0/edit
14:17:54 <fifieldt> requires a quick run through the "alive" SDKs to see if any new support added
14:18:03 <fifieldt> that was last done in December, so probably not too out of date
14:18:05 <annegentle> fifieldt: is that spreadsheet sorta derived from the firstapp?
14:18:05 <shamail> That is harder to automate than I expected, was anticipating a doc or repo source.
14:18:15 <jproulx> packaging I think we need to search some (yet to be?) defined repos
14:18:22 <annegentle> shamail: yeah was thinking of something we could tie it to
14:18:40 <shamail> annegentle: +1
14:18:43 <fifieldt> annegentle, naw
14:18:50 <fifieldt> basically I've been digging into the SDKs themselves
14:19:03 <annegentle> fifieldt: ah yeah it's wider net casting which is also good
14:19:16 <fifieldt> key feature - noting "dead" SDKs :)
14:19:23 <shamail> #link https://docs.google.com/spreadsheets/u/1/d/1nRSzWo4I-isHGL67ROuqdj5My9rrSsMhlf5eSqSntn0/htmlview
14:19:38 <annegentle> yeah wonder if it could be automated from github api then if that's the main piece of data
14:19:39 <shamail> Just to record it when I go looking for it again.
14:19:48 <fifieldt> Craig S at Intel has been contacting the SDK-builder communities of late
14:20:03 <fifieldt> he may have good enough contacts now that we can just send one email per SDK saying "any updates?"
14:20:31 <fifieldt> but I expect few over the past months
14:20:58 <annegentle> ok
14:21:00 <shamail> I can investigate the repos for the SDKs to see if there are any clues in commits
14:21:16 <fifieldt> AppEcoWG might be able to help, too
14:21:23 <fifieldt> they have a few more peeps than us
14:21:26 <shamail> Craig is a good starting point though to update/refresh for the new release
14:21:40 <shamail> fifieldt: +1, I'll contact Flanders and John
14:21:46 <fifieldt> cheers
14:22:01 <fifieldt> #action shamail to ask flanders and john for help on updating sdk-support tag
14:22:01 <shamail> The packaging one is interesting...
14:22:11 <fifieldt> always :)
14:22:26 <fifieldt> basically looking for new projects and new major bugs
14:22:32 <fifieldt> a key source of the latter is the docs team :)
14:22:38 <fifieldt> from their install guide testing
14:22:55 <shamail> Ah.
14:23:04 <jproulx> seems we should be able to search some well know repos (though bugs are trickier to identify than does it exist)
14:23:16 <shamail> yeah
14:23:20 <fifieldt> with that one, I'd start by copy/pasting the previous releases json and submitting it as-is for mitaka
14:23:41 <fifieldt> then taking a look at what new and upcoming projects are listed in the user survey
14:23:52 <fifieldt> see if they're freshly packaged for mitaka in a couple distros
14:23:58 <shamail> the other thing I wanted to mention is that we only tag it as "good" and don't provide specifics on which distributions have packages.  Any reason for that?
14:24:00 <fifieldt> the old ones should not really go away
14:24:12 <fifieldt> shamail, maintenance
14:24:22 <fifieldt> the effort to create an exhaustive matrix is significant
14:24:36 <jproulx> don't have the tag in  front of me, do you know if it defines where we look for packages?  probably should so people can point us at new places if the pop up
14:24:50 <fifieldt> and, if you want to know if a package exists for your distro, apt-cache search :)
14:24:51 <shamail> Got it.
14:25:06 <fifieldt> it doesn't define where we look
14:25:06 <shamail> thanks fifieldt
14:25:10 <fifieldt> it's a general quality indicator
14:25:21 <fifieldt> that also provides some info about any traps
14:25:42 <fifieldt> key things we're looking for are mostly to do with new projects
14:25:48 <fifieldt> or badly botched packaging for older ones
14:25:57 <jproulx> fifieldt +1
14:26:13 <fifieldt> ops-production-use is directly from the user survey
14:26:14 <shamail> Alright, so I wanted to refresh that one then I would need to talk to the docs team to find out the condition and look for new things in the packaging repos?
14:26:44 <fifieldt> yeah, matt kassawara is a good individual, but basically the docs install guide specialty team
14:26:56 <shamail> Got it.  Thanks.
14:26:59 <annegentle> fifieldt: which is just Matt right now :)
14:27:00 <fifieldt> and https://wiki.openstack.org/wiki/Documentation/MitakaDocTesting
14:27:07 <shamail> I'll try to get a head start on this before Austin
14:27:13 <shamail> So we can review in person
14:27:14 <annegentle> shamail: he's Sam-i-am on irc
14:27:27 <shamail> Thanks annegentle
14:27:35 <fifieldt> to kill two birds with one stone, docs team should also be contacted about the install guide tag
14:27:45 <fifieldt> see if they have any new projects
14:27:54 <fifieldt> any worries or concerns about existing ones
14:28:07 <shamail> https://wiki.openstack.org/wiki/Documentation/MitakaDocTesting
14:28:17 <annegentle> fifieldt: should that be discoverable without having to contact a person?
14:28:25 <fifieldt> of course :)
14:28:31 <jproulx> I think it *should* be
14:28:33 <annegentle> okay
14:28:35 <annegentle> yeah
14:28:36 <fifieldt> just cuz politeness
14:28:36 <fifieldt> :)
14:28:52 <shamail> Agree
14:29:05 <annegentle> I'd love to automate docs completeness somehow, ideas?
14:29:15 <annegentle> filenaming is pretty consistent
14:29:36 <annegentle> could scrape the docs repos ever release?
14:29:56 <annegentle> every
14:30:00 <fifieldt> and the guides in individual project repos ? :)
14:30:09 <fifieldt> where's that xkcd comic about automation
14:30:10 <fifieldt> :D
14:30:13 <shamail> Is there a classification matrix anywhere today that shows the doc types per project and branch?
14:30:21 <annegentle> yes fifieldt it wouldn't be too bad since we're consistent in dir names and tox jobs
14:30:35 <annegentle> shamail: thinking it wouldn't be hard to make, but there isn't one now
14:30:55 * fifieldt is all for automation
14:31:07 <annegentle> only headache is sometimes there's a folder but there's garbage docs
14:31:18 <fifieldt> in the case where automation doesn't exist, just cp liberty/install-guide.json mitaka/install-guide.json
14:31:20 <shamail> annegentle: +1
14:31:22 <annegentle> but still if it has build jobs, someone somewhere thought it was ready to publish? I am asking not telling :)
14:31:24 <fifieldt> very few changes between releases
14:31:50 <shamail> Yeah, fifieldt I was going to wait until the user survey report is released and snag the data.
14:32:04 <annegentle> I also envision automating "docs drift" where you measure the time between code commits and docs commits difference
14:32:06 <shamail> For install guide, I'll take your approach.
14:32:21 <fifieldt> oh, I like that one annegentle :)
14:32:22 <annegentle> do you mind posting a link to the repo here?
14:32:30 <shamail> that would be interesting annegentle
14:32:37 <fifieldt> this one annegentle ? https://github.com/openstack/ops-tags-team/blob/master/liberty/ops-docs-install-guide.json
14:32:43 <annegentle> fifieldt: yes thank you
14:32:47 <annegentle> saves me a lookup :)
14:33:03 <fifieldt> shamail - will have to see when the survey is released
14:33:10 <fifieldt> if it isn't released before release day
14:33:13 <fifieldt> I have pre-release data we can use
14:33:16 <shamail> So far, I am struggling to find where we could automate validation for ops tags.  Sadly.
14:33:39 <shamail> fifieldt: +1
14:33:52 <shamail> I don't think it will be... The plan is in next couple of weeks
14:34:05 <shamail> But release week is stretch
14:34:06 <fifieldt> ok, I'll take the action on production-use then
14:34:08 <annegentle> shamail: what mechanisms are in place now? Basically "take the json from the ops-tags-team repo and publish to matrix?"
14:34:13 <fifieldt> #action fifieldt to udpate project-use tag
14:34:21 <fifieldt> that's how it happens, annegentle :)
14:34:24 <shamail> Yes...
14:34:26 <annegentle> fifieldt: ok thought so.
14:34:32 <fifieldt> one thing I would love is some json validation in the gate :)
14:34:38 <shamail> I was also referring to updating the JSON on the tag itself
14:34:46 <annegentle> fifieldt: oh so like writing a schema?
14:34:51 <fifieldt> naw
14:34:55 <annegentle> fifieldt: or "this is json yes it is"
14:34:56 <fifieldt> just plain json test
14:34:59 <fifieldt> yes :)
14:35:02 <shamail> In these conversations, we identified very little things that could be scraped (except for docs)
14:35:03 <fifieldt> at the moment lacking even that
14:35:05 <annegentle> yeah we have that
14:35:12 <fifieldt> it totally exists
14:35:16 <annegentle> I mean, we have that gate test, just a matter of patching infra
14:35:17 <fifieldt> just need to write the jobs
14:35:17 <annegentle> yeah
14:35:32 * fifieldt is the worst offender of broken json
14:35:44 <annegentle> prolly me :)
14:35:48 <fifieldt> ok. is everyone happy with the updates of existing tags for mitaka?
14:35:52 <shamail> 😂
14:35:56 <jproulx> +1
14:35:58 <shamail> fifieldt: +1
14:36:03 <fifieldt> ok, goodo
14:36:05 <fifieldt> moving on
14:36:10 <fifieldt> #topic planning for Austin
14:36:23 <fifieldt> So,
14:36:32 <fifieldt> ops-tags-team was submitted as a WG for Monday's ops summit stuff
14:36:49 <shamail> Yeah, saw that.
14:36:50 <fifieldt> is a 40 minute slot enough
14:36:50 <fifieldt> ?
14:37:09 <shamail> Last time we went over but it was a meeting worth having popcorn for!
14:37:28 <jproulx> I suspect a longer slot would be better
14:37:42 <fifieldt> ok
14:37:48 <shamail> jproulx: +1, we usually have more participants
14:38:04 <fifieldt> I will wrangle a schedule so we get a good slot that both of you are available for
14:38:06 <jproulx> We've got containerizable and ha on the table already each of thos is probably good for 30min
14:38:08 <fifieldt> this is harder than normal
14:38:21 <fifieldt> since we're doing working sessions and general sessions in parallel this time
14:38:30 <shamail> thanks
14:39:08 <shamail> I can email you my schedule for Monday later this evening
14:39:18 <fifieldt> awesome
14:39:26 <fifieldt> so, schedule for the double session then?
14:39:32 <shamail> +1
14:39:34 <jproulx> +1
14:39:58 * jproulx should probably look at his summit schedule real soon now
14:40:13 <fifieldt> as in, what's our agenda?
14:40:21 <shamail> I'll start populating ideas on the etherpad
14:40:24 <fifieldt> as a start, I think we need some brainstorming
14:40:36 <fifieldt> how about https://etherpad.openstack.org/p/AUS-ops-tags
14:41:03 <shamail> Yep, that's the one
14:41:07 <fifieldt> excellent
14:41:07 <jproulx> well we have the two hot tags we've carried forward
14:41:33 <shamail> Yeah.
14:41:41 <jproulx> and we should think & work hard on automating what we've got
14:41:50 <shamail> "Revising automation"
14:41:51 <fifieldt> kewl
14:42:09 <jproulx> so maybe 1st 1/2 tag review second 1/2 tag automation roughly
14:42:17 <shamail> expand/discuss maturity components
14:42:23 <shamail> And missing tags based on definition
14:42:34 <fifieldt> need some time to try and come up with new tagsd
14:42:43 <fifieldt> leech the newcomer's ideas
14:42:48 <shamail> +1
14:43:19 <jproulx> +1
14:43:38 <shamail> jproulx: sounds good... 1st half could be discussion on tags content and 2nd half could be focused on procedures and processes
14:44:02 <fifieldt> +1
14:44:06 <jproulx> Yes that's a better description
14:44:09 <shamail> Documenting sources for all the tags (maybe in the tag itself) might be a useful thing to consider in 2nd half as well
14:44:19 <fifieldt> also, at some point, we need to define how core review rights on the repo works
14:44:27 <fifieldt> at the moment it's just the UC
14:44:32 <fifieldt> and I think shamail should be included :)
14:44:37 <jproulx> +1
14:44:40 <fifieldt> so someone just needs to make a decision about how we do that
14:44:45 <shamail> :-) definitely an in person topic
14:45:05 <fifieldt> aye, aye
14:45:23 <fifieldt> ok, so we have the skeleton of the agenda for austin
14:45:26 <fifieldt> anything else?
14:45:48 <shamail> That's all from me!
14:45:55 <jproulx> Sounds good
14:46:02 <fifieldt> ok
14:46:06 <fifieldt> #topic anything else?
14:46:13 <fifieldt> 14 minutes left in this slot
14:46:14 <shamail> I'll start on action items and report out (if you don't see changes coming in)
14:46:19 <fifieldt> want it back?
14:46:20 <jproulx> meetings are so fast and efficient with 4 participants :)
14:46:28 <fifieldt> big props to guest star annegentle
14:46:29 <shamail> haha
14:46:33 <annegentle> haha
14:46:34 <shamail> ++
14:46:39 <annegentle> docs drift. it's a thing
14:46:47 <shamail> Thanks Ann!
14:46:50 <shamail> Anne*
14:47:00 <fifieldt> ok then
14:47:02 <jproulx> thanks all
14:47:03 <shamail> My phone... :(
14:47:13 <fifieldt> #endmeeting