19:00:55 <SotK> #startmeeting storyboard
19:00:56 <openstack> Meeting started Wed Jan 11 19:00:55 2017 UTC and is due to finish in 60 minutes.  The chair is SotK. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:57 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:00 <openstack> The meeting name has been set to 'storyboard'
19:01:00 <rhochmuth> o/
19:01:13 <diablo_rojo> Hello :)
19:01:21 <diablo_rojo> Thanks for coming rhochmuth :)
19:01:35 <SotK> #link https://wiki.openstack.org/wiki/Meetings/StoryBoard Agenda (that is semi-updated)
19:01:38 <rhochmuth> thanks for inviting me
19:01:44 <SotK> welcome folks!
19:02:07 <SotK> I don't think we have any announcements or urgent items today
19:02:14 <SotK> #topic In Progress Work
19:04:01 <zara_the_lemur__> I've finally sent a patch to get rid of our placeholder operator deployment docs
19:04:01 * SotK hasn't managed to get much done in this last week
19:04:25 <diablo_rojo> zara_the_lemur__, nice :)
19:04:33 <zara_the_lemur__> I'd rather a really nice operator guide, but the current thing was worse than nothing
19:04:45 <zara_the_lemur__> since it made a user think there was a guide, then they'd click it and get 'guide todo'
19:05:21 <zara_the_lemur__> story for that more generally is: https://storyboard.openstack.org/#!/story/2000535
19:05:52 <zara_the_lemur__> patch is:
19:05:55 <zara_the_lemur__> #link https://review.openstack.org/#/c/418122/
19:06:21 * SotK is still vaguely conflicted about it
19:06:28 <zara_the_lemur__> fair enough
19:06:37 <diablo_rojo> SotK, conflicted why?
19:07:06 <zara_the_lemur__> (I partly sent it to prompt discussion on it and maybe someone would take on the task of making a better ops guide)
19:07:07 <SotK> I worry that we will remove it and then never get round to putting a proper one there
19:07:38 <SotK> but then the counter argument is that we are equally likely to have the misleading stub there and never get round to putting a proper one there
19:07:51 <zara_the_lemur__> yeah, that was my reason for leaving the stub up so long
19:07:53 <diablo_rojo> Ha ha fair enough
19:07:54 <zara_the_lemur__> and I didn't get round to it
19:08:36 <zara_the_lemur__> is anyone here atm able or interested to make a better ops guide?
19:08:45 <zara_the_lemur__> *able to or interested in
19:08:56 * zara_the_lemur__ gives up on fixing that sentence
19:09:01 <persia> Folk who are current operators would be the most exciting of volunteers
19:09:02 <diablo_rojo> zara_the_lemur__, maybe we could work on that in conjunction with a summit talk for storyboard?
19:09:18 <diablo_rojo> Unless someone else wants it :)
19:09:41 <zara_the_lemur__> diablo_rojo: my trouble atm is setting up a good environment for the deployment
19:10:18 <diablo_rojo> zara_the_lemur__, Ah, fair enough.
19:11:53 * SotK is not seeing many volunteers, so I guess I will think more and become unconflicted :)
19:11:59 <zara_the_lemur__> heh
19:12:03 <zara_the_lemur__> #link https://review.openstack.org/#/c/416070/
19:12:06 <zara_the_lemur__> also awaits reviews
19:12:29 * SotK will will review after this meeting
19:12:33 <zara_the_lemur__> either of the 'fix the logic a bit first' kind or the 'merge this' kind, not sure if people are waiting for me there
19:12:52 <diablo_rojo> I will review too. I need to get in on all that fun :)
19:13:22 <SotK> Anything else in progress that needs attention brought to it?
19:14:02 <zara_the_lemur__> my misc install instructions fixes are still awaiting further +1s
19:14:13 <zara_the_lemur__> I apparently sent a bunch of install instruction things this week
19:14:31 * SotK cannot offer extra +1s
19:14:40 <zara_the_lemur__> #link https://review.openstack.org/#/c/415942/
19:14:53 <zara_the_lemur__> #link https://review.openstack.org/#/c/415996/
19:15:44 <zara_the_lemur__> oh, and AJaeger's docs change:
19:15:47 <zara_the_lemur__> #link https://review.openstack.org/#/c/415493/
19:15:59 <zara_the_lemur__> that's one I've reviewed, awaiting more eyes
19:16:03 * diablo_rojo opened a bunch of tabs
19:16:07 <zara_the_lemur__> brb plugging in computer
19:16:32 <zara_the_lemur__> nothing else to note there
19:16:50 * SotK merges AJaeger's patch
19:17:10 <SotK> ok, lets move on
19:17:17 <SotK> #topic Migration scripts
19:17:56 <SotK> I believe that it is a good time to see if our LP migration scripts still work?
19:18:02 <diablo_rojo> So, do we know the last time they were used for sure?
19:18:40 <SotK> I think they were used when craige was investigating phabricator stuff
19:19:03 <SotK> so probably late 2015/early 2016
19:19:45 <diablo_rojo> Not so bad then
19:20:19 <diablo_rojo> So, we need to test those to see what all breaks so we can hopefully get Monasca rolling on moving over.
19:20:41 <zara_the_lemur__> if it's any help, they were used to import infra stories to storyboard before that; some unusual characters weren't converted properly
19:20:58 <SotK> that is help, we need to fix that bug
19:21:03 <diablo_rojo> unusual characters?
19:21:13 <zara_the_lemur__> it's something we should test on a small scale first, hang on, will find the story
19:21:39 * zara_the_lemur__ remembers it because she went through removing the oddness so GETs worked properly
19:21:46 <SotK> terminal colour codes and stuff iirc
19:22:25 * zara_the_lemur__ finds https://storyboard.openstack.org/#!/story/292
19:22:45 <diablo_rojo> Interesting. Okay well if its just a few things to get cleaned up that shouldn't be bad. Does anyone have something we could test it on?
19:23:12 <zara_the_lemur__> I would imagine storyboard-dev.openstack.org would be a good place?
19:23:21 <SotK> I suggest we try importing a real LP project into storyboard-dev
19:23:37 <persia> There were a few classes of characters that were not converted properly.  These at least included A) ANSI control codes, B) some multibyte characters (but not all multibyte characters, strangely), and C) some cases where different sorts of markup seemed to be overly escaped.
19:25:08 <diablo_rojo> SotK, so we could hypothetically just try it with Monasca to storyboard-dev and so long as it all looks happy then we can move forward with the real migration?
19:25:13 <persia> Specifically related to Monasca: what is the post-migration plan for things like https://bugs.launchpad.net/monasca/+bug/1368661 ?
19:25:13 <openstack> Launchpad bug 1368661 in Glance Client "Unit tests sometimes fail because of stale pyc files" [Undecided,In progress] - Assigned to Shu Muto (shu-mutou)
19:25:30 <SotK> it strikes me we will also need to decide how to encode the priorities opf of tasks from LP
19:26:14 <SotK> diablo_rojo: yes, I reckon so
19:26:20 <persia> rhochmuth: How do you think the monasca team will want to encode priorities in the future?
19:26:33 <diablo_rojo> persia, good question :)
19:26:35 <rhochmuth> i'm not sure
19:27:00 <rhochmuth> are you referring to high, undecided, …
19:27:36 <rhochmuth> for example, the priority to resolve a bug or implment a feature
19:27:53 <persia> There are two different common ways of expressing priority in Storyboard, used today.  One is to use tags (e.g. "high-priority").  The other is to use worklists (gather the most important into a worklist, with developers expected to pull their next activity from that worklist).
19:28:23 <persia> Yes.  LP has a "Priority" field, with Critical, High, Medium, Low, Undecided.  Storyboard doesn't have that field.
19:28:30 <rhochmuth> i think the tags would be the best approach
19:28:53 <rhochmuth> for bugs
19:29:20 <rhochmuth> the worklists i see being more useful for tasks associated with a new feature or blueprint
19:29:47 <rhochmuth> So, I think the Priority field would be useful
19:29:57 <rhochmuth> without that we can't order bugs for resolution
19:29:59 <persia> Specifically for the migration, do you think it makes sense to create tags as part of migration?
19:30:20 <persia> Or do you think it makes sense to populate a worklist in Priority order.
19:30:21 <rhochmuth> i see, i think tags make sense
19:30:36 <fungi> also, with automatic worklists based on story tags, you can sort of emulate the old bug importance implementation in lp
19:30:57 <persia> Ah, good point
19:31:00 <rhochmuth> well, now i might want to change my mind since you put it that way
19:31:34 <fungi> you can have an automatic worklist for high-importance bugs, and tag them with monasca-bug-high or something
19:31:34 <rhochmuth> so, if you can create worklists by based on tags, that sounds like what we woudl want
19:31:58 <rhochmuth> sounds like it would work well
19:32:14 <persia> rhochmuth: Note that automatic worklists aren't editable directly, so if you go that way, and you want to change a priority, you'd go to the individual story, and change the tag.
19:32:25 * SotK returns and complains about internet trouble
19:32:27 <fungi> and then their tasks would automatically appear in some "monasca's high importance bugs" worklist
19:32:50 * SotK would also recommend some kind of worklist approach
19:33:00 <persia> But I think that's the easiest to implement in a migration script, especially tags with names like project-importance (e.g. monasca-high)
19:33:28 <rhochmuth> i think that flow would work
19:33:41 <persia> And if a team wants to change later, it's not that hard to create a manual worklist with the most important things on the top.
19:33:54 <rhochmuth> so, within a project, you have several worklists
19:34:02 <rhochmuth> one list for high, another for medium, ...
19:34:05 <persia> Whereas, I think if the migration script goes straight to worklists, it's hard for teams to migrate to something using tags.
19:34:19 <SotK> persia: I concur
19:34:20 <fungi> rhochmuth: worklists aren't tied to projects. anyone can create one
19:34:23 <rhochmuth> and then the bugs are tagged and automatically put into a worklist
19:34:40 <rhochmuth> ohhh,
19:34:46 <persia> rhochmuth: And if you need to add cross-project things to the worklist, you can use the same tag, even on another project.
19:34:53 <zara_the_lemur__> bear in mind that you can filter a tag by project and name, so there's no need to use something like 'monasca-high' if you filtered by 'monasca' and 'high-priority'
19:34:54 <SotK> though, I'm not sure the project name would be needed, you could just have the worklist contain tasks in the Monasca project in stories with the relevant tag
19:35:08 <fungi> zara_the_lemur__: great point
19:35:16 <rhochmuth> i think i understand
19:35:27 <persia> zara_the_lemur__: That makes it hard to emulate something like https://bugs.launchpad.net/monasca/+bug/1368661 , where different projects have different priorities for the same story.
19:35:27 <openstack> Launchpad bug 1368661 in Glance Client "Unit tests sometimes fail because of stale pyc files" [Undecided,In progress] - Assigned to Shu Muto (shu-mutou)
19:35:39 <persia> Err, different importance.
19:35:47 <fungi> also true
19:35:52 <SotK> hm, good point
19:36:08 <rhochmuth> well, i'm not sure that case where there are different priorities is as important
19:36:22 <rhochmuth> seems like an outlier
19:36:41 <persia> I think that long-term, if people like to use tags for importance, some common tags may come into use, but for the migration script, I think we want to preserve as much information from LP as possible.
19:37:12 <SotK> seems sensible to me
19:37:13 <zara_the_lemur__> makes sense
19:38:03 <zara_the_lemur__> (also, that reminds me, I believe we need task-tag filtering in the webclient)
19:38:06 <persia> rhochmuth: I've worked on a number of differnt projects that use LP (both inside and outside openstack), and find it fairly normal that different projects have different priorities for the same issue (depending on what else needs doing, whether there are deadlines, etc.)
19:38:49 <rhochmuth> got it, a lot of learnings for me
19:39:00 <rhochmuth> thx persia
19:39:44 <SotK> so to summarise, we need to fix the priority handling in the migration script to use tags, and we need to fix the encoding bug
19:39:55 <SotK> and we will then run a test import against storyboard-dev
19:40:00 <persia> SotK: Yes, but I have another issue to raise :)
19:40:05 <SotK> persia: go on
19:40:16 <fungi> longer term having fine-grained ordering in team-specific and personal worklists may provide a more usable model, though i expect it will take a while for our community to acclimate since it is a fairly significant paradigm shift
19:40:32 <persia> If there is not a "flag day", where all the projects migrate, what is the right thing to do with LP bugs that affect multiple projects, where one of those projects is migrating, and another project is not migrating?
19:40:40 <persia> fungi: +1
19:41:13 <SotK> ah, like the one you linked
19:41:18 <persia> Yes.
19:41:28 <fungi> persia: if there are only a handful, those bugs will diverge, but comments in lp that are missing from sb will get added at reimport later
19:41:36 * persia spent the beginning of the meeting poring through Monasca bugs to find an example like that
19:41:54 <persia> fungi: And comments in sb will never be in LP?
19:41:58 <fungi> right
19:42:11 <fungi> projects with lots of shared bugs, we've recommended waitibng for a big-bang import
19:42:40 <fungi> monasca has stated that they feel the number of cross-project bugs involving their project are minimal enough to accept that bit of pain
19:42:43 <persia> My concern with that is that it might impact cross-project coordination, but I suppose if individual projects are willing to take on the extra load of watching both places while waiting for the big bang, that choice should be available.
19:43:02 <persia> And that implementation is easy enough that it shouldn't complicate things too much (other than a little extra work for early migration teams)
19:43:14 <fungi> and yeah, any additional cross-project pain might help increase momentum/adoption ;)
19:43:32 <diablo_rojo> Ha ha ha a silver lining
19:43:44 <zara_the_lemur__> :D
19:44:14 <persia> In that case, I'm happy with SotK's earlier summary, unless rhochmuth is surprised and worried about the extra cross-project effort.
19:44:29 <rhochmuth> persia: i made an assumption we didn't have too many cross-project issues
19:44:49 <persia> rhochmuth: You don't, based on looking at about 40% of your bugs :)
19:45:01 <rhochmuth> i don't think the monasca team is worried about it then
19:45:03 <rhochmuth> thx
19:46:54 <SotK> excellent
19:47:07 <SotK> thanks rhochmuth and the monasca team :)
19:47:21 <rhochmuth> that you storyboard team
19:47:23 <rhochmuth> :-)
19:47:36 <rhochmuth> is there anything you need me to do
19:47:43 <rhochmuth> or any help with
19:48:17 <diablo_rojo> SotK, ?
19:49:26 <diablo_rojo> More internet issues I expect..
19:49:55 * SotK can't think of anything immediately off the top of his head except "feel free to help out with reviews and code and stuff if you like/have the time"
19:50:06 <zara_the_lemur__> yeah, I'm not aware of anything migration-specific
19:50:14 <zara_the_lemur__> though we'll let you know if we think of anything
19:50:23 <rhochmuth> ok, sounds good
19:50:25 <SotK> indeed
19:50:30 <rhochmuth> will look for revews
19:50:56 <fungi> rhochmuth: once the import script is tested/working, we'll want to schedule a window to import your stories in production and turn off lp bug tracking (and you'll probably want to coordinate some documentation updates and announcements about that for the same time)
19:51:14 <rhochmuth> got it
19:51:19 <rhochmuth> thanks fungi
19:51:40 <zara_the_lemur__> if you have any more questions, feel free to ask as they come up! :)
19:51:41 <rhochmuth> do you have a time-frame?
19:51:53 <fungi> there's one other annoyance we've noticed, which is that disabling bug reporting in lp doesn't make existing bugs immutable, it only prevents opening new bugs... so people may still update old lp bugs from time to time if they didn't "get the memo"
19:52:14 <rhochmuth> ok, so we'll have to watch for that
19:52:30 <persia> In other contexts, I've seen teams set all open bugs in LP to invalid with a comment pointing to the new tracker.
19:52:38 <rhochmuth> but, it sounds reasonable that we'll be able to track that
19:52:45 <rhochmuth> persia: that sounds like a good idea
19:52:49 <persia> This isn't the most friendly transition, but it has been used in the past by teams migrating from LP to other (non-SB) things.
19:53:00 <fungi> persia: yep, that may be a good tack to take, though could be messy for the handful of shared bugs
19:53:22 <diablo_rojo> Easy enough to do for all the non shared bugs though I would think
19:53:48 <persia> fungi: Also popular bugs with many subscribers.  The debate over whether some project should use LP or not can overflow into unrelated issues, and people can become unhappy about bug subscriptions.
19:53:49 <fungi> yeah, the lp api, even with its warts, is pretty easy to use for something like that
19:54:04 <persia> diablo_rojo: Yep.
19:54:23 <fungi> oh, though bugs with many subscribers have a tendency to run into internal lp timeouts when trying to update them
19:54:45 <persia> #launchpad can often help with those special cases.
19:54:47 <fungi> so there could be a few that end up not getting updated with the pointer, but that's better than nothing
19:55:06 <fungi> yeah, also sometimes the smtp api will work for those where the http api and webui don't
19:56:21 <SotK> certainly sounds like a good idea to me
19:56:35 <SotK> can anyone think of anything else to discuss about this right now?
19:57:13 <diablo_rojo> Nope?
19:57:41 <SotK> then I guess we'll have open discussion for a couple of minutes
19:57:50 <SotK> #topic Open Discussion
19:58:19 <diablo_rojo> Got the second blog post basically done so we can post that whenever. Decision on what we are changing the domain to?
19:59:06 <SotK> my favourite one in the etherpad is unavailable :(
19:59:53 <SotK> we can continue this discussion in #storyboard though, since we should probably pick something and are out of time
20:00:02 <SotK> #endmeeting