19:01:28 #startmeeting storyboard 19:01:29 Meeting started Wed Apr 26 19:01:28 2017 UTC and is due to finish in 60 minutes. The chair is SotK. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:30 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:01:32 The meeting name has been set to 'storyboard' 19:01:58 #link https://wiki.openstack.org/wiki/Meetings/StoryBoard#Agenda_for_next_meeting Agenda 19:02:18 #topic Announcements 19:02:27 Nothing here 19:02:35 #topic Urgent Items 19:02:44 Also nothing urgent right now 19:02:57 #topic In Progress Work 19:03:33 I am slowly but steadily reviewing things; somewhat distracted by summit prep 19:04:10 some reviews waiting over here: 19:04:11 looking forward to the survival guide presentation 19:04:19 #link https://review.openstack.org/#/q/project:openstack-infra/storyboard 19:04:23 haha, thanks :) 19:04:37 #link https://review.openstack.org/#/q/project:openstack-infra/storyboard-webclient 19:04:41 I'm still in panic mode 19:05:44 fungI: also I've been keeping track of questions that seem to come up a lot and answers to those; I may run the list by you some time at the end of this week in case you can think of any we've missed 19:05:54 erm sorry about that strange capital I 19:06:32 off the top of my head, 'will bug ids clash', 'what happens to blueprints', 'what's the syntax for gerrit' 19:06:35 seem to come up the most 19:06:47 * SotK has also been reviewing things, but sadly failed to test any projects again 19:06:56 fungi: thanks :) 19:07:09 That "where is importance" doesn't come up more often is a testament to the existing documentation. 19:07:32 :) 19:07:47 I did have a chance to test one import for a bigger project 19:08:10 it took a very long time (so long I stopped the import before it finished), so the migration scripts will require optimisations 19:08:13 for bigger projects 19:08:22 zara_the_lemur__: happy to review your faq skeleton once you have it 19:08:32 i agree it's a great idea 19:08:54 thanks :) 19:09:58 I did some profiling of the scripts a while back because they were painfully slow, pretty much all of the slowness is waiting on the launchpad API 19:10:18 yeah, I wonder if we can do fewer requests? 19:10:56 if that's not possible, one alternative might be to only import a subset of bugs immediately (eg: only open bugs), and then stagger the others (eg: merged bugs) over a period of days or something 19:11:16 it's (very roughly) 1hr per 1000 bugs imported 19:11:45 I couldn't see any obvious places to do less requests given how launchpadlib works, but I didn't look entirely thoroughly 19:12:03 (though I expect it not to be as straightforward as that, because part of the slowness is around importing users, etc) 19:12:20 (and there are likely to be fewer new users to import as the script goes down the list of bugs) 19:13:04 (for scale, the project I tested was openstack-manuals; it's about 7000 bugs total. Nova is about double) 19:13:24 (that's open, merged and invalid) 19:13:24 we'll definitely need to come up with something then 19:13:49 yeah 19:14:01 I don't think it was a limitation with the hardware 19:14:18 (my vm had 4gb ram, on a machine with an ssd and an i7) 19:14:45 so there probably aren't easy ways to speed it up around that. :( 19:15:33 could you split the import sets? and have several machines operating on the same db (assuming its an api issue) 19:16:05 I expect that would work actually, yeah 19:16:29 we'd need to add some way of doing it to the scripts first though 19:16:37 * zara_the_lemur__ wonders how we'd tell the scripts to split it 19:16:40 ah snap-ish 19:16:44 given that the import is meant to be idempotent, does rerunning to get additional bugs later go more quickly since it can skip the content it already imported? 19:16:55 yes, it should be 19:17:36 i don't personally care hos long the import takes to complete. if it means we leave the script running for a week to seem most of the bugs and then do some faster passes afterward to get the updates before telling people to switch, that seems good enough to nme 19:17:41 er, to me 19:17:49 s/hos/how/ 19:18:10 okay, if you're okay with that, then I'm also fine with leaving it :) 19:18:15 * fungi isn't typing well this afternoon 19:18:16 me too 19:18:31 so we have 3 known issues with the scripts: handling unusual unicode character codes, erroring out when given any sort of error response code (including a 410 if, say, a user has been banned from lp) and slowness for big projects (which maybe isn't an issue) 19:18:38 it's a one-time event, more or less, so not like it needs extreme optimization 19:18:46 indeed 19:19:33 I thought '16 hours to import a project' might be pushing it, but if it's fine to leave it running then that works for me 19:20:05 we had one project import fail because it tried to import an lp user whose name includes a robot emoji 19:20:18 and one fail because it tried to import a user who had been banned from lp for spam, raising a 410 19:20:48 so we need better handling for that sort of thing 19:20:53 the emoji thing should be fairly simple to fix, I'm fairly sure we just need to change the charset the database is using 19:21:45 *nod* 19:22:14 likely just need to make sure it's using 4-byte instead of mysql's 3-byte default 19:22:37 indeed 19:22:44 if it only chokes on 4-byte codepoints, that's a sure sign 19:23:10 etherpad had a similar issue up until a couple years ago 19:23:54 or, rather, our etherpad deployment did (because we were using mysql defaults) 19:24:40 yeah, I'd naively assumed that utf8 meant utf8 rather than utf8-that-is-3-bytes-or-less 19:24:47 that sounds like fun to discover :) 19:26:04 talk planning, review, import script testing... I don't think I have anything else in-progress 19:26:22 * SotK neither 19:26:28 I haven't got to kendall's latest branch patch updates 19:26:37 * SotK has, just this evening 19:26:38 so will mention those since she's not able to attend today 19:26:41 ah, nice 19:27:25 ah you should probably -1 so it updates her 19:27:34 good point, will do 19:27:46 anything else in progress? 19:28:19 nope, I'm behind enough on the things I've mentioned already \o/ 19:28:24 #topic Open Discussion 19:29:38 tallllllllllllk 19:29:52 as in, boston talk 19:30:28 it will be on weds 10th 19:30:29 the title is catchy 19:30:53 yay, that was diablo_rojo's work :D 19:30:57 she'll be pleased 19:31:06 i've got it on my shortlist (for sessions i will definitely attend) 19:31:12 \o/ 19:32:09 I'll find the details now for anyone interested, since I'm procrastinating from actually making material for the talk 19:32:42 two weeks today! 19:32:49 I seem to recall it's weds early afternoon 19:33:15 yep, 1:50-2:30pm local time 19:33:24 according to my notes anyway 19:33:37 #link https://www.openstack.org/summit/boston-2017/summit-schedule/events/17632/storyboard-101-survival-guide-to-the-great-migration 19:34:06 we should also be generally around for questions etc over the summit, and available in #storyboard 19:34:10 with the usual summit lag ;) 19:34:22 (how do people look at irc and talk in person?) 19:34:26 (never figured it out) 19:35:08 portable computing devices 19:35:17 and a lot of inconvenience ;) 19:35:34 we live in the future~ 19:35:57 also, if anyone has been receiving a lot of questions about storyboard and would like them covered, feel free to send them our way 19:36:07 +1 19:36:55 i'll see if i can round up some i received/answered for you 19:37:02 thank you :) 19:37:16 I see the 'frequently asked questions' that make it into #storyboard or the mailing list, but I feel like those might not be that representative 19:37:42 if people are already informed enough to find their way there 19:37:48 diablo_rojo_phon and i have received some questions in private e-mail discussions 19:37:58 so happy to help get those merged in 19:38:08 cool, I think that would help 19:42:14 * zara_the_lemur__ doesn't think she has anything else to mention 19:42:31 * SotK also has nothing 19:42:39 * diablo_rojo_phon catches up from dentist parking lot 19:42:49 that's dedication! 19:42:53 hey diablo_rojo_phon :) 19:43:10 try not to type with your teeth! 19:43:10 Haha if you say so. 19:43:53 Haha I'll avoid that. Not sure it would work great on the phone anyway. 19:44:23 I saw SotK made comments on my patch that I will address before the end of the week. 19:44:40 I also need to finish writing up the release goal before the end of next week. 19:44:49 And work on slides for our presentation. 19:45:39 sounds like a busy week :) 19:45:51 That's only the sb stuff lol 19:46:05 heh 19:46:18 well we should all be around for the slides 19:46:19 I haven't gotten confirmation from cloudkitty that they want to move, but we can go ahead and do craton if we want. 19:46:31 I just need to give them a heads up about when it will happen. 19:46:51 do we want to decide a time/day? 19:47:15 guessing we want to rule out summit week 19:47:32 I think we do 19:47:36 Add stuff to slides and we should have them done eod Friday so we can go through them next week and break it up. 19:47:49 yeah, i'm travelling may 4 through 14 so will likely need to find another infra-root volunteer if they want it on one of those dates 19:48:06 fungi: noted :) 19:48:09 I too am afks from like the 5th to the 16th 19:48:26 I expect the end of this week to be a big rush of 'ACK WE ACTUALLY HAVE TO DO A TALK' on our side 19:48:42 i already have that x3 or 4 anyway 19:48:49 haha 19:49:01 need to decide which talks i should panic about first ;) 19:49:11 ah, an important decision 19:49:20 that will consume all mental energy 19:49:22 Yeah I have sb, upstream training, keynote, foundation fundamentals and ask a stacker shifts... 19:49:23 until it's time for the first talk 19:49:27 leaving no time to prepare for a talk 19:50:02 fungi: not all simultaneously? 19:50:26 diablo_rojo_phon: my multitasking skills are sub-par 19:50:38 even panics must be serialized 19:51:05 solution: panic over which thing to panic about instead 19:51:24 all else fails, improv 19:51:32 SotK: +1 19:52:27 anything else to discuss? 19:53:04 Don't think so. 19:53:19 at least we're not doing a musical 19:53:22 * zara_the_lemur__ also has nothing 19:53:29 ha 19:53:39 thanks for coming folks! 19:53:42 #endmeeting