Summary of IRC Meeting in #couchdb-meeting at Thu Dec 18 20:05:23 2014: Attendees: awenkhh, Kxepal, jan____, robertkowalski, chewbranca, strmpnk
- Preface - Info: CouchDB Day 2015 Hamburg - CouchDB Day 2015 in Hamburg, Germany - Info: if you attend and would like to talk about CouchDB 2.0 or something please get in touch with awenkhh ([email protected]) - Link: http://day.couchdb.org - Link: https://github.com/andywenk/couchdb_day_hamburg/blob/master/couchdb_day_january_2015.md - CouchDB 2.0 - Action: strmpnk to post on dev@ ML his plan about dependencies reorganization - Link: https://issues.apache.org/jira/browse/COUCHDB-2513 - Cloudant query aka mango contribution - Fauxton - Link: https://issues.apache.org/jira/browse/COUCHDB-2390 - Branding routines: logo, slogan - Happy New 2015 Year IRC log follows: ## Preface ## [Thu Dec 18 20:06:13 2014] <awenkhh>: #info CouchDB Day 2015 Hamburg [Thu Dec 18 20:06:23 2014] <Kxepal>: ok, so what will be the topics? CouchDB 2.0, CouchDB Day in Hamburg I suppose, Mango, some year over report...what else? [Thu Dec 18 20:07:45 2014] <awenkhh>: well Ican start with a short info about CouchDB Day 2015 in Hamburg, Germany [Thu Dec 18 20:07:51 2014] <Kxepal>: ok, please propose any other topic on the way ## CouchDB Day 2015 in Hamburg, Germany ## [Thu Dec 18 20:08:20 2014] <Kxepal>: awenkhh: so while you'd started...go on! (: [Thu Dec 18 20:08:26 2014] <awenkhh>: So far the event is very well received. The page can be found here day.couchdb.org [Thu Dec 18 20:08:40 2014] <Kxepal>: #link http://day.couchdb.org [Thu Dec 18 20:08:54 2014] <awenkhh>: robertkowalski an me are organising the day with a lot of help from the CouchDB Community [Thu Dec 18 20:09:14 2014] <awenkhh>: we now have very cool sponsors and there are around 25 tickets left atm [Thu Dec 18 20:09:33 2014] <awenkhh>: we expect to give them away in erly January [Thu Dec 18 20:09:48 2014] <awenkhh>: jan____ will be the keynote speaker [Thu Dec 18 20:10:02 2014] <Kxepal>: \o/ [Thu Dec 18 20:10:06 2014] <awenkhh>: and as we plan to have two more talks, we are looking for someone to speak [Thu Dec 18 20:10:17 2014] <awenkhh>: we would really love to have a CouchDB 2.0 talk [Thu Dec 18 20:10:32 2014] <awenkhh>: so if you attend and would like to talk about CouchDB 2.0 [Thu Dec 18 20:10:35 2014] <Kxepal>: awenkhh: is there any agenda about? [Thu Dec 18 20:10:50 2014] <awenkhh>: please get in touch with me ([email protected]) [Thu Dec 18 20:10:52 2014] <awenkhh>: yes ... [Thu Dec 18 20:11:02 2014] <Kxepal>: cannot see it on website... [Thu Dec 18 20:11:12 2014] <awenkhh>: https://github.com/andywenk/couchdb_day_hamburg/blob/master/couchdb_day_january_2015.md [Thu Dec 18 20:11:33 2014] <awenkhh>: we will post it soon but we want to fill the talk gap before that [Thu Dec 18 20:11:40 2014] <Kxepal>: #link https://github.com/andywenk/couchdb_day_hamburg/blob/master/couchdb_day_january_2015.md [Thu Dec 18 20:11:53 2014] <jan____>: here now, sorry [Thu Dec 18 20:12:04 2014] <Kxepal>: welcome to the party jan____ ! [Thu Dec 18 20:12:05 2014] <awenkhh>: hey jan____ ;-) [Thu Dec 18 20:13:14 2014] <Kxepal>: #info if you attend and would like to talk about CouchDB 2.0 or something please get in touch with awenkhh ([email protected]) [Thu Dec 18 20:13:31 2014] <Kxepal>: thanks awenkhh ! ## CouchDB 2.0 ## [Thu Dec 18 20:13:52 2014] <Kxepal>: ok, here we are [Thu Dec 18 20:14:20 2014] <Kxepal>: strmpnk: your turn (: [Thu Dec 18 20:14:35 2014] <strmpnk>: I've been doing a little bit of development around 2.0 and found the build system lacking so I've got some work in progress on cleaning it up. [Thu Dec 18 20:14:59 2014] <strmpnk>: With 2.0 we transitioned to using rebar to manage most things with a simple configure script to setup a few settings. [Thu Dec 18 20:15:32 2014] <strmpnk>: I've mostly focussed on cleaning up the couch sub-repo and started working on getting the configure script to pass the proper settings through. [Thu Dec 18 20:15:53 2014] <strmpnk>: What I'm interested in is people who have interesting platforms or settings that they'd like to have supported. [Thu Dec 18 20:16:14 2014] <Kxepal>: strmpnk: what do you mean by "interesting platforms"? [Thu Dec 18 20:16:15 2014] <strmpnk>: Part of it should be good defaults for common systems. The other is appropriate flags to make it easy to override. [Thu Dec 18 20:16:19 2014] <jan____>: I can do a FreeBSD test run [Thu Dec 18 20:16:30 2014] <strmpnk>: Something people use. Doesn't have to be exotic. [Thu Dec 18 20:17:10 2014] <Kxepal>: ah. can help with Gentoo, NixOS, FreeBSD too, CentOS [Thu Dec 18 20:17:12 2014] <strmpnk>: Windows support is still up in the air since I haven't tested rebar's port building system on that platform so I don't know if it will be trivial but I'm hoping to give that a shot too. [Thu Dec 18 20:17:55 2014] <Kxepal>: and also I think I'll could get a help for Fedora [Thu Dec 18 20:18:15 2014] <strmpnk>: Beyond that stage, I'm also looking at reorganizing dependencies and .app dependencies to allow pulling in couchdb as a rebar dep as well as building it into a relx release. [Thu Dec 18 20:18:36 2014] <strmpnk>: That will likely happen after 2.0 but if others are interested in it I can try to push that through faster. [Thu Dec 18 20:19:04 2014] <strmpnk>: It just requires enough code to be shifted around that it might be safe to wait. [Thu Dec 18 20:19:38 2014] <strmpnk>: The end goal is to make CouchDB a much better citizen in the Erlang side of the world. [Thu Dec 18 20:19:48 2014] <Kxepal>: strmpnk: do you have a plan for reorganization? [Thu Dec 18 20:20:42 2014] <Kxepal>: I think it's worth to post it on dev@ to raise the discussion and let someone pick it if you don't have a time for [Thu Dec 18 20:20:59 2014] <jan____>: yeah, +1 [Thu Dec 18 20:21:04 2014] <strmpnk>: It makes sense to have a clear DAG for the .app files (each OTP application or repo under the top-level src/ dir after configure). The current couch app is split between low and high level stuff. [Thu Dec 18 20:21:28 2014] <jan____>: this is great work and worth sharing to the larger dev team to see if anyone wants to chip in [Thu Dec 18 20:21:39 2014] <awenkhh>: +1 cool! [Thu Dec 18 20:21:40 2014] <jan____>: strmpnk: yes, that, and also document this, like with a diagram and all :) [Thu Dec 18 20:21:45 2014] <strmpnk>: The other changes are mostly missing deps and possibly adding a src/ for the top level repo so it can bring the apache couchdb together as intended while subcomponents can be reused. [Thu Dec 18 20:22:12 2014] <strmpnk>: jan____: I've got some dot file generators that look pretty scary in the current state but it's improving with time. [Thu Dec 18 20:22:18 2014] <strmpnk>: One more thing, I was chatting with jan____ about the ease of building couchjs. It might be worth bringing up node or io.js support for that. [Thu Dec 18 20:22:35 2014] <strmpnk>: It specifically might make windows support that much easier to deal with. [Thu Dec 18 20:23:03 2014] <jan____>: good point, but significant work to get going [Thu Dec 18 20:23:09 2014] <jan____>: so would not block on this for 2.0 [Thu Dec 18 20:23:35 2014] <strmpnk>: Yes. It's definitely a conversation for dev [Thu Dec 18 20:23:54 2014] <strmpnk>: For 2.0 I'm focussed on the configure + rebar defaults stuff. [Thu Dec 18 20:24:03 2014] <jan____>: yeah, excellent [Thu Dec 18 20:24:22 2014] <strmpnk>: But I'd be happy to include some of the other work if people thing there are reasons to include it in the first 2.x release. [Thu Dec 18 20:25:17 2014] <Kxepal>: #action strmpnk to post on dev@ ML his plan about dependencies reorganization [Thu Dec 18 20:25:28 2014] <Kxepal>: strmpnk: I believe you don't mind for that (: [Thu Dec 18 20:26:11 2014] <strmpnk>: Definitely don't mind. I've been meaning to but time has been short since I started looking at this. [Thu Dec 18 20:26:22 2014] <strmpnk>: Also I have one ticket for the rebar side open here: https://issues.apache.org/jira/browse/COUCHDB-2513 [Thu Dec 18 20:26:44 2014] <strmpnk>: The repos need to be updated with some more work but I'll track work there. [Thu Dec 18 20:27:21 2014] <Kxepal>: #link https://issues.apache.org/jira/browse/COUCHDB-2513 [Thu Dec 18 20:27:56 2014] <strmpnk>: That should be good. The rest of the discussion can be on the ML or IRC if anyone wants to ping me directly. [Thu Dec 18 20:28:45 2014] <jan____>: cool, thank you strmpnk, this is exciting stuff! [Thu Dec 18 20:29:08 2014] <awenkhh>: yes - super cool strmpnk - thank you! [Thu Dec 18 20:29:12 2014] <Kxepal>: strmpnk: great! I believe you'd like also to contact with djc, plemenkov and other package maintainers to receive feedback from them about building [Thu Dec 18 20:30:12 2014] <strmpnk>: Thank you. Yes. [Thu Dec 18 20:31:20 2014] <Kxepal>: cool. now about my part... [Thu Dec 18 20:32:05 2014] <Kxepal>: I have a two main tasks for 2.0: fix eunit tests and documentation update [Thu Dec 18 20:33:03 2014] <Kxepal>: for eunit I'm slowly, but following chewbranca review notes made for 1.x PR + backporting all etap changes made for 2.0 to fix broken tests which currently simply disabled [Thu Dec 18 20:33:44 2014] <Kxepal>: actually, review notes I'd already covered, now it's the fixing part in progress [Thu Dec 18 20:33:58 2014] <jan____>: wow! [Thu Dec 18 20:34:19 2014] <Kxepal>: also I received a lot of valuable feedback about eunit tests from Peter Lemenkov, who already picked them for 1.6.1 on Fedora [Thu Dec 18 20:34:30 2014] <chewbranca>: Kxepal: nice! [Thu Dec 18 20:34:57 2014] <Kxepal>: they works great, except few bugs + they runs on very very very slow build servers causing failures by timeouts [Thu Dec 18 20:36:17 2014] <Kxepal>: so I come to the idea to make two checks in make file: one for common usage, which everyone will run and check-slow for build farms which injects some timeout modifier [Thu Dec 18 20:36:19 2014] <jan____>: nice [Thu Dec 18 20:36:37 2014] <jan____>: thatâs a good idea [Thu Dec 18 20:36:54 2014] <awenkhh>: great work Kxepal +1 [Thu Dec 18 20:37:05 2014] <Kxepal>: so we'll be able to track QA for our code while let it build on slow hosts where you actually don't care how fast it is...just need to build it and test it [Thu Dec 18 20:37:26 2014] <jan____>: yeah, thatâs excellent [Thu Dec 18 20:38:05 2014] <Kxepal>: I don't think I'll make a final PR on this year, but will try for that, but we'll have two weeks holidays after so that's a lot of time to finish all the stuff [Thu Dec 18 20:38:21 2014] <jan____>: :D [Thu Dec 18 20:38:26 2014] <Kxepal>: yea (: [Thu Dec 18 20:38:55 2014] <jan____>: Kxepal: is that it for you end, or is there more? [Thu Dec 18 20:39:22 2014] <Kxepal>: as about documentation it's also in slow progress since I have to check every existed statement, add new ones, notice differences...it's just in progress for now, but no results so far to share [Thu Dec 18 20:39:42 2014] <Kxepal>: jan____: yes, I think I'm done [Thu Dec 18 20:39:57 2014] <awenkhh>: Kxepal unfortunately I do not have enough time to really work on the docs [Thu Dec 18 20:40:07 2014] <awenkhh>: but I am very happy to review what you have [Thu Dec 18 20:40:28 2014] <jan____>: cool, thanks Kxepal [Thu Dec 18 20:40:42 2014] <Kxepal>: awenkhh: that would be awesome because sometime I feel lack on second pair of eyes (: will call you then [Thu Dec 18 20:40:42 2014] <jan____>: I donât have much to report, but Iâd like to move discussion on shipping 2.0 a little further. [Thu Dec 18 20:41:00 2014] <awenkhh>: Kxepal please do!!!! [Thu Dec 18 20:41:06 2014] <jan____>: I see that Cloudant has been tireless at moving this forward on all fronts and thatâs great to see [Thu Dec 18 20:41:14 2014] <jan____>: they started even working on PouchDB today. [Thu Dec 18 20:41:25 2014] <Kxepal>: oh wow [Thu Dec 18 20:41:27 2014] <awenkhh>: wow !!! [Thu Dec 18 20:41:30 2014] <strmpnk>: Nice [Thu Dec 18 20:41:56 2014] <jan____>: specifically to work on CouchDB 2,0 compat [Thu Dec 18 20:41:56 2014] <Kxepal>: jan____: are they going to conquer the world?(: [Thu Dec 18 20:42:05 2014] <jan____>: dunno :) [Thu Dec 18 20:42:10 2014] <jan____>: I hope not :P [Thu Dec 18 20:42:30 2014] <jan____>: that probably should have been [off] [Thu Dec 18 20:42:37 2014] <Kxepal>: haha [Thu Dec 18 20:42:37 2014] <awenkhh>: :D [Thu Dec 18 20:43:11 2014] <Kxepal>: jan____: so what about release dates? sorry for distraction [Thu Dec 18 20:43:29 2014] <jan____>: One thing Iâm keen on, and we touched on it already is understanding what a release would look like [Thu Dec 18 20:44:03 2014] <jan____>: that includes: what bytes to ship, and it what form, what the installation procedure looks like and what upgrades from 1.0 look like [Thu Dec 18 20:44:48 2014] <jan____>: we kinda have the developer-preview branch, but thatâs not a good âreleaseâ and was put together in a hurry for ApacheCon EU. [Thu Dec 18 20:46:03 2014] <jan____>: any thoughts? :) [Thu Dec 18 20:47:11 2014] <Kxepal>: I think it's good idea to set it down what a release would like [Thu Dec 18 20:47:43 2014] <Kxepal>: collect all the todo stuff and set priorities to them which goes with 2.0 and which are for 2.1 etc [Thu Dec 18 20:48:05 2014] <jan____>: yeah, that sounds like a plan. [Thu Dec 18 20:48:34 2014] <awenkhh>: +1 [Thu Dec 18 20:49:11 2014] <Kxepal>: call be pessimist, but I don't believe that 2.0 will be an ideal and a lot of people will go to upgrade onto. life showed that two-or-three releases after major bump goes with a lot of hotfixes and are more suitable for production usage [Thu Dec 18 20:49:47 2014] <Kxepal>: so probably we can account that moment to not try to push everything to the single deadline [Thu Dec 18 20:50:33 2014] <jan____>: in general that is true, but I think we have the advantage that BigCouch is a production system, so a lot of small issues are already ironed out [Thu Dec 18 20:50:42 2014] <jan____>: not saying this isnât gonna be hard, but yeah [Thu Dec 18 20:50:58 2014] <Kxepal>: devil is in details, you know that [Thu Dec 18 20:51:05 2014] <awenkhh>: but we also don't wanna hava a half baked release - we wanna make a lot of "marketing loundness" with 2.0 [Thu Dec 18 20:51:07 2014] <strmpnk>: I'd also caution the 1.6.x upgrades to try 2.0 separately first before moving things directly over anyway. We might be able to do a 2.1 release after to support upgrades to make this intention clearer. [Thu Dec 18 20:51:09 2014] <jan____>: but generally I agree, we should get something out that is good enough, but ca be free to add more things quickly in point releases [Thu Dec 18 20:51:45 2014] <jan____>: strmpnk: what intention? [Thu Dec 18 20:51:52 2014] <jan____>: why not support upgrades in 2.0? [Thu Dec 18 20:52:04 2014] <jan____>: thatâs something Iâd say should be supported :) [Thu Dec 18 20:52:23 2014] <awenkhh>: I personally would better wait one month to have a maybe better release - when the feedback sucks ... that would be really unfortunate [Thu Dec 18 20:52:38 2014] <awenkhh>: but waiting too long is for sure not a plan on the other hand [Thu Dec 18 20:52:48 2014] <strmpnk>: It's a lot of subtle breaking changes is all. Our intention is to make clustering a first class feature so a lot of old couch-isms become underscored with gotchas. [Thu Dec 18 20:53:13 2014] <jan____>: I have a rough idea of getting a group of beta testers that can help during the run up to 2. [Thu Dec 18 20:53:17 2014] <jan____>: 2.0 [Thu Dec 18 20:53:30 2014] <strmpnk>: It's better to get the software a good "this is 2.0" release. The 1.x branch still works great for those that haven't been on clusters. [Thu Dec 18 20:53:46 2014] <jan____>: maybe 10-30 people who use CouchDB in production that helps us find issues with prerelease builds so we can make lifes for upgrades as easy as possible [Thu Dec 18 20:53:49 2014] <Kxepal>: would like to agree with strmpnk [Thu Dec 18 20:53:50 2014] <jan____>: *lives [Thu Dec 18 20:54:16 2014] <jan____>: strmpnk: thatâs a good point. it might be that 2.0 users are more likely to start fresh [Thu Dec 18 20:54:18 2014] <awenkhh>: heavy testing with production useres would be a big advantafe ! [Thu Dec 18 20:54:20 2014] <Kxepal>: we also have open discussion about fate of os_daemons, externals, query servers [Thu Dec 18 20:54:26 2014] <Kxepal>: which people tend to use a lot [Thu Dec 18 20:54:46 2014] <strmpnk>: Granted, if someone writes some awesome upgrade logic I wouldn't block it. It just shouldn't block 2.0. [Thu Dec 18 20:54:48 2014] <jan____>: but Iâm not sure this is the right dichotomy: on the one hand it is, seamless ugprades are a feature and as such push out hte release date [Thu Dec 18 20:55:32 2014] <jan____>: but on the other hand, I am *very* afraid of the Python 3 syndrome, so I want 2.0 to be as friendly as 1.x is for everyone, regardless of where they are coming from [Thu Dec 18 20:55:41 2014] <jan____>: strmpnk: dunno, what upgrade logic is required? [Thu Dec 18 20:56:17 2014] <strmpnk>: mem3 shard mapping for legacy dbs mostly. [Thu Dec 18 20:56:27 2014] <strmpnk>: Probably some configuration stuff that needs to be checked. [Thu Dec 18 20:57:21 2014] <jan____>: maybe a shortcut would be that in 2.0 the âupgradeâ is using the local interface and in 2.x we can add support ofr the cluster layers [Thu Dec 18 20:57:32 2014] <chewbranca>: I still think recommending replication as an upgrade mechanism is a good safe way to get people migrated, not saying it should be the only but it could be useful as a recommendation for people testing things out or who want to get up and running quickly [Thu Dec 18 20:57:39 2014] <strmpnk>: Doing so is costly enough that it's probably good to treat it as a one-way step but we could also support merging shards to go back. [Thu Dec 18 20:58:00 2014] <strmpnk>: I'll have to check the disk format too but we could do a release on 1.6.x to support anything new. [Thu Dec 18 20:58:04 2014] <jan____>: strmpnk: uh, going back, heh *cough* [Thu Dec 18 20:58:11 2014] <Kxepal>: jan____: we already have python3 syndrome cos of breaking changes for details and different setup. same /_config endpoint for example. but still our situation is much better since you don't have to rewrite from scratch all couchdb client [Thu Dec 18 20:58:11 2014] <jan____>: would not make that a priority :) [Thu Dec 18 20:59:29 2014] <jan____>: Kxepal: right, we have minor breakages (although Iâd seee a full list of old/new/changed endpoints as they are distributed over cluster and local interface, because it is a lot more than originally proposed), but I want to make sure it stays that way [Thu Dec 18 21:00:51 2014] <jan____>: cool then, anything else? [Thu Dec 18 21:01:20 2014] <Kxepal>: jan____: I post some list already on ML (without new ones, just about breaking changes), but it need to be updated since things are going to change [Thu Dec 18 21:03:04 2014] <jan____>: Kxepal: thanks, yeah, I kinda envision something like a spreadsheet or something in table format that just lists it all to see at once [Thu Dec 18 21:04:16 2014] <Kxepal>: trying to find it in archives... [Thu Dec 18 21:06:09 2014] <Kxepal>: ETIMEOUT, sorry. will share after the meeting or in the middle to not block it [Thu Dec 18 21:06:38 2014] <jan____>: cool, letâs wrap this up [Thu Dec 18 21:07:00 2014] <awenkhh>: everything sounds great! [Thu Dec 18 21:07:43 2014] <Kxepal>: jan____: I think we need to make this comparison in wiki in anyway [Thu Dec 18 21:07:51 2014] <Kxepal>: ML is bad place for it [Thu Dec 18 21:07:59 2014] <jan____>: Kxepal: yea [Thu Dec 18 21:09:12 2014] <Kxepal>: ok, so..anything else about 2.0? [Thu Dec 18 21:10:25 2014] <Kxepal>: good, moving forward [Thu Dec 18 21:10:27 2014] <strmpnk>: Do we have any 2.0 specific way to get feedback on preview releases? It might be nice to know if anyone fails to get it working &c. [Thu Dec 18 21:10:50 2014] <jan____>: strmpnk: not yet [Thu Dec 18 21:10:58 2014] <Kxepal>: strmpnk: I saw some reports about broken fauxton, but we fixed that [Thu Dec 18 21:11:14 2014] <Kxepal>: but nothing else [Thu Dec 18 21:11:34 2014] <strmpnk>: I assume we could add a link to the setup endpoint but maybe welcome could include a link. Anyway, just a last minute thought. [Thu Dec 18 21:12:18 2014] <Kxepal>: strmpnk: like curl -XPOST /_bug -d '"it doesn\'t works at all!"' ?(: [Thu Dec 18 21:12:56 2014] <Kxepal>: or /_feedback [Thu Dec 18 21:13:49 2014] <Kxepal>: but anyway people who get devpreview knows how to contact with us. or should know (: ## Cloudant query aka mango contribution ## [Thu Dec 18 21:14:05 2014] <awenkhh>: strmpnk I like that idea !!! [Thu Dec 18 21:14:51 2014] <Kxepal>: ok, we have a lot of requests to see cloudant query feature in couchdb 2.0 and cloudant folks also would like to contribute it [Thu Dec 18 21:15:18 2014] <Kxepal>: so to make this real we need to: [Thu Dec 18 21:16:22 2014] <Kxepal>: 1. start the vote about (this didn't happened) [Thu Dec 18 21:16:22 2014] <Kxepal>: 2. submit mango snapshot to IP Clearance by ASF member [Thu Dec 18 21:16:22 2014] <Kxepal>: 3. make it accept [Thu Dec 18 21:16:22 2014] <Kxepal>: 4. ... [Thu Dec 18 21:16:22 2014] <Kxepal>: 5. integrate it with couchdb 2.0 [Thu Dec 18 21:16:38 2014] <Kxepal>: jan____: is that correct? [Thu Dec 18 21:17:22 2014] <robertkowalski>: hi just returned from christmas market [Thu Dec 18 21:17:31 2014] <robertkowalski>: i am on that topic sorry for the delay [Thu Dec 18 21:17:52 2014] <robertkowalski>: i was not aware that there are still things missing [Thu Dec 18 21:17:52 2014] <Kxepal>: welcome robertkowalski! Merry Christmas (: [Thu Dec 18 21:18:16 2014] <awenkhh>: do we need to vote on this? and where? [Thu Dec 18 21:18:28 2014] <awenkhh>: what kind of concensus do we need? [Thu Dec 18 21:18:35 2014] <Kxepal>: according our policy yes, we need the vote, on dev@ ML [Thu Dec 18 21:18:55 2014] <robertkowalski>: regarding the 2.0 release (sorry for beeing late): the fauxton team has some blockers where we currently wait if they will settle before the 2.0 release [Thu Dec 18 21:18:57 2014] <Kxepal>: and probably it's better to be started by someone from Cloudant to make the things looks sane [Thu Dec 18 21:18:58 2014] <robertkowalski>: https://issues.apache.org/jira/browse/COUCHDB-2390 [Thu Dec 18 21:19:46 2014] <robertkowalski>: i will send the vote soon, i am just picking up the loose ends on the cloudant and couchdb side [Thu Dec 18 21:20:09 2014] <awenkhh>: unfortunately this link is dead: https://incubator.apache.org/ip-clearance/index.html [Thu Dec 18 21:21:12 2014] <Kxepal>: oh. need to be reported on INFRA [Thu Dec 18 21:21:17 2014] <robertkowalski>: i started to work for cloudant in october, and i will take care of this topic [Thu Dec 18 21:22:09 2014] <awenkhh>: robertkowalski very cool - but you will need assistance from the VP or a member for the IP clearance stuff [Thu Dec 18 21:22:26 2014] <awenkhh>: would be cool to write to the thread Joan opened [Thu Dec 18 21:22:53 2014] <awenkhh>: ... in private@ [Thu Dec 18 21:23:14 2014] <robertkowalski>: awenkhh: yes, jan____ will help me [Thu Dec 18 21:23:33 2014] <jan____>: Iâm on it no worries :) [Thu Dec 18 21:23:35 2014] <robertkowalski>: awenkhh: i started yesterday [Thu Dec 18 21:24:03 2014] <awenkhh>: *respond [Thu Dec 18 21:24:49 2014] <Kxepal>: robertkowalski: cool! congrats (: if you can push it forward that would be awesome [Thu Dec 18 21:24:56 2014] <robertkowalski>: basically i don't have anything further to report as i just started [Thu Dec 18 21:25:06 2014] <awenkhh>: robertkowalski +1 [Thu Dec 18 21:26:36 2014] <Kxepal>: ok, moving to Fauxton topic? robertkowalski ## Fauxton ## [Thu Dec 18 21:29:01 2014] <robertkowalski>: ok [Thu Dec 18 21:29:11 2014] <robertkowalski>: so we are making good progress in fauxton [Thu Dec 18 21:29:22 2014] <robertkowalski>: there are some issues for 2.0 tat are blockers [Thu Dec 18 21:29:33 2014] <robertkowalski>: https://issues.apache.org/jira/browse/COUCHDB-2390 is one of them [Thu Dec 18 21:29:53 2014] <Kxepal>: #link https://issues.apache.org/jira/browse/COUCHDB-2390 [Thu Dec 18 21:30:04 2014] <Kxepal>: yea, fauxton becomes better day by day [Thu Dec 18 21:30:14 2014] <robertkowalski>: they basially all deal with multiple nodes and fauxton posting to them in front of a (round robin) load balancer [Thu Dec 18 21:30:52 2014] <robertkowalski>: the last idea for the configuration section was to use etcd (from coreOS) - a raf consensus key/value store for the config [Thu Dec 18 21:31:05 2014] <robertkowalski>: the topic is discussed at https://issues.apache.org/jira/browse/COUCHDB-2390 [Thu Dec 18 21:31:52 2014] <Kxepal>: sounds cool, but I don't think we could get etcd everywhere [Thu Dec 18 21:31:53 2014] <jan____>: didnât aphyr rip their failure modes into shreds? :) [Thu Dec 18 21:32:06 2014] <Kxepal>: also good point (: [Thu Dec 18 21:32:10 2014] <robertkowalski>: and currently our options are: deleting the module from fauxton, writing a message next to it that it does not work or waiting until the release if mabe someone with more backend experience solved our problem [Thu Dec 18 21:32:25 2014] <Kxepal>: https://aphyr.com/posts/316-call-me-maybe-etcd-and-consul [Thu Dec 18 21:32:28 2014] <jan____>: https://aphyr.com/posts/316-call-me-maybe-etcd-and-consul might not apply to our use case [Thu Dec 18 21:32:32 2014] <Kxepal>: P: [Thu Dec 18 21:32:35 2014] <jan____>: they also might have fixed things [Thu Dec 18 21:32:48 2014] <robertkowalski>: and we decided to wait, as we don't see a release in the next week currently. [Thu Dec 18 21:33:37 2014] <Kxepal>: robertkowalski: you seems miss couchdb 2.0 topic....((: [Thu Dec 18 21:34:34 2014] <robertkowalski>: Kxepal: can you explain this? [Thu Dec 18 21:34:39 2014] <Kxepal>: I think the only idea which might works fine there is to make /_config as database which will replicate over the cluster node with some builtin filter which strips node specifics settings [Thu Dec 18 21:34:46 2014] <Kxepal>: robertkowalski: nvm [Thu Dec 18 21:35:25 2014] <Kxepal>: but still a work to do [Thu Dec 18 21:36:15 2014] <robertkowalski>: yes. we decided to wait for a final decision, as deleting or adding a message ("this does not work") is a task of 1hr including review [Thu Dec 18 21:36:21 2014] <strmpnk>: I can see a lot of edge cases popping up like lack of write permission on the config file but the db showing the proper value. [Thu Dec 18 21:36:47 2014] <Kxepal>: strmpnk: we also have an idea to move away from config files [Thu Dec 18 21:37:23 2014] <Kxepal>: so this eventually could help with [Thu Dec 18 21:37:40 2014] <strmpnk>: Kxepal: I have a bit to say about that with regard to release embedding but I'd be curious to see this discussion develop in the open. Possibly better on the ML? [Thu Dec 18 21:37:52 2014] <robertkowalski>: yes, i like that idea [Thu Dec 18 21:38:04 2014] <Kxepal>: strmpnk: sure, that's need to be discussed on ML (: [Thu Dec 18 21:39:02 2014] <robertkowalski>: just to get some opinions [Thu Dec 18 21:39:25 2014] <robertkowalski>: what wuld the attendees of this meeting think about a fauxton without the option to configure couch? [Thu Dec 18 21:39:40 2014] <robertkowalski>: (we delete the module) [Thu Dec 18 21:39:50 2014] <awenkhh>: hm ... [Thu Dec 18 21:40:02 2014] <awenkhh>: so just configure it via the ini files? [Thu Dec 18 21:40:10 2014] <strmpnk>: +1 though it'd be create to see it be read-only at least and the possibility to make it work in the future with the proper installation. [Thu Dec 18 21:40:11 2014] <Kxepal>: robertkowalski: that would be bad since onfly configuring from web is an important feature [Thu Dec 18 21:40:28 2014] <jan____>: Iâd say this would be an acceptable thing for 2.0 â and something we can add in 2.1 or later. [Thu Dec 18 21:40:38 2014] <jan____>: good call on read-only with instructions on how to change things [Thu Dec 18 21:40:46 2014] <awenkhh>: jup! [Thu Dec 18 21:40:59 2014] <strmpnk>: s/create/great/ [Thu Dec 18 21:41:15 2014] <robertkowalski>: i am +1 on deleting the feature as a serious production would include configuration management anyway for the nodes which would place the config files on the nodes [Thu Dec 18 21:41:54 2014] <Kxepal>: robertkowalski: but we also have a single node mode, remember? [Thu Dec 18 21:42:20 2014] <robertkowalski>: good call! [Thu Dec 18 21:42:39 2014] <robertkowalski>: maybe config via web just for single node setups? [Thu Dec 18 21:42:57 2014] <awenkhh>: I don't know the eco system of 2.0 good enough but wouldn't it be great to see the configuration of each node in Fauxton? Read only ? [Thu Dec 18 21:43:22 2014] <awenkhh>: not sure if that's doable at all [Thu Dec 18 21:43:47 2014] <Kxepal>: robertkowalski: that's makes a sense. and disable it when Fauxton is able to detect that it runs on cluster [Thu Dec 18 21:43:51 2014] <awenkhh>: I mean it's an admin interface still where I expect info about my cluster [Thu Dec 18 21:44:05 2014] <robertkowalski>: awenkhh: from the browser (where fauxton lives) we can not post to 1-x (maybe 130) nodes to get a config [Thu Dec 18 21:44:29 2014] <Kxepal>: awenkhh: summary dashboard is an future feature. to make it we should first fix /_stats for cluster iface [Thu Dec 18 21:44:30 2014] <awenkhh>: not in one shot no [Thu Dec 18 21:44:55 2014] <awenkhh>: but you will know which nodes you have somewhere [Thu Dec 18 21:45:18 2014] <Kxepal>: awenkhh: behind haproxy you wouldn't be able reach them [Thu Dec 18 21:45:18 2014] <awenkhh>: and then you can get details per node as requested I imagine ... but that's maybe a discussion for the ML [Thu Dec 18 21:45:39 2014] <strmpnk>: /_healthcheck would be nice but I don't see it blocking early adopters. [Thu Dec 18 21:45:52 2014] <awenkhh>: hm ... as I said ... not sure if know enough yet [Thu Dec 18 21:45:55 2014] <awenkhh>: so nvm [Thu Dec 18 21:46:35 2014] <robertkowalski>: regarding 2.0 taht's it from the fauxton team [Thu Dec 18 21:46:43 2014] <strmpnk>: For what it's worth, we should try hard to get clusters setup easily and correctly from the installation process. Doing it in fauxton is not a requirement. [Thu Dec 18 21:47:23 2014] <robertkowalski>: we are also looking into other options to render our views, and are in love with react.js for our views - garren will post to the ML in the next days. [Thu Dec 18 21:47:42 2014] <robertkowalski>: strmpnk: i agree [Thu Dec 18 21:47:56 2014] <Kxepal>: strmpnk: yes, that's what jan____ /_setup does [Thu Dec 18 21:48:04 2014] <strmpnk>: :-) [Thu Dec 18 21:48:12 2014] <Kxepal>: Fauxton just provides more nicer way to work with it [Thu Dec 18 21:48:20 2014] <strmpnk>: /____jan is a 3.0 API. [Thu Dec 18 21:48:30 2014] <Kxepal>: robertkowalski: are you going to replace backbone with react.js in near future? (: [Thu Dec 18 21:48:40 2014] <Kxepal>: strmpnk: ahaha! great! [Thu Dec 18 21:49:17 2014] <robertkowalski>: we have some issues regarding the classic backbone views (bugs, alot of bugs and complexity) and figuring out how we could iterativley release better versions of fauxton [Thu Dec 18 21:49:31 2014] <jan____>: lol [Thu Dec 18 21:50:18 2014] <Kxepal>: robertkowalski: I like this idea [Thu Dec 18 21:50:23 2014] <robertkowalski>: basically backbone views, creating other subviews create giant trees of views in the current setup we have [Thu Dec 18 21:50:25 2014] <jan____>: /____jan would be a nice easter egg [Thu Dec 18 21:51:12 2014] <robertkowalski>: and fauxton got too complex over the years, we bascially also outgrown our framework and our development process got slower as we want to include more complex features [Thu Dec 18 21:51:12 2014] <awenkhh>: ha ha :D [Thu Dec 18 21:52:06 2014] <robertkowalski>: we will post to the mailinglist very soon but we are aware of the issues you have with fauxton (hi alex ;) ) and we want to improve [Thu Dec 18 21:52:36 2014] <Kxepal>: robertkowalski: hi Robert (: look forward for it! [Thu Dec 18 21:53:17 2014] <robertkowalski>: the nice thing is that we could do that iterativly and don't have to do a full rewrite [Thu Dec 18 21:53:48 2014] <robertkowalski>: </fauxton team> [Thu Dec 18 21:53:49 2014] <Kxepal>: ok, we're beein running for two hours lets run shortly over one last topic and take it over [Thu Dec 18 21:55:03 2014] <Kxepal>: thanks, robertkowalski! that's sounds really good plan. hope you'll make it (: [Thu Dec 18 21:55:20 2014] <awenkhh>: yeah - very cool !!! +1 ## Branding routines: logo, slogan ## [Thu Dec 18 21:56:15 2014] <Kxepal>: that's holywar topic, so I'll be short [Thu Dec 18 21:56:20 2014] <robertkowalski>: garren did all the research! [Thu Dec 18 21:56:33 2014] <awenkhh>: Kxepal oh thats another two hours ;-) [Thu Dec 18 21:56:39 2014] <Kxepal>: we're collecting ideas about new CouchDB logo and slogan on marketing@ ML [Thu Dec 18 21:56:57 2014] <Kxepal>: so if you have anything - join there! [Thu Dec 18 21:57:02 2014] <robertkowalski>: yes, sean barclay did some really cool work! [Thu Dec 18 21:57:12 2014] <awenkhh>: I appreciate the work on it a lot [Thu Dec 18 21:57:17 2014] <robertkowalski>: and sent some proposals [Thu Dec 18 21:57:24 2014] <awenkhh>: the suggestions have become better and better [Thu Dec 18 21:57:36 2014] <jan____>: I like Noahâs latest summary on the why/how/what [Thu Dec 18 21:57:41 2014] <awenkhh>: although I am not sure if we are near a target or miles away [Thu Dec 18 21:57:48 2014] <awenkhh>: jan____ yes ! [Thu Dec 18 21:57:55 2014] <jan____>: The logo suggestions are getting better as well [Thu Dec 18 21:58:11 2014] <awenkhh>: but we are not there imho [Thu Dec 18 21:58:15 2014] <jan____>: but I also feel that just the couch with the name, without the Dude is a good iteration [Thu Dec 18 21:58:25 2014] <Kxepal>: yea, hope finally we'll find our sweet spot [Thu Dec 18 21:58:28 2014] <robertkowalski>: jan____: i also like that! [Thu Dec 18 21:58:28 2014] <jan____>: maybe we can do the more radical rebrand for later [Thu Dec 18 21:58:33 2014] <jan____>: -for [Thu Dec 18 21:58:39 2014] <awenkhh>: good idea [Thu Dec 18 21:58:44 2014] <Kxepal>: jan____: iterative rebranding is also the way to handle it [Thu Dec 18 21:58:45 2014] <awenkhh>: on the other hand ... [Thu Dec 18 21:59:03 2014] <awenkhh>: with 2.0 we would have a massive reachout [Thu Dec 18 21:59:05 2014] <jan____>: THERE IS NO OTHER HAND [Thu Dec 18 21:59:13 2014] <awenkhh>: pah! :D [Thu Dec 18 21:59:17 2014] <Kxepal>: there are only two hands (: [Thu Dec 18 21:59:22 2014] <jan____>: (sorry ) [Thu Dec 18 21:59:31 2014] <jan____>: awenkhh: sure, but we also have a lot to get across [Thu Dec 18 21:59:39 2014] <awenkhh>: on the other side ... [Thu Dec 18 21:59:40 2014] <jan____>: aybe keep the logo for more radical changes down the road [Thu Dec 18 21:59:43 2014] <awenkhh>: ;-) [Thu Dec 18 21:59:44 2014] <jan____>: *maybe [Thu Dec 18 22:00:01 2014] <jan____>: (I have some ideas :) [Thu Dec 18 22:00:04 2014] <awenkhh>: I personally see no need to change the logo at all [Thu Dec 18 22:00:08 2014] <jan____>: how many sides does your hand have? :D [Thu Dec 18 22:00:19 2014] <awenkhh>: but I understand why people would like to change it [Thu Dec 18 22:00:25 2014] <jan____>: (sorry, itâs late here) [Thu Dec 18 22:00:26 2014] <awenkhh>: many many - belive me :D [Thu Dec 18 22:00:33 2014] <jan____>: nah, getting rid of the dude isa good idae [Thu Dec 18 22:00:37 2014] <jan____>: *idea [Thu Dec 18 22:00:42 2014] <awenkhh>: I know [Thu Dec 18 22:00:56 2014] <awenkhh>: but it's hard to get something new [Thu Dec 18 22:01:12 2014] <awenkhh>: I just mean that when we change sth then the 2.0 release would be great [Thu Dec 18 22:01:24 2014] <awenkhh>: changing just a little is not a good idea I think. [Thu Dec 18 22:01:45 2014] <awenkhh>: mybe we leav it for later all together ... but that's a discussion for the ML again [Thu Dec 18 22:01:59 2014] <awenkhh>: *maybe [Thu Dec 18 22:02:25 2014] <Kxepal>: agree, let's discuss it on ML. it's not a simple topic for IRC (: [Thu Dec 18 22:02:26 2014] <jan____>: maybe the couch without the dude that repeats into infinity [Thu Dec 18 22:02:31 2014] <robertkowalski>: maybe we don't need something complete new [Thu Dec 18 22:02:32 2014] <jan____>: but yeah, this is ML material [Thu Dec 18 22:02:37 2014] <robertkowalski>: just iterating on that what we have [Thu Dec 18 22:02:42 2014] <robertkowalski>: in small steps [Thu Dec 18 22:03:01 2014] <robertkowalski>: (for the current timeframe) [Thu Dec 18 22:03:07 2014] <awenkhh>: I am always a bit of a all or nothing guy [Thu Dec 18 22:03:13 2014] <awenkhh>: but let's see .... [Thu Dec 18 22:03:20 2014] <awenkhh>: I like it LOUD !!! [Thu Dec 18 22:03:30 2014] <awenkhh>: with a lot of BANG [Thu Dec 18 22:03:51 2014] <awenkhh>: I guess that's it for this topic ok? [Thu Dec 18 22:04:04 2014] <Kxepal>: no, it was a reminder about (: [Thu Dec 18 22:04:26 2014] <Kxepal>: moving to the last one [Thu Dec 18 22:04:32 2014] <awenkhh>: well ... ;-) ## Happy New 2015 Year ## [Thu Dec 18 22:05:41 2014] <Kxepal>: since this is our the last meeting in 2014 I'd like to congrats everyone with Merry Christmas and happy new year and say that 2014 was amazing year for CouchDB [Thu Dec 18 22:06:14 2014] <Kxepal>: also Lena collecting stuff for weekly news about most amazing things happened in CouchDB universe in this year [Thu Dec 18 22:06:26 2014] <Kxepal>: if you have something, share it with marketing@ ML [Thu Dec 18 22:06:49 2014] <awenkhh>: yeah thanks to everybody for making this an awesome CouchDB year! [Thu Dec 18 22:07:34 2014] <Kxepal>: great work everyone! let's make 2015 even more awesome with 2.0 release and more features! [Thu Dec 18 22:07:49 2014] <awenkhh>: +1 [Thu Dec 18 22:07:55 2014] <Kxepal>: wish you and us the best (: [Thu Dec 18 22:08:40 2014] <Kxepal>: ASFBot: meeting stop Meeting ended at Thu Dec 18 22:08:40 2014
