Summary of IRC Meeting in #couchdb-meeting at Wed Jun 25 19:07:22 2014: Attendees: davisp, nslater, awenkhh, Kxepal, chewbranca, rnewson
- Preface - 1.6.0 released - Link: https://blogs.apache.org/couchdb/entry/apache_couchdb_1_6_0 - welcome new committer - Action: everyone welcome our new committer Lena also known as ffffux (; - bigcouch merge - Link: https://github.com/chewbranca/test_couch_file_migrations - couchdb blog - Info: the best blog is the blog with interesting stuff which people likes to read. if you'd like to write interesting posts about couchdb we're welcome you! contact with us via marketing@ ML - l10 - AdvocateHub - Link: http://www.mail-archive.com/marketing@couchdb.apache.org/msg00539.html IRC log follows: ## Preface ## ## 1.6.0 released ## [Wed Jun 25 19:08:13 2014] <Kxepal>: #link https://blogs.apache.org/couchdb/entry/apache_couchdb_1_6_0 [Wed Jun 25 19:08:38 2014] <Kxepal>: CouchDB 1.6.0 finally released! everyone encouraged to get their couchdb update (: [Wed Jun 25 19:10:00 2014] <Kxepal>: and thanks everyone, especially djc, for making this release ## welcome new committer ## [Wed Jun 25 19:10:48 2014] <awenkhh>: hi [Wed Jun 25 19:11:14 2014] <Kxepal>: #action everyone welcome our new committer Lena also known as ffffux (; [Wed Jun 25 19:11:52 2014] <awenkhh>: ACTION waaaaaaves for ffffux :) [Wed Jun 25 19:13:26 2014] <rnewson>: ffffux: welcome. ## bigcouch merge ## [Wed Jun 25 19:16:00 2014] <Kxepal>: heading to the hot topics. rnewson, chewbranca is there any updates for merge status? [Wed Jun 25 19:16:39 2014] <Kxepal>: ACTION probably topic should be renamed to couchdb 2.0 or something since actually bigcouch is merged [Wed Jun 25 19:17:07 2014] <chewbranca>: I've been working on testing .couch file migrations from 1.1.2 onwards for the latest point releases [Wed Jun 25 19:17:32 2014] <chewbranca>: I've got some generated .couch files up here: https://github.com/chewbranca/test_couch_file_migrations [Wed Jun 25 19:17:56 2014] <Kxepal>: #link https://github.com/chewbranca/test_couch_file_migrations [Wed Jun 25 19:18:13 2014] <chewbranca>: and a hacked together script that generates a db with a variety of doc types in it, copies it into bigcouch on the single node interface and makes sure the results are the same [Wed Jun 25 19:18:35 2014] <chewbranca>: I stopped at 1.1.2 because 1.0.x requires messing around with spidermonkey versions [Wed Jun 25 19:19:16 2014] <chewbranca>: I'm working on doing some final tests with compacting the copied over db in the bigcouch side and still verifying it works, and that will be it for tests unless anyone else has any other ideas of interesting doc types to test [Wed Jun 25 19:19:44 2014] <awenkhh>: chewbranca sounds great - how about testing efforts - shall we start a thread in dev@ to ask people to test? [Wed Jun 25 19:20:20 2014] <awenkhh>: a 'how to' would be cool for that purpose ... I guess [Wed Jun 25 19:20:43 2014] <chewbranca>: awenkhh: the best thing we can do for people to test is to grab a database they have, copy it into the bigcouch local server folder and see if it works [Wed Jun 25 19:21:29 2014] <chewbranca>: although having some people grab the .couch files in that repo and test those would be useful as well, but I think the important bit is making sure we test a larger variety of databases than what I have here [Wed Jun 25 19:21:55 2014] <awenkhh>: chewbranca ok - would you write a call for test :) ? [Wed Jun 25 19:22:05 2014] <chewbranca>: sure [Wed Jun 25 19:22:21 2014] <awenkhh>: chewbranca thank you :) great! [Wed Jun 25 19:22:33 2014] <Kxepal>: chewbranca: does anything is need to be prepared/configured before migration? [Wed Jun 25 19:23:05 2014] <chewbranca>: I think I'll grab make a little script in that repo that lets you run the assertions directly against a db, and just add some instructions for copying into the local bigcouch dev server [Wed Jun 25 19:23:43 2014] <chewbranca>: Kxepal: there's still some questions about the ideal upgrade path from 1.x -> 2.0 [Wed Jun 25 19:24:08 2014] <Kxepal>: chewbranca: compaction vs replication? [Wed Jun 25 19:24:22 2014] <chewbranca>: the file format is different for some of the btrees, which means all dbs need to be compacted [Wed Jun 25 19:24:36 2014] <chewbranca>: Kxepal: well even before that point, right now you can copy in an older .couch file and it "just works" [Wed Jun 25 19:24:51 2014] <chewbranca>: but, it's still not the proper format, because it needs to be compacted [Wed Jun 25 19:25:57 2014] <chewbranca>: the main difference is that #full_doc_info records are stored in the seq_tree rather than #doc_info records [Wed Jun 25 19:26:19 2014] <chewbranca>: so the question is what approach to take for the upgrade there, and it would be good to have more people weigh in on the issue [Wed Jun 25 19:26:51 2014] <chewbranca>: my preference is to do something like bump the disk format version number and enforce a compaction on all dbs before allowing them to be used in 2.0 [Wed Jun 25 19:27:14 2014] <chewbranca>: but it would be good to hear from davisp, rnewson and others as well [Wed Jun 25 19:27:41 2014] <chewbranca>: or the compact issue entirely and just replicate in [Wed Jun 25 19:27:50 2014] <chewbranca>: *or dodge the .. [Wed Jun 25 19:27:53 2014] <Kxepal>: sounds good. it worth to be discussed. [Wed Jun 25 19:28:30 2014] <davisp>: chewbranca: It basically has to be compaction. Replication is a pretty terrible upgrade strategy to throw at people [Wed Jun 25 19:28:36 2014] <chewbranca>: that's all from me, I'll get this compact test finished up, write up a few things and shoot out an email to dev@ [Wed Jun 25 19:28:48 2014] <awenkhh>: may I ask about what time period we are speaking when the merge is completely doen? I ask because we also need to write documentation and we need to start at one point [Wed Jun 25 19:29:37 2014] <chewbranca>: davisp: or far simpler if you can spin up multiple nodes, especially if you want to switch over to clustered with Q > 1 [Wed Jun 25 19:30:04 2014] <chewbranca>: awenkhh: I think rnewson and wohali have a better idea than myself on the full scope of things remaining [Wed Jun 25 19:30:37 2014] <rnewson>: oh, hi [Wed Jun 25 19:30:40 2014] <awenkhh>: ok ... we should keep in mind that we also need documentation - [Wed Jun 25 19:30:44 2014] <davisp>: chewbranca: I don't think spinning up multiple nodes is far simpler than anything [Wed Jun 25 19:30:59 2014] <rnewson>: actually, not so much. davisp and I are going to get together to thrash out a complete list of what's left. [Wed Jun 25 19:31:31 2014] <rnewson>: I will say that it's easy to wedge the bigcouch branch with a replication of a large source db with many attachments. [Wed Jun 25 19:31:32 2014] <awenkhh>: sounds great! [Wed Jun 25 19:31:46 2014] <rnewson>: and that needs looking into because it seems worse than what cloudant experiences. [Wed Jun 25 19:32:05 2014] <rnewson>: though we also have 18 months of fixes to also pull forward in another IP clearance. (and the last) [Wed Jun 25 19:33:34 2014] <davisp>: yay ip clearance [Wed Jun 25 19:34:32 2014] <chewbranca>: davisp: do you have any thoughts on enforcing a compaction into the bigcouch file format that I mentioned above? [Wed Jun 25 19:35:28 2014] <davisp>: Enforcing? [Wed Jun 25 19:35:39 2014] <davisp>: I haven't thought about it [Wed Jun 25 19:35:47 2014] <davisp>: Why would we do that? [Wed Jun 25 19:35:53 2014] <davisp>: Once we read it from disk it'll upgrade in place [Wed Jun 25 19:36:06 2014] <chewbranca>: davisp: yeah, I'm not a fan of allowing older file format versions to be floating in bigcouch, the whole #full_doc_info vs #doc_info thing is a good example of why [Wed Jun 25 19:36:09 2014] <davisp>: chewbranca: there is a huge thing for users that once you open it wiht the new code it can't be read by old code [Wed Jun 25 19:36:43 2014] <davisp>: chewbranca: I dunno how we'd make that forceful without lots of changes [Wed Jun 25 19:37:17 2014] <chewbranca>: one more reason I'm a fan of replicating in [Wed Jun 25 19:37:44 2014] <chewbranca>: true, figuring out how to allow users to trigger compaction and nothing else would be awkward [Wed Jun 25 19:39:14 2014] <chewbranca>: either way, there's some open questions on how to actually approach upgrading a local install that need to get fleshed out [Wed Jun 25 19:39:16 2014] <Kxepal>: replication causes disk usage overhead and manual cleanup things after + two couchdb servers on board. forced compaction is need to be queued somehow since mass compaction of dozen huge databases may cause some issues with undestanding the processes [Wed Jun 25 19:39:23 2014] <davisp>: sure [Wed Jun 25 19:39:32 2014] <davisp>: to both [Wed Jun 25 19:41:17 2014] <Kxepal>: chewbranca: just to ensure the things, you'll start the discussion thread on dev about db upgrade with your thoughts? [Wed Jun 25 19:41:30 2014] <chewbranca>: Kxepal: sure [Wed Jun 25 19:41:46 2014] <Kxepal>: chewbranca: cool! thanks (: how we'll figure something [Wed Jun 25 19:42:14 2014] <awenkhh>: yes thanks to all of you working hard on the merge - awesome!!! ## couchdb blog ## [Wed Jun 25 19:42:53 2014] <chewbranca>: yeah we're getting close, looking forward to 2.0!! [Wed Jun 25 19:43:12 2014] <awenkhh>: ACTION waves for 2.0 :) [Wed Jun 25 19:43:29 2014] <Kxepal>: \o/ [Wed Jun 25 19:43:39 2014] <Kxepal>: so we're making a copy of our blog with fresh design and friendly for writers: http://blog.couchdb.org/ [Wed Jun 25 19:43:48 2014] <awenkhh>: there have been problems with the roller couchdb blog recently [Wed Jun 25 19:44:00 2014] <awenkhh>: so I think this is a really good descision [Wed Jun 25 19:44:03 2014] <chewbranca>: nice! looks good [Wed Jun 25 19:45:39 2014] <Kxepal>: yes, it's a good step forward. now I have to try to migrate the posts from asf roller blog, after we'll make small cleanup and make the announcement of blog.couchdb.org [Wed Jun 25 19:46:07 2014] <awenkhh>: Lena has made a good list of what to do - so yeah Alex I can alos help although my time is limited ... [Wed Jun 25 19:46:27 2014] <awenkhh>: if you tell me what to do I'll jump in [Wed Jun 25 19:47:42 2014] <awenkhh>: the problem is a bit that we need to roll out the weekly news - if the period without actual news is becoming too long people loose interest [Wed Jun 25 19:47:48 2014] <Kxepal>: awenkhh: ok, will keep this in mind, thanks (: also Dave wanted to help with migration [Wed Jun 25 19:48:32 2014] <awenkhh>: we should maybe start a thread to user@ and dev@ that we do have a problem but also a solution [Wed Jun 25 19:48:58 2014] <awenkhh>: better to speak about a problem then trying to hide it [Wed Jun 25 19:49:44 2014] <Kxepal>: if the migration will lasts a bit long, yes, it worth to [Wed Jun 25 19:50:34 2014] <awenkhh>: I don't think we should wait longer with an informal post [Wed Jun 25 19:50:52 2014] <awenkhh>: if there are no objections I am happy to write that email [Wed Jun 25 19:51:26 2014] <Kxepal>: awenkhh: it good to discuss on marketing@ about. at least the text need to be defined [Wed Jun 25 19:51:42 2014] <awenkhh>: yup [Wed Jun 25 19:52:15 2014] <Kxepal>: so that's about our blog [Wed Jun 25 19:52:45 2014] <Kxepal>: oh, btw [Wed Jun 25 19:53:22 2014] <Kxepal>: #info the best blog is the blog with interesting stuff which people likes to read. if you'd like to write interesting posts about couchdb we're welcome you! contact with us via marketing@ ML [Wed Jun 25 19:53:51 2014] <Kxepal>: minute of adv (: ## l10 ## [Wed Jun 25 19:54:48 2014] <awenkhh>: my part then [Wed Jun 25 19:55:01 2014] <awenkhh>: well - I am a bit unlucky about the progress [Wed Jun 25 19:55:39 2014] <awenkhh>: the work is hard, time is limited and there are not many people helping [Wed Jun 25 19:55:54 2014] <awenkhh>: so I have the question how we handle this ... [Wed Jun 25 19:56:00 2014] <Kxepal>: and https://translate.apache.org/ is down at the mean time.. [Wed Jun 25 19:56:27 2014] <awenkhh>: I know that L10n is not the topic prio 1 but would be a shame when the work already done is lost [Wed Jun 25 19:56:59 2014] <awenkhh>: oh yeah it's down - I will get in touch with janI from infra [Wed Jun 25 19:57:29 2014] <awenkhh>: so actually I have no idea how to make better progress or progress at all - I just wanted to mention it and am happy about any ideas that help [Wed Jun 25 19:58:32 2014] <Kxepal>: awenkhh: iirc there was only German localization done by you and robertkowalski . do you recall the status of the others? [Wed Jun 25 19:59:18 2014] <awenkhh>: yes it was Robert and me and I am not sure about the progress of other translators but I am afraid it is not much [Wed Jun 25 20:00:53 2014] <Kxepal>: so we need to use that moment. ping me when t.a.o will get back, we'll setup docs browsing via rtd [Wed Jun 25 20:01:27 2014] <awenkhh>: Kxepal ok ## AdvocateHub ## [Wed Jun 25 20:02:34 2014] <Kxepal>: #link http://www.mail-archive.com/marketing@couchdb.apache.org/msg00539.html [Wed Jun 25 20:03:29 2014] <Kxepal>: nslater published timeline schedule for Influitive AdvocateHub work [Wed Jun 25 20:06:45 2014] <Kxepal>: mmm..couldn't find anything to add about, but if you're would like to help with notice the schedule and get in touch with nslater about the details [Wed Jun 25 20:07:02 2014] <nslater>: as you might've noticed, i plan to do a hangout with you folks [Wed Jun 25 20:07:13 2014] <nslater>: i'll give a presentation of the advocatehub and answer any questions you may have [Wed Jun 25 20:08:30 2014] <awenkhh>: nslater yeah awesome! [Wed Jun 25 20:11:29 2014] <awenkhh>: oh we lost Alex :) [Wed Jun 25 20:14:41 2014] <awenkhh>: well are there any topics left ? [Wed Jun 25 20:16:03 2014] <awenkhh>: ok I'd say that's it for this week ... I will close the meeting then [Wed Jun 25 20:16:22 2014] <awenkhh>: ASFBot: meeting end [Wed Jun 25 20:22:44 2014] <rnewson>: ASFBot: meeting end Meeting ended at Wed Jun 25 20:22:44 2014
