Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2015-05-10

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:16 jimallman jar286: here now, hope things are going well.
00:17 jar286 oh, I’m sure they are… just watching columbo on netflix
00:17 jar286 will try that update now…
00:17 jar286 first, will back up downloads/
00:18 jimallman ah, columbo. it’s been many years since i’ve seen it. does it hold up?
00:20 jar286 well, yes, if you’re into that kind of thing. what I like about 1960s TV is the pace, not so frenetic as now
00:23 jimallman yep, it takes a bit of effort to slow down for lots of older films, too. they seem to linger, or savor, or (in some cases) drag.
00:25 jimallman and then there are the screwball comedies, where the snappy one-liners still come faster than i can process them.
00:26 jar286 it’s taking a long time to copy treemachine.db.tgz.  I guess I don’t have to wait, will do deploy to devapi / devtree in parallel
00:26 jimallman yeah, that’s a huge file. but josephwb still thought it was slower this week. disk (almost) full?
00:27 jar286 no, downloads/ is on the 2TB drive
00:28 * jimallman nods
00:28 jar286 weird, the copy I’m doing is MIT to MIT, in the same rack.
00:30 jar286 cpu bound (rsync)
00:31 jar286 deploying opentree master to devtree now
00:33 jar286 40% copied
00:33 jar286 deployment seems fine (not surprising)
00:34 jar286 I’m copying the database from ot10 to my other server (varela), and will copy it from there to api (ot20?)
00:34 jar286 this could take a while.  the second copy is MIT to EC2 so could take even longer
00:34 jimallman yes, dig says api.opentreeoflife.org is ot20
00:37 jar286 I don’t see much point in doing the production deploy until that database is in place.
00:37 jar286 so I think I’ll go back to watching columbo.  will set a timer for 10 minutes
00:37 jimallman sounds right. ok, i’ll be hanging around…
00:37 jar286 it’s saturday night, you should be out partying
00:38 jimallman i can goof off perfectly well right here!
00:38 jar286 the episode i’m watching has leslie nielson and patrick mcgoohan
00:38 jimallman i had my big adventure for the day: off to the trash collection site, then Wal-Mart, then a tiny sushi place that just opened nearby. now i’m a barnacle.
00:39 jimallman sweet!
00:39 jimallman (that was for Falk + Nielsen + McGoohan)
00:42 jimallman and now i’m sucked into Wikipedia: “The first episode of Columbo as a series was directed in 1971 by a 25-year-old Steven Spielberg in one of his earliest directing jobs.”
00:46 jar286 ok, so 70s, not 60s
00:47 jar286 yes, it’s a remarkable series in ways… I liked it when I came out, and still do, though am reluctant to admit it
00:47 jimallman i believe the original pilot was in 1968, so you’re technically correct. but that’s the best kind of correct.
01:00 jar286 starting 2nd copy…
01:17 jar286 looks like I don’t have a backup copy of the current (sept 2014) database. will have to copy that too.
01:20 jar286 then when I’ve saved that, maybe in 2 places, I’ll delete it from ot20, which is short on space…
02:59 jar286 copies all made. continuing in the a.m.
03:00 jimallman sounds good, i’ll probably be around. g’nite!
03:00 jar286 night
13:23 jar286 joined #opentreeoflife
14:03 jar286 upgarded ot16. rebooting ot16, to make sure that this is still possible
14:03 jar286 s/gard/grad/
14:13 jar286 jimallman, the supporting studies aren’t showing up on dev…
14:14 jar286 checking oti logs
14:16 jar286 last oti init was Aug 23 2014
14:39 jar286 e.g. edge from Theria to Eutheria should show as supported by pg_1428
14:41 jar286 Indexing failed for URL "http://devapi.opentreeoflife.org/phylesystem/v1/../default/v1/study/pg_17.json", study http://devapi.opentreeoflife.org/phylesystem/v1/../default/v1/study/pg_17.json. Message = "Attempt to get the root of an empty NexsonTree object."
15:38 codiferous joined #opentreeoflife
15:59 jimallman jar286: sorry, getting a late start here. yeah, i’m not sure what to make of this message.
16:16 codiferous joined #opentreeoflife
16:39 codiferous joined #opentreeoflife
18:26 codiferous joined #opentreeoflife
20:57 jar286 jimallman, I’m less worried about the indexing problem, which was never tested in the latest round of changes, than I am about the missing supporting studies lists, which was.
20:57 jimallman ah, i assumed the former was causing the latter (since the supporting studies are fetched from oti)
20:57 jar286 unrelated.
20:57 jimallman gotcha
20:57 jar286 well, i don’t know.
20:58 jar286 i expect they’re unrelated
20:58 jimallman only if we’re expecting to get stale (but accurate) information for most supported studies.
20:58 jimallman i’m taking a closer look now at the arguson
20:59 jar286 the study in the example in question, Theria to Eutheria, is pg_1428, and it ought to be in the dev oti index.
20:59 jar286 I can double check that
21:00 jar286 it’s in phylesystem-0…
21:00 jar286 I browsed around a bit and didn’t find any supporting study lists, so I’m pretty sure it’s a logic error
21:01 jar286 hey, maybe your fix didn’t get merged?
21:01 jar286 no… that’s not it… #637
21:02 jimallman i see the problem. we’re trying to fetch supporting-study info from devapi instead of api
21:02 jimallman (this is due to a mismatch between the synthetic tree and the dev* phylesystem)
21:02 jar286 I see, so the two problems *are* related...
21:02 jar286 had I been able to reindex dev oti, we would be seeing the studies lists
21:02 jimallman in a sense. but we can fix it by updating the server-config file (oti base URL)
21:03 jar286 right?
21:03 jar286 hmm, I’d rather do this the right way
21:03 jimallman if you could re-index dev oti, it would still (if i understand correctly) not match the latest synthetic tree, which was built from the production phylesystem
21:03 jar286 fix oti, re-index, then deploy
21:04 jar286 not sure I understand… the trees are identified by study id and tree id, right?  and those are the same in both phylesystems (dev and prod)
21:05 jimallman yes, with the caveat that dev is very stale; uses bare (no-prefix) study ids for pg studies; and missing lots of recently added studies+trees
21:05 jimallman so a search for (for example) pg_2542 comes up empty in dev-oti, but finds the proper study in prod-oti
21:06 jar286 and that wouldn’t be fixed by re-indexing?… I don’t see unprefixed ids in the repo itself https://github.com/OpenTreeOfLife/phylesystem-0/tree/master/study/pg_28
21:06 jimallman i feel this is a temporary mismatch, and the best solution is probably to flush the data in dev-oti and dev-phylesystem and copy the latest production data there.
21:06 jar286 oh… the ids are in the files...
21:07 jar286 sigh.
21:07 jar286 ok, can you commit the config change to master?
21:07 jimallman yeah, i believe someone (mtholder?) swept the production phylesystem and udpated all of these
21:07 jimallman yes, looking at server-config changes now...
21:08 jar286 hang on, there is no unprefixed study id in pg_1428
21:08 jar286 so it looks to me like reindexing ought to work
21:08 jar286 "^ot:studyId": "pg_1428",
21:09 jimallman hmm, i pushed the change to deployed-systems already. but yes, let’s keep looking..
21:10 jar286 that’s ok
21:11 jar286 the problem is that if someone edits a study, the thing will blow up on reindexing that study
21:12 jar286 maybe
21:13 jar286 I guess I can try it, although I’m not sure where the error would show up
21:14 jimallman i doubt the user will see it, since re-indexing is triggered by a post-commit hook on GitHub.
21:15 jimallman we’ll see the result in the GitHub webhooks page for the phylsystem repo (phylesystem-0)
21:15 jar286 invsibility that makes the problem even worse
21:17 jimallman https://github.com/OpenTreeOfLife/phylesystem-0/settings/hooks/2656717
21:17 jar286 thanks!
21:17 jar286 family arriving home, will check back later, sorry to abandon you
21:17 jimallman we’ve talked (several times) about adding notification when this fails.. not sure if mtholder has addressed this.
21:17 jimallman no problem; i’ll be around later, probably
21:24 jar286 joined #opentreeoflife
22:09 jar286 joined #opentreeoflife
22:44 jar286 joined #opentreeoflife

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary