Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-06-10

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

All times shown according to UTC.

Time Nick Message
00:15 josephwb joined #opentreeoflife
00:28 josephwb joined #opentreeoflife
00:39 josephwb joined #opentreeoflife
00:40 jimallman towodo: i figured just a 301 redirect, no?
00:41 jimallman ..or an information page?
00:41 towodo i thought a stub with an explanatory sentence and a link to develoopment curator
00:41 towodo a chance to set expectations
00:42 jimallman agreed. i like the idea of explaining things, esp. since we expect to have a production tool at that URL in the future.
00:42 towodo i.e. replace views/default/index.html
00:43 jimallman it might be best to override the main curation controller (study.py) so that all URLs go to this page for now..
00:43 towodo not sure about automation… I can write a script to make the switch, and I guess put it somewhere in the deploy/ directory
00:43 jimallman just in case people share curation URLs (from us or from their own notes)
00:43 towodo I wouldn't know how to change study.py … let me take a look
00:43 jimallman i just figured we’d treat this as a one-time (manual) intervention
00:44 jimallman happy to help with the controller. you’ll see some redirects there already, mapping actions to non-eponymous views
00:44 towodo yeah, it's reflex that I hate doing anything that's not scripted
00:44 towodo can't find any file named study.py
00:45 towodo what path?
00:45 jimallman https://github.com/OpenTreeOfLife/opentree/blob/master/curator/controllers/study.py#L41
00:46 jimallman this is the case i was thinking of… not a redirect per se, but a diversion to another view
00:46 jimallman i hear you on automation. i’d treat this as a branch then, with the necessary mods in the curation app, and deploy that branch to production.
00:47 towodo oops. i had typed the wrong shell command, 'cd' instead of 'find'. ...
00:47 towodo I understood once how this kind of thing works
00:48 towodo you're thinking replace the entire study.py ?
00:48 towodo not sure with what
00:49 jimallman i was thinking that  and the default’ controller, just modify all existing actions to redirect to the info-and-link page
00:49 towodo that's a lot of editing.  what i had in mind is a two-liner
00:50 jimallman i see. apache is probably the point of greatest leverage, just on that /curator path.
00:50 towodo i guess we have multiple entries into the curator app… there's a link from the browser, yes?
00:50 towodo yes, i think i'm agreeing with you now
00:50 jimallman yes, but that goes (as expected) to default/index, the study list
00:51 jimallman if we were quite confident that nobody will use a “deep link” that would suffice. i’m not confident.
00:52 towodo hmm. apache it is. it can redirect to the ad hoc info page
00:53 jimallman ok. shall i whip up a bare-bones info-and-link page with the OToL header? or do you want to go truly bare-bones?
00:53 towodo i think truly bare bones…
00:54 jimallman cool. let me know if i can help. otherwise i’m working through some GitHub issues (incl. making optional deployment requirements truly optional, as you suggested).
00:55 jimallman towodo: we haven’t talked in detail about how/when to move changes (esp. bug fixes) to production branches. would you prefer to freeze these for now, or go ahead and fix known issues?
00:57 towodo you mean in the webapp (not curator or api or oti)?
00:57 towodo i'm sort of … miffed… that we had these two releases… nobody really warned me
00:58 jimallman ah, good point. i guess the webapp is the only visible piece on production.
00:58 jimallman two releases?
00:58 towodo api and browser/tree
00:59 jimallman i’m sorry, not sure what you mean. have we deployed more recent changes than you expected?
01:00 jimallman i’ll admit to being a little dizzy at the number of domains and branches at the moment, but i’m getting my sea legs now.
01:00 towodo i thought we were going to do a launch on publication.  instead we had the browser/tree launch (for reviewers), and the api launch (for hackathon), and both came as surprises to me
01:03 towodo you know, i wonder if you could do the disable-curator hack, now that i think of it… i've got to revive the RAID array tomorrow, and start setting up the new dev system
01:03 towodo is that ok? the apache stuff is simple and i've added a deploy command called 'apache' that just restarts apache (after copying the config file)
01:04 jimallman sure, i’ve actually done lots of apache-config wrangling over the years.
01:04 jimallman i can do this tonight.
01:04 jimallman towodo: ^    (and now i must break for dinner)
01:04 towodo the link can go to 'devtree' which is all set up except maybe for the github setup
01:05 towodo oh wow, devtree works
01:05 jimallman ok. shall i register the webapp and curator apps on github? or have you done this already? (or have i and forgotten?)
01:05 towodo hmm… curator seems to not work there...
01:06 jimallman ok, a few possible causes. i can probably chase that down.
01:06 towodo I've registered something. would have to check to see what I did
01:06 jimallman is the apache config under version control? if not, i’ll probably make a <samename>.ORIG copy of the file
01:07 towodo apache config is under version control. it's in the opentree repo
01:07 jimallman cool! i’ll chase down the registered app, or (re)create one as needed. i’m slightly hampered by not having admin privileges on the GitHub org (OpenTreeOfLife), but so far we’ve worked around that by just transferring ownership of registered apps.
01:07 towodo it's copied every time you do a push.sh command
01:07 jimallman oh, of course!
01:07 towodo it's in the deploy/setup dir
01:07 jimallman i knew this but forgot
01:07 towodo I think I know what to do. Get as far as you can tonight I guess, send me status, and I'll try to finish up in the morning
01:08 jimallman will do!
01:08 towodo but maybe not turn off ot3 curator app until the ot11 one works.  (I'll be changing ot11 to point to ot10 instead of ot12, but the content will be identical)
01:09 towodo thanks a bunch
01:09 towodo we really need tests.
11:58 mtholder joined #opentreeoflife
12:28 towodo joined #opentreeoflife
13:08 kcranstn joined #opentreeoflife
13:26 blackrim joined #opentreeoflife
13:28 kcranstn ah, I see we are bouncing the dev-curator: http://dev.opentreeoflife.org/opentree/static/bounce-to-devtree-curator.html
13:28 kcranstn is this one pointing at the most recent phylesystem? http://devtree.opentreeoflife.org/curator
13:29 mtholder phylesystem-0 (testing repo)
13:29 mtholder and it may have trouble pushing to GitHub. I have not pulled from pg in a while.
13:30 mtholder i just forwarded you an email from jim this morning about the curation redirection.
13:30 kcranstn trying to decide when I can plan to send out an email to the curators about moving to Curator from pg
13:30 kcranstn (thanks for the mail)
13:34 kcranstn is there a plan for pulling from pg and making devtree-curator point to phylesystem?
13:34 kcranstn or deploying a non-dev curator app
13:36 kcranstn @jimallman - it would be great if the curator app listed the underlying github repo
13:36 mtholder I think that we need to decide where we are going to deploy first.
13:36 kcranstn sure
13:37 mtholder towodo was talking about a new instance yesterday.
13:37 towodo i thought it over and you convinced me that using ot10 was ok
13:37 towodo for development
13:37 mtholder cool.
13:37 towodo so no new server, but ot10 is being repurposed
13:38 mtholder perhaps, I'll sync the phylesystem-1 (real repo) with pg and then deploy it on ot10
13:38 mtholder then we won't have two instances using the same repo.
13:38 towodo i was getting close to the point of redeploying but i'll wait until your sync.
13:38 mtholder (phylesystem repo that is)
13:38 kcranstn let me know if I can help
13:39 towodo I was just updating the config files (deployed-systems)
13:39 towodo just pushed an updated opentree-servers list that I believe reflects the new world view
13:40 kcranstn @towodo - are you and I talking at 11?
13:40 mtholder is ot10 big enough for treemachine, taxomachine, + phylesystem
13:41 towodo yes
13:41 mtholder cool
13:41 towodo it's bigger than ot12 and ot12 handles them all just fine
13:41 kcranstn cool - perhaps we can walk though the SSL activation
13:42 towodo yes - hang on - about dev curator -
13:42 mtholder yes
13:42 towodo we bounce to the curator tool on ot11 / ot12
13:42 towodo this is temporary.  soon it will be ot11 / ot10
13:42 mtholder we point ot11's curator at ot10
13:42 towodo because ot12 is 'production' and curator tool is 'development'
13:43 towodo someone changed that.
13:43 mtholder it is sort of next-release candidate, right?
13:44 towodo yes I was hoping to avoid having distinct development and staging servers. (but we can do that if we need to)
13:44 mtholder I think production vs dev is better (just 2 not 3), but it seems like the manuscript being frozen at a different version than the hackathon api means we need 3
13:44 towodo ok, so ot11 goes to ot10, and both are development.  so I just need to make ot11 go to ot10 for tree and taxo as well, and we'll be able to update synthetic trees
13:45 mtholder sounds good.
13:45 towodo I'm just saying that ot12 is part of production.
13:45 towodo so production has redundant copies of tree & taxo.
13:45 mtholder sounds good. (but it is not on the production branch)
13:45 towodo api goes to ot12 for tree & taxo, browser goes to ot9 for tree & taxo.
13:46 towodo right, i guess that should be fixed.. that will be a pain, i can do it later.   I want to do ssl now
13:46 towodo kcranstn, you still there?
13:46 kcranstn yup
13:46 towodo I need to generate a CSR but I don't know how 'wildcards' are handled at namecheap...
13:47 towodo and I read the docs and they don't say...
13:47 kcranstn this: https://www.namecheap.com/support/knowledgebase/article.aspx/467/39/
13:47 towodo so the info I need would have to come from looking at the activation dialog on their site...
13:47 kcranstn hangout so we can screenshare?
13:47 mtholder I'll be offline for ~30 minutes, then I'll start syncing phylesystem-1 with pg...
13:48 towodo I know how to make a csr...
13:48 towodo just a question of whether to put a '*' or a particular host name
13:48 towodo sure we can screen share
13:48 towodo if we need audio i need to move
13:49 kcranstn no audio.
13:49 kcranstn video + chat ok
13:51 towodo right.
13:51 towodo no specific directions, so i will assume we generate a separate csr for each host name.
13:51 kcranstn I am over at g+
13:51 kcranstn screensharing
13:51 towodo yes i see
13:52 towodo thanks
13:52 kcranstn need anything else/
13:52 towodo that was the information i needed.
13:52 kcranstn ok
13:52 kcranstn server type?
13:52 towodo oh right, what are the options?
13:52 kcranstn apache + openSSL?
13:53 towodo yes.  funny, i'm not seeing that on the screen share
13:53 towodo i will prepare a csr and email it to you
13:53 towodo we need two actually, one for devtree and one for devapi
13:53 towodo generating a csr will take me a few minutes
13:53 kcranstn ok
13:53 kcranstn signing off g+, then
13:54 josephwb joined #opentreeoflife
13:54 towodo ok. generating key now, then will go through csr prompts
13:57 lcoghill joined #opentreeoflife
14:01 towodo Organization Name (eg, company) [Internet Widgits Pty Ltd]:
14:01 kcranstn National Evolutionary Synthesis Center
14:02 jimallman kcranstn, re: showing repo in curation app.. i’ll see if i can show this and more (API and other servers) in the web apps.. maybe somewhere discreet, like the footer?
14:02 kcranstn @jimallman - thanks!
14:02 kcranstn footer is perfect
14:03 jimallman i like this idea. helps to illuminate the architecture (for the curious) and keep us (me) oriented.
14:04 towodo kcranstn, CSR emailed to you, for activation.
14:04 kcranstn ok
14:04 kcranstn thanks
14:04 towodo maybe we should see if this one works, before requesting the second one
14:04 towodo (need a separate one for devapi)
14:04 josephwb @jimallman: is there a way for curators to run the 1) treemachine and 2) MIAPA validators, or is this a scheduled thing?
14:04 towodo (then later we will need tree & api)
14:05 kcranstn “ Invalid Wildcard CSR
14:05 kcranstn The CSR provided is invalid to generate a wildcard certificate. Please make sure that you generate a CSR with a Common Name (CN) that starts with '*.' (ex: *.yourdomain.com)."
14:05 towodo ah!
14:05 towodo that's what i was trying to find out before.  AFAICT it's undocumented
14:05 kcranstn who needs documentation when you have error messages ;)
14:05 towodo ok, I'll make another one
14:06 towodo the email address i'm using is certmeister@opentreeoflife.org , which I set up to forward to me
14:06 kcranstn ok
14:09 towodo emailed...
14:09 kcranstn approver email? admin@opentreeoflife.org?
14:10 towodo I don't know...
14:10 towodo oh.
14:10 kcranstn probably doesn’t matter
14:10 towodo it does
14:10 towodo i'm trying to remember...
14:11 towodo I think there wil be followup email and it might be important.  i'd set it to an address that definitely goes to you, then you can forward to me if it matters, ok?
14:14 towodo it might be that the cert gets emailed to you, or that a url or password that lets you access the cert gets emailed to you
14:15 kcranstn no matter what email I pick, I am getting an error message
14:15 towodo hmm.
14:15 kcranstn We are unable to validate the settings of the approver email you have chosen. This may be an indication of incorrectly setup MX records, invalid mailbox, in-accessible mail server, invalid domain etc.
14:15 kcranstn We strongly recommend that you fix the issue above before proceeding. If you are certain that the above findings are invalid and you still want to proceed, please choose the option below.
14:16 towodo let me think
14:17 towodo weird.  you tried gmail? nescent? opentreeoflife.org?
14:17 kcranstn yes, tried gmail and opentreeoflife.org
14:17 towodo maybe it's trying to do some kind of handshake that's not generally supported. it's important enough i'd check their support knowledgebase. will do so now
14:18 kcranstn too late...
14:18 towodo ok...
14:19 kcranstn got confirmation, now just waiting for SSL
14:20 kcranstn forwarded to you
14:20 towodo cert gets emailed to that address. https://www.namecheap.com/support/knowledgebase/article.aspx/794/67/how-to-activate-ssl-certificate
14:23 mtholder joined #opentreeoflife
14:26 mtholder josephwb: (re query about validators). validation by peyotl happens with every save. no other validation is supported AFAIK
14:32 josephwb mtholder: thanks. do you know why this study fails? http://devtree.opentreeoflife.org/curator/study/edit/2926
14:35 josephwb @blackrim: http://treemachinerecords.com
14:38 jimallman josephwb: i’m able to save this (after trivial changes) in curation app, with no visible errors. how do you know it’s failing?
14:39 blackrim josephwb: nice site. they were probably before us i suppose
14:40 josephwb jimallman: from the "Tools" tab
14:40 josephwb "This study fails" the treemachine test
14:41 jimallman D’oh! My bad, that’s placeholder text (in both Tools) meant to demonstrate that they *could* have their own validation logic. I’ll remove this today.
14:42 jimallman FWIW, i see a bogus warning ‘UNRECOGNIZED_KEY’ … apparently the validator isn’t aware of study[‘^ot:agents’]
14:43 * jimallman is adding a peyotl ticket for this now...
14:43 josephwb ok, thanks
14:51 mtholder jimallman: that will be an easy fix. I'm in the middle of something else, but will get to it soon.
14:52 jimallman no sweat, just a little housekeeping
15:09 kcranstn joined #opentreeoflife
15:24 kcranstn why does this show up a curator = anonymous? http://devtree.opentreeoflife.org/curator/study/view/ot_19
15:25 mtholder I think that is a test created via the cmdline with the bogus testing GitHub OAUTH token
15:26 kcranstn temi is the developer at atlas of living australia
15:26 mtholder though I'm thinking of changing my name to anonymous to make life more interesting (and increase my CV length)
15:26 mtholder Ah. Perhaps his/her profile has no name info... I'll check
15:28 mtholder yeah. I think that is it. We get name from GitHub profile. If it is empty, we substitute anon. We could change that behavior.
15:28 kcranstn https://github.com/temi
15:28 kcranstn yup
15:29 mtholder Actually I'm not seeing "anonymous" as a default value on the server side...
15:37 mtholder towodo, jimallman: I'm about to redeploy a bugfix to api on ot10. this will also entail switching the branches for oti, phylesystem-api and opentree to "master" in that config (and putting that config file back in the development dir of the deployed systems repo). Everyone OK with that?
15:39 jimallman This matches my understanding of things, yes. Though I’m concerned that ot10 might be serving both production and development at this point…?
15:40 mtholder I don't think that it is used by ot3, is it? does the browser use oti?
15:40 mtholder the curator app now redirects.
15:40 mtholder sorry browser == opentree/webapp
15:41 jimallman ah, right. lemme check production configs…
15:42 mtholder I think that the configs may indicate that we look to ot10 for api and oti, but I don't think that ot3 uses those features.
15:43 jimallman pretty sure you’ve got it right. the tree-viewer (webapp) shouldn’t need oti or phyesystem-api.
15:43 jimallman i thought maybe it used oti for digging up source-tree or -study info, but that seems far-fetched.
15:44 mtholder I stand corrected. opentree/webapp/views/about/references.html
15:44 mtholder <a target="_blank" href="{{= oti_domain[:-4] }}/api/v1/study/{{= study.get('ot:studyId')}}">Fetch study NexSON</a>
15:45 mtholder perhaps we can just remove that link in the production version of ot 3 ?
15:47 mtholder or leave it. It is unlikely that we'll tolerate a version of the code on ot10 that does not support that basic v1/study/### operation.
15:47 jimallman https://github.com/OpenTreeOfLife/opentree/blob/9ddb7587d461234b8aaf761a92137db3183986a0/webapp/controllers/about.py#L47
15:48 mtholder well damn
15:48 jimallman here i believe we use oti to gather source studies from synthesis. this feeds the Bibliographic References page.
15:49 jimallman no! i’m mistaken… this uses treemachine:
15:49 jimallman https://github.com/OpenTreeOfLife/opentree/blob/8e55a6384792a52c9a34c4935d3b2fcc038e8ea4/webapp/private/config.example#L46
15:50 jimallman but the following method ‘findAllStudies_url’ uses oti… lemme see if this is used anywhere
15:51 jimallman ah yes, we do use oti to retrieve DOIs and full reference text for the Bib.Refs page
15:51 mtholder I don't think that treemachine has the refs....
15:51 blackrim josephwb: nice site. they were probably before us i suppose
15:51 jimallman https://github.com/OpenTreeOfLife/opentree/blob/9ddb7587d461234b8aaf761a92137db3183986a0/webapp/controllers/about.py#L63
15:51 mtholder damn and damn again.
15:52 mtholder towodo, did you see all this ^ ? looks like we do need another machine...
15:52 mtholder and I like ot10 so much.
15:52 * mtholder sobs uncontrollably
15:52 towodo no i didn't see all this.
15:53 mtholder the opentree/webapp uses oti
15:53 mtholder so we can't steal ot10 for development purposes.
15:53 towodo you mean production opentree/webapp …? we can point it at ot12
15:54 towodo needs to be tested, but I know no reason why it shouldn't do the job
15:54 mtholder that should work...
15:54 mtholder it uses the old urls, bu they are still maintained with the new mapping, right?
15:54 towodo for example?
15:55 towodo ot12 = api.opentreeoflife.org
15:55 mtholder .../api/v1/study... instead of .../phylesystem/v1/study
15:55 mtholder ot3 will use the api call
15:55 towodo the old urls should still work
15:55 towodo the old paths that is.
15:56 towodo i didn't change the web2py app name for the [phylesystem] api
15:56 towodo I was just about to do some shuffling in deployed-systems but you can go ahead
15:56 mtholder OK. So I'll point ot3 at ot12 for oti and api. and then deploy the master (with bug fix) on ot10.
15:57 towodo ot12.config  needs to be moved from development/ to production/ and ot10.config needs to be moved from production/ to development/
15:57 towodo not sure about branch choices for ot12.  could set to production but important stuff might be missed
15:57 towodo i would test ot12 first to make sure it's working for webapp purposes
15:58 towodo ot10.config needs to have its branch choices changed probably
15:58 towodo for a minute I was worried we had hardward problems on ot10
15:59 towodo i need to go work on the RAID.  will try to check in before NESCent meeting which is 2-4 but if I don't answer immediately these 2 things are the explanation
15:59 mtholder I'm checking ot12's suitability for the webapp. /about/references looks OK. jimallman, do you think that is the only spot?
16:00 jimallman yes, any service URLs used should be listed in the private/config file, and that’s the only use of oti I can see.
16:01 mtholder great. thanks. I'm first going to redeploy ot11 to point to ot12 (to make sure that ot12's api and oti do what the webapp needs). then I'll proceed.
16:03 jimallman mtholder,kcranstn: meanwhile, re: the “ANONYMOUS” curator name… the curation app tries to build a sensible name from the GitHub profile. I didn't think to add their simple username like 'jimallman', so it jumps straight to 'ANONYMOUS' instead of using 'jimallman' as a fallback for logged-in curator. I'm adding this now..
16:03 towodo there's some ambitguity as to whether when you deploy the opentree repo locally is refreshed from the production branch or the development branch… I would look at the commit log to try to anticipate confusions
16:04 mtholder the key is that opentree is on the master branch when I deploy for development machines, and opentree is on the production branch when I deploy on ot3, right?
16:09 towodo yes, that's right… in theory… I haven't tested the theory though, and i'm paranoid
16:09 towodo i guess jim has been doing it, right jimallman?
16:09 mtholder opentree on ot11 is being redeployed now..
16:10 mtholder compiling NCL (for curator app) which is a bit odd as that should not need to be recompiled...
16:10 mtholder god I hate C++ compilation times....
16:11 jimallman towodo: still trying to parse this… “whether when you deploy the opentree repo locally is refreshed from the production branch or the development branch”
16:11 mtholder you break ot12 if your local copy of opentree repo is "production"
16:12 mtholder that is what I did this weekend.
16:12 mtholder so there is some dependence on what version of opentree repo you have locally checked out (because the deploy scripts there get copied to the server).
16:12 jimallman if i understand you correctly, you’re using the push/deployment tools for local testing. i haven’t been doing that… just pulling/compiling repos manually.
16:13 jimallman i see what you mean.. the local branch of opentree will determine some deployment details. that’s why i made the funky ’deploy-production-with-curator-bounce-to-devtree' branch of opentree last night.
16:14 towodo I messed up typing that comment. let me try again
16:15 towodo "whether, when you deploy [to ot11], your local opentree repo is refreshed from the production branch or the development branch" - goes to contents of deploy/setup directory
16:15 towodo i don't think there's an issue but i like to think these things through and just haven't finished doing so
16:16 towodo because deploy/setup is copied out from the local repo every time you run push.sh
16:16 jimallman agreed.. is that what you mean by “refreshed”?
16:16 towodo even if the config file specified a different branch
16:16 jimallman right
16:16 towodo refreshed = checked out
16:17 mtholder hmm. changing OTI_BASE_URL=http://ot10.opentreeoflife.org/oti to OTI_BASE_URL=http://ot12.opentreeoflife.org/oti on ot11 was not a happy thing
16:17 mtholder reverting...
16:17 jimallman the local opentree repo (incl. push/deploy tools) doesn’t change in response to the specified branches in server-config files, no. ie, if a server-config specifies opentree ‘production’ branch, that won’t affect your (pusher’s) local opentree
16:20 mtholder OK ot11 is functioning again. I think that I messed setting both OPENTREE_API_BASE_URL and OTI_BASE_URL hopefully that is why it wasn't happy.
16:20 mtholder meaning that I should have set both. stand by while I try that....
16:22 jimallman mtholder: no, something’s borked on ot12/oti...
16:22 jimallman note that this works (curl example) on ot10:
16:22 jimallman curl -X POST http://ot10.opentreeoflife.org/oti/ext/QueryServices/graphdb/findAllStudies
16:22 mtholder that would certainly explain it...
16:23 jimallman but fails with a Java “NoClassDefFoundError” on ot12:
16:23 jimallman curl -X POST http://2.opentreeoflife.org/oti/ext/QueryServices/graphdb/findAllStudiesies
16:23 towodo that last is misspelled
16:24 mtholder ok. I guess I'll redeploy oti on ot12 first...
16:24 jimallman re: misspelling, it’s a shell glitch. the real URL also fails:
16:24 jimallman curl -X POST http://ot12.opentreeoflife.org/oti/ext/QueryServices/graphdb/findAllStudies
16:25 mtholder yeah. thanks. fails for me too.
16:25 mtholder redeploying oti on ot12
16:25 * mtholder still sobbing uncontrollably
16:26 jimallman possible clues… i can see the findAllStudies method here:
16:27 jimallman curl  http://ot12.opentreeoflife.org/oti/ext/QueryServices
16:28 jimallman …and here (self-documenting GET request):
16:28 jimallman curl  http://ot12.opentreeoflife.org/oti/ext/QueryServices/graphdb/findAllStudies
16:29 mtholder redeployed oti to ot12. didn't help...
16:29 * jimallman passes a tissue  *honk*
16:30 mtholder towodo: any thoughts on why oti is broken on ot12
16:31 mtholder the GET works, but PUT fails with
16:31 mtholder "message" : "org/opentree/oti/QueryRunner",
16:31 mtholder "exception" : "NoClassDefFoundError",
16:31 mtholder "fullname" : "java.lang.NoClassDefFoundError",
16:31 mtholder "stacktrace" : [ "org.opentree.oti.plugins.QueryServices.findAllStudies(QueryServices.java:37)", "java.lang.reflect.Method.invoke(Method.java:606)", "org.neo4j.server.plugins.PluginMethod.invoke(PluginMethod.java:57)", "org.neo4j.server.plugins.PluginManager.invoke(PluginManager.java:168)", "org.neo4j.server.rest.web.ExtensionService.invokeGraphDatabaseExtension(ExtensionService.java:300)", "org.neo4j.server.rest.web.ExtensionService.
16:31 mtholder invokeGraphDatabaseExtension(ExtensionService.java:122)", "java.lang.reflect.Method.invoke(Method.java:606)", "org.neo4j.server.rest.security.SecurityFilter.doFilter(SecurityFilter.java:112)"
16:31 jimallman is QueryRunner the missing class? (is ot-base up to date?)
16:31 mtholder I just ssh-ed in and ran a git pull origin master on it, no changes came down.
16:32 mtholder perhaps I need to trigger a rebuild of it.
16:32 jimallman https://github.com/OpenTreeOfLife/oti/blob/master/src/main/java/org/opentree/oti/QueryRunner.java
16:32 towodo oh yeah, there is the problem that taxomachine doesn't get rebuilt, but I didn't think oti was vulnerable
16:33 mtholder does that mean that I should just deploy taxomachine ?
16:33 towodo (still need to do 'make'-based installations for deployment, I've made a start on that but not fully working yet)
16:33 towodo well you're not using taxomachine so shouldn't matter.
16:33 towodo just wondering if there was an analogous failure on oti.
16:34 mtholder i meant as a hack to fix it
16:34 mtholder is there a workaround?
16:34 towodo you ssh in and re-run maven (using the shell script)
16:34 jimallman maybe trash the compiled results on ot10?
16:35 jimallman then re-deploy, i mean (to force a recompile)
16:36 mtholder i'm on ot12 for this. several stacks down in my deployment hell.
16:36 mtholder currently trying to get ot11 pointed at ot12 to work (for browsing about/references).
16:36 mtholder if that works then I'll point ot3 to ot12 and point ot11 back to ot10,
16:37 mtholder then i'll redeploy the bug fix to peyotl from 2 hours ago on ot10.
16:37 mtholder then I'll drink heavily.
16:38 jimallman mtholder: there are three likely-looking shell scripts in the root of oti repo… can’t recall which is the most likely to do the job:
16:38 jimallman setup_oti.sh
16:38 jimallman mvn_serverplugins.sh
16:38 jimallman mvn_recompile_install_restart_oti_server.sh
16:38 mtholder I'm looking at that too.
16:38 jimallman the last looks promising
16:38 mtholder it has a nice long name
16:39 jimallman yeeesssss
16:40 mtholder README indicates that ./setup_oti.sh --recompile-plugin is the winner. I'll try...
16:41 mtholder that one started asking me questions and I got scared and aborted it.
16:41 mtholder I'll try mvn_recompile_install_restart_oti_server.sh
16:41 jimallman if he gives you any trouble, shoot the .JAR file
16:42 * jimallman is all about Thug Life, now that i’ve got my parrot
16:43 mtholder ooo, we've got a bad-ass over here...
16:46 mtholder I manually moved the newly compiled oti jar to its home and still no dice.
16:46 mtholder I'll see if I get the same behavior when I compile oti on my laptop
16:47 * mtholder lists the things he hates today as java, c++, maven, and life
16:51 jimallman life gives you bacon.. those other things, not so much
16:54 towodo mtholder, you picked up where i left off: ot11 + ot12
16:54 towodo I was waiting on a treemachine bug, and that got fixed
16:54 mtholder oti is not happy.
16:54 mtholder I can't get it to compile on my local machine - looks like I have an old taxomachine jar
16:55 towodo um. I think that's the wrong shell script
16:56 towodo well, i never used it, that is.
16:56 towodo the deploy scripts use mvn_serverplugins.sh
16:57 mtholder that worked but I had to blow away my old ~/.m2/repository/org/opentree dir.
16:59 towodo right, oti depends on taxomachine… that one threw me
16:59 towodo just remembered: the shell scripts all blow away any existing jar file
17:00 towodo and the jar needed as a dependency is different from the one needed as a plugin...
17:02 towodo the shell scripts all blow away any existing jar file --> in the target/ directory that is
17:02 towodo maybe that's not a problem.
17:03 towodo ok, it's not a problem because the plugins .jar gets copied to the neo4j plugins directory, and the library .jar gets copied into the oti maven .jar
17:03 towodo so false alarm.
17:03 mtholder OK. it is working on my machine. I'm tempted to move ~/.m2/repository/org/opentree  on the server and recompile
17:04 mtholder that seems to have worked for me.
17:04 mtholder (on my laptop)
17:04 towodo so you don't want to go through github...
17:04 mtholder ?
17:05 towodo you are going to commit, push to github, then run push.sh ?  or just do a raw copy ?
17:05 mtholder I haven't changed any code.
17:05 towodo oh.
17:05 mtholder just blew cached jars away and recompiled.
17:05 towodo we can do that on the server too.
17:05 mtholder I'll try. I'll move the old m2 cache rather than deleting it.
17:06 mtholder here goes...
17:06 towodo oh… I see why… it's because Cody & Stephen don't version their stuff
17:06 towodo all of their .jar files are version 0.0.1
17:06 mtholder that would make sense.
17:07 towodo the version number is in each pom.xml.  We need a script to bump it on commit, or something...
17:08 towodo there's nothing precious on ot11.  it can be re-created easily
17:08 mtholder I'm on ot12
17:08 mtholder oti was broken there
17:08 mtholder and still is.
17:09 towodo oh. sorry
17:09 mtholder here is what I did "mv .m2/repository/org/opentree dotm2-repository-org-opentree" and then redeployed. no dice.
17:10 mtholder maybe I need to manually pull taxomachine changes...
17:10 towodo you need to force recompilation of taxomachine
17:10 towodo delete the .jar file and redeploy, I think
17:10 towodo (will check)
17:11 mtholder trying that now.
17:12 towodo the logic around taxomachine is all messed up.
17:12 towodo it's not going to work… the test isn't sensitive enough
17:12 towodo if git_refresh OpenTreeOfLife taxomachine || [ ! -d ~/.m2/repository/org/opentree/taxomachine ]; then
17:13 towodo the test should be fixed.  what's there now is an ineffective sledgehammer.
17:13 mtholder so git pull from opentree and try again?
17:13 towodo no that won't work yet, i haven't made the change
17:14 towodo oh. if you cleared out .m2 then the test will pass
17:16 towodo if you cleared .m2, then taxomachine library will get rebuilt … but not clear that oti will get rebuilt.  also need to remove the .jar from the neo4j-oti/plugins directory
17:16 towodo to force creation of an oti linked with new taxomachine.
17:16 towodo that should do it.
17:16 towodo if git_refresh OpenTreeOfLife $APP || [ ! -r neo4j-$APP/plugins/$jar ]; then
17:17 towodo mtholder, gotta get lunch before my 2:00 meeting. will check in in maybe 20 mins ok?
17:18 mtholder it works thanks!
17:19 mtholder ps I still hate maven
17:25 mtholder OK verified that ot12 works wrt serving the needs of opentree/webapp
17:26 mtholder I tested that on ot11, but I'm about to point ot11 back to ot10
17:26 lcoghill left #opentreeoflife
17:31 towodo I hate maven too.
17:33 mtholder I know that you (and all right thinking people) do.
17:40 mtholder jimallman: Alas, I fear that I have broken your very nice redirect from ot3's curator to devtree's curator app.
17:41 jimallman …ok, hold on...
17:41 jimallman did you switch opentree to the “special” branch before deploying?
17:42 mtholder no. sorry. I'll do that.
17:42 towodo why not just merge special to production? we can always undo it later
17:42 jimallman ’deploy-production-with-curator-bounce-to-devtree' branch of opentree
17:42 mtholder OK. I'll do the branch merge..
17:43 jimallman towodo: OK, i can do that. still a weird wrinkle, since we ordinarily wouldn’t need to switch *local* opentree to production, just to push production… right?
17:43 towodo well that's what I'm not sure of.
17:43 mtholder I think that we do have to do that.
17:43 * jimallman stands back while mtholder merges branches
17:43 mtholder Or at least make sure that when pushing to ot12 we use master of opentree..
17:44 towodo if what you say is true, then the whole deploy/ system is in the wrong repo, should be in deployed-systems.
17:44 towodo but that can't be right.
17:44 jimallman NOTE that i also specified the special branch in the server-config for ot3, since it includes the tiny redirect page. so please reset that to ‘production’ as well.
17:44 towodo oh!
17:45 towodo well as long as the local repo and the branch are close, there should be no problem.
17:45 jimallman https://github.com/OpenTreeOfLife/deployed-systems/blob/master/production/ot3.config#L28
17:45 jimallman after merging the special branch to production, this should also revert to simply ‘production’
17:46 towodo what I wonder about is using local development-ish opentree repo to push to production-ish server, or vice versa.
17:47 towodo if the deployment system had less knowledge of the suite of apps it wouldn't be an issue
17:47 mtholder I merged opentree's deploy-production-with-curator-bounce-to-devtree to production and pushed to GitHub
17:47 jimallman the deployment tools should be orthogonal to the deployed systems, at least they were until my recent tweaks in the special branch. it does feel weird having them gummed together in the ‘opentree’ repo
17:47 mtholder see https://github.com/OpenTreeOfLife/opentree/tree/production
17:48 towodo I dislike the proliferation of repos. But I see what you mean.
17:48 jimallman mtholder: understood. for now (because of the curation redirect), we’ll need to deploy production *from* the production branch, and deploy development from master branch of opentree.
17:48 mtholder I have changed "opentree_branch opentree production" in the ot3 config. and I have production version of opentree repo.
17:48 mtholder I'll redeploy opentree now...
17:49 jimallman i tried to make this a domain-specific tweak (to avoid redirecting the curation URLs on devtree), but that would require mod_rewrite, so i bailed from adding more complications.
17:49 kcranstn joined #opentreeoflife
17:50 towodo looking at https://github.com/OpenTreeOfLife/opentree/network
17:51 mtholder dp you want me to hold off on the deploy?
17:51 mtholder dp == do
17:51 towodo no. looks fine to me.
17:52 jimallman agreed
17:52 mtholder it does have several recent changes other than the deploy stuff
17:52 mtholder but it looks good to me
17:53 towodo jim has already done the risky part.  deployed today at 6:08 UTC
17:53 mtholder good point. we'll all blame jim if i break anything
17:53 towodo boy am i glad i added that install logging feature
17:55 towodo (pats self on back. awkward)
17:55 mtholder particularly awkward w/ your bad shoulder
17:56 mtholder http://dev.opentreeoflife.org/ seems to be happy
17:56 towodo actually shoulder has been great since cortisone shot
17:56 towodo what's the test for happiness?
17:56 mtholder curator bounces to devtree and bibliographic refs from ot12
17:56 mtholder tree still browsable...
17:57 towodo what was the symptom of the oti dependence?
17:57 mtholder looking at source code
17:58 mtholder (that was the symptom)
17:58 towodo i gotta go to my nescent meeting.  will be gone for up to 2 hours.  will check in after
17:58 mtholder http://dev.opentreeoflife.org/about/materials-and-methods is still lorem ipsum, but that seems appropriate.
17:58 mtholder ok thanks for the help
17:59 towodo jimallman, we have our certificate now and can do apache config.  i wanted to talk to you about that after my mtg
18:00 jimallman sounds good. i’ll be here, fixing bugs. approx time?
18:00 jimallman oh, i see above
18:00 mtholder left #opentreeoflife
18:00 jimallman i’ll be here around 4pm
18:00 mtholder joined #opentreeoflife
18:04 mtholder git hub is acting really flaky...
18:04 blackrim mtholder: yes it is. can't pull, push
18:05 mtholder that makes me sad
18:06 * towodo waits for meeting to start
18:11 josephwb working for me
18:12 blackrim seems to be flaky for me. can't pull
18:12 jimallman https://status.github.com/
18:14 blackrim aha! got my push in there right before it flaked again
18:18 kcranstn joined #opentreeoflife
18:19 mtholder joined #opentreeoflife
18:29 mtholder joined #opentreeoflife
18:44 kcranstn joined #opentreeoflife
18:53 kcranstn joined #opentreeoflife
19:38 towodo joined #opentreeoflife
19:39 kcranstn joined #opentreeoflife
19:56 josephwb Looking at the API examples
19:57 josephwb Says "Parameter treeID = id of the parent synthetic tree (string); defaults to current synthetic tree", but treeID is apparently required.
19:58 jimallman josephwb: ah, good to know. i cribbed the note about default behavior from source-code comments, but it sounds like they’ve gone out of date.
20:08 josephwb what should we change: the documentation of the method?
20:08 josephwb OR the method
20:08 kcranstn method, if possible
20:08 josephwb ok, looking at it right now. should be easy
20:08 kcranstn thanks, joseph!
20:25 towodo jimallman, you there?
20:26 jimallman towodo: yes, just wrapping something up.
20:26 towodo ok lemme know when done
20:26 jimallman new topic: certs!
20:27 jimallman ok, i’m ready
20:29 jimallman towodo: here? or hangout?
20:29 towodo here is ok.
20:29 jimallman sounds good. so we have wildcard certs?
20:30 towodo yes. i was hoping you could help with apache config for ot11 & ot10
20:30 jimallman i think so, sure
20:30 kcranstn I am also here - let me know if you need anything from me
20:30 jimallman thanks! how to y’all feel about all-HTTPS-all-the-time?
20:30 josephwb just pushed the change to make treeID argument optional
20:31 jimallman or would you prefer HTTPS-only-when-logged-in, or HTTPS-only-when-passing-credentials?
20:31 towodo I don't like all-https for information that doesn't need to be kept private.
20:31 towodo but I could concoct cases where it would make a difference to someone
20:31 towodo basically if they care about security they shouldn't be visiting our site… :)
20:31 jimallman yeah, tough call. once someone is logged in, session hijacking is always an issue.
20:32 towodo and there are privacy concerns.
20:32 jimallman and login will happen implicitly (after they’ve logged in once) even in the tree browser
20:32 towodo after login, yes, https.  the credentials need to be kept safe
20:33 towodo see /etc/ssl/private/
20:33 jimallman OK, i’ll see if I can come up with a simple “bump” to HTTPS (when logged-in only) that works across all apps.
20:33 jimallman joined #opentreeoflife
20:33 * jimallman got bumped momentarily...
20:33 towodo I'm getting the cert now from email, will put it in /etc/ssl/certs/
20:34 jimallman on ot11 and ot12, you mean?
20:35 towodo right now I'm just putting stuff on ot11.   should you decide to accept this mission you'll copy them to ot12
20:35 jimallman right, wildcard!
20:35 * jimallman is changing his irc nick to wildcard…
20:37 towodo I think we can get away with a single apache vhost file
20:38 jimallman looks like there’s a header for GitHub session, i’ll try to sniff this in apache config and choose HTTP/S accordingly
20:38 jimallman X-GitHub-Session-Id:17951574
20:38 towodo so the main thing is to merge the ssh config from default-ssh into apache-config
20:39 towodo the <virtualhost> port spec will have to be changed… I guess *:80 and *:443 ? I don't know if that will work
20:39 towodo just looking for something really simple
20:40 jimallman got it. i suppose some API calls will also need HTTPS if they have sensitive stuff, at least the ones from client-side code.
20:40 towodo I guess we can put certs in github - no reason to keep them secret I think ?
20:40 jimallman i’ll review some of my older apache config files and refresh my memory.
20:40 josephwb jimallman: just a note I changed the treemachine method getSyntheticTree to match the documentation. so, you will probably want that code up.
20:41 towodo yes, any API call containing credentials of any kind
20:41 towodo I can do that
20:43 towodo ok, certs and authority chain stuff is in /etc/ssl/certs/opentree/
20:43 towodo on ot11
20:43 jimallman great, thanks! i’ll try to get this set up tonight.
20:44 towodo i'm wondering about the advisability of putting the certs in github… surfing now...
20:45 towodo this is an excellent example of why dumb text search loses.
20:46 towodo 'protect ssl certificates' doesn't get anywhere close. the subject/object distinction matters a great deal
20:46 jimallman i’m looking for this too… it makes me itch, not least because it seems like someone could set up evilstuff.opentreeoflife.org and use our wildcard, right?
20:46 towodo nope.  they would need to know our private key
20:47 towodo we certainly wouldn't put the private key in github.  just the cert.
20:47 jimallman i’m having better luck going to general background stuff (wikipedia, etc). i tried “opps accidentally pushed cert to github”… figured that would get something..
20:47 jimallman ah yes, the cert itself is easy to fetch
20:47 towodo namecheap sent the cert in email.  they can't care too much about it
20:48 jimallman but the email was marked “TOP SECRET”
20:48 kcranstn when the nsa sees that, they immediately stop surveillance
20:49 towodo well, up to you i guess… i think if it were me i'd copy the files that I put in /etc/ssl/certs/opentree/ to deploy/certs/ or deploy/setup/certs/ and add them to the repo
20:50 towodo and the private key becomes yet another of our secrets… this one has to be treated especially carefully so I would do it manually, with instructions somewhere.  only has to be done once per server
20:51 towodo and it needs admin / sudo
20:52 towodo because we don't want the 'opentree' user to be able to see it.  goes in /etc/ssl/private/
20:52 jimallman so far we’ve consolidated all secret stuff in each pusher’s private ~/.ssh/opentree, with instructions on where to fetch them via scp. ok to continue with this method?
20:52 jimallman hm, can we move (and place) these in the as-admin scripts?
20:52 jimallman deployment scripts, i mean
20:53 * jimallman sees that you’ve already mentioned admin/sudo… makes sense.
20:54 towodo hmm.  if the key goes in ~/.ssh/opentree that would just be a convention, each developer could put it wherever they want, since it won't be scripted...
20:54 towodo in my proposal that is...
20:55 towodo the certs can go in deploy/setup/ so will get copied, no problem there
20:55 jimallman so you’re proposing a truly manual operation, maybe using scp for private keys. can we put this in the as_admin script?
20:56 towodo I would rather not.
20:56 towodo I really don't even want people storing the key on their computers, unless under a passphrase (as I have done).
20:57 jimallman OK, i’ll keep this out of the general deployment tools.
20:57 jimallman do we have a server-setup checklist? if so, i could add a note there.
20:57 towodo yep.  it wouldn't hurt to put the correct scp command in a comment or in documentation
20:57 jimallman towodo: ^
20:57 towodo I've been talking about checklists for a long time...
20:57 jimallman sounds like i should start one!
20:57 towodo maybe time to create one. not sure where
20:58 towodo deployed-servers maybe? or opentree/deploy/docs ?
20:58 jimallman i fear it will get lost in opentree repo, but maybe in the local README of deploy/
20:58 towodo or wiki? ...
20:58 towodo sure
20:59 jimallman deployed-systems feels kind of right, since that’s where we list all available machines… i’ll play it by ear and see where it feels the most natural.
20:59 towodo sure
21:00 towodo I've been the one doing initial server setup, so private key install is one of the things I'd do
21:02 towodo if you want to talk further about security particulars maybe better over hangout or phone than on a public irc that will be archived forever… would rather not expose unplugged vulnerabilities in a public forum
21:02 jimallman true that. i’ll dig into this tonight and let you know if i hit a wall.
21:03 towodo ok.  did you look at the ssl vhost file in apache2/available-servers ?
21:03 jimallman not yet, lemme see...
21:04 jimallman sites-available, you mean?
21:04 towodo yes
21:04 towodo I think we just take the ssl-related directives from there, and put them in apache-config, and add *:443 at the top
21:04 jimallman looking at default-ssl…
21:05 towodo and fix the key and cert paths. then ssl should work. i'd start with that, then fix the web2py stuff
21:05 jimallman yes, that sounds right to me. maybe add some directives here to bump from HTTP <=> HTTPS based on headers, etc.
21:06 towodo again I'd rather not do that in apache-admin. we have to support opentree servers that have no credentials.
21:06 towodo sorry apache-config
21:07 towodo not sure what you mean by "here", I assume you meant apache-config
21:07 jimallman yes
21:07 towodo trying to avoid complexity
21:08 jimallman i’ve been pondering that (opentree servers w/o credentials)… my thinking is that lots of functionality will just stop working w/o the expected “secrets”. would you prefer we try to keep things working, but in an insecure fashion?
21:09 towodo the only things we lose are the ability to edit a study, or make a comment, right? what else?
21:09 jimallman i’m expecting these to mostly be sandbox systems, so maybe not exposed to the wider net..but playing with tree and study data is something i thought they’d want to do.
21:10 jimallman i suppose if they’re more interested in synthesis and the existing corpus it’s no biggie.
21:10 jimallman so yes, i think you’ve summed it up. no curation and no comments.
21:10 towodo there's a lot to the system that's not those things.
21:11 towodo we want to encourage forking and experimentation.
21:11 jimallman no opentreeapi bot user… it seems like that will break something more. maybe not.
21:11 jimallman it’s definitely a desirable change, just trying to think through the desired behavior.
21:12 towodo if someone want to do curation or comments, they have two routes: (1) join our club, (2) fork phylesystem as well.
21:12 towodo but it's enough of a pain we can't make secret management be part of the barrier to entry
21:12 jimallman agreed
21:14 towodo foo, looks like ot6 has a github client id.  also it points to the wrong phylesystem.  need to change both those things
21:17 jimallman i’m betting that CLIENT_ID is unused. it’s got an old name
21:17 towodo and it looks like I never filed my issue against gcmdr about the asterales build failing…
21:17 jimallman more recent config files always specify the registered app, eg TREEVIEW_GITHUB_CLIENT_ID
21:17 towodo and the server with the RAID isn't booting… too much to do
21:19 towodo jimallman, please read this: http://www.crossref.org/02publishers/doi_display_guidelines.html
21:19 towodo we need to be displaying  http://dx.doi.org/10.1234, not 10.1234
21:20 towodo i'll make an issue, sorry for the distraction
21:20 jimallman ah, OK. yes, this should be easy, but please start a ticket.
21:22 towodo now I can't find the study that had one of these...
21:23 towodo maybe you already fixed this.  I definitely saw one, recently, somewhere
21:25 jimallman hm, i’ll double-check my code. this should be grep-able in phylesystem-0… i’m having a hard time including literal quotes in a GitHub search. their search seems pretty hit-or-miss in my experience.
21:26 towodo don't worry about it.
21:26 towodo i tried a bunch of cases that were 10.xxx in the phylografter db and they all look fine
21:26 jimallman OK, but let me know if you see something like that.. could just be old/test data
21:26 jimallman ah, sounds like we’re cleaning it up then
21:29 towodo ok. next I think I'm going to finish setting up ot10 with tree & taxo, so we can point ot11 at it.
21:33 towodo don't know what phylesystem should be used for ot10.  ot12 points at repo-phylesystem-0, so need something that's not that
21:34 towodo maybe I'll just hold off on deploying the api, for now
21:35 jimallman just ‘phylesystem-0’, right? unless ‘repo-phylesystem-0’ is a branch or something…
21:36 jimallman i think mark is working on allowing multiple api instances to push to a common repo.. not sure how close he is to done.
21:36 towodo yes.  i was looking at
21:36 towodo OPENTREE_GH_IDENTITY=~/.ssh/opentree/repo-phylesystem-0.pem
21:36 towodo i'm losing my brain. i've made way too many mistakes recently
21:37 towodo now need a GITHUB_CLIENT_ID for devapi.  looking to see if I made one
21:37 jimallman it’s not just you. we’re a bit of a flying circus right now… looking forward to business as usual.
21:38 towodo ok, don't need GITHU_CLIENT_ID … that's for the front end
21:39 mtholder I have not tested the multiple branches using the same phylesystem repo.
21:39 mtholder I thought that there was a rush to get the next round of synth done (not sure why there is a rush...) so that we were probably going to deploy phylesystem-1 to ot10 soon.
21:40 jimallman mtholder: thanks for the update
21:40 towodo what i was saying was that both ot10 and ot12 currently both use phylesystem-0.  need to exclude ot10 until mark's extension is done
21:40 towodo or until another repo is chosen, or created...
21:42 * jimallman needs to step away for awhile… will check back here within the hour.
21:45 towodo or we could make ot12 access to phylesystem-0 be read-only, to a snapshot?  or none at all?  (we have no documented api calls that write)
21:45 towodo ok. hmm. so the exclusion is to keep ot12 from modifying phylesystem-0.
21:46 mtholder we can certainly disable it from pushing. we don't have logic for it to pull from GH. so it won't stay in sync.
21:46 mtholder that may not be a prob.
21:47 mtholder or a crontab would work
21:47 * towodo ok, oti, treemachine, taxomachine updated on ot10.  now will deploy tree & taxo databases
21:48 towodo i've been wrestling with how to disable access… could manually delete the github secret, but it might get reinstalled when someone does a push
21:49 towodo if the name of the secret file were in the config, it would be easy
21:49 mtholder It is that PKEY bit in the api../private/config
21:50 towodo the push script copies to there from a fixed file location on the pusher
21:54 towodo I'll need to change push.sh ...
21:54 mtholder We could hold off (expecting folks to not push to ot12 because the docs are not that prominent).
21:54 mtholder hopefully very soon we'll have phylesystem-1 on ot10
21:54 mtholder and no contention.
21:57 towodo I could point ot10 to phylesystem-1 now… it won't work, and will be a reminder that work is in progress…
21:57 mtholder it will work, I think.
21:58 mtholder unless you put the wrong key in the config.
21:58 towodo oh.
21:58 towodo i'll put phylesystem-2, or better not-assigned.
21:59 mtholder bear in mind that ot11 is pointing to ot10 for api.
21:59 mtholder and that is how we are testing curation.
21:59 towodo umm…
22:00 towodo right. ok. so it *has* to point to phylesystem-0 … ?
22:00 towodo so need to change ot12. (or not since it won't be trying to write.)
22:01 mtholder we could "go live" with phylesystem-1, but I'd prefer to wait until I get that repo updated from pg and gcmdr
22:01 mtholder I'm working on a script now to tag the candidateTreeForSynthesis info (currently in gcmdr) into the nexsons, where it belongs.
22:01 mtholder then I think that we can use it.
22:02 towodo not sure what "go live" means here.
22:02 towodo I think I just confused, yet again.  will compare my new ot10 config with previous one before pushing
22:02 mtholder I'd prefer to debug a while on phylesystem-0, but it sounded like some folks wanted to be curating for the next round of synthesis before Friday.
22:03 mtholder by "go live" I meant - use the repo that we don't plan on later trashing.
22:04 towodo so, ot10 should be phylesystem-0 right now, changing to phylesystem-1 when you say we're ready. yes?
22:04 mtholder yes.
22:05 mtholder it could cause oddness on ot12, but (as you pointed out) we don't really expect api users to be pushing.
22:05 towodo my top level goal right now is to remove the reference from ot11 to ot12, so we don't screw up the released api
22:05 mtholder "oddness" here is just failure to push to GH and an out of date oti db
22:06 towodo ot11.config: "OPENTREE_NEO4J_HOST=ot12.opentreeoflife.org"  -- wrong.
22:06 mtholder should be 10?
22:06 towodo so, installing tree and taxo on ot10.  then change to ot10, yes.
22:07 josephwb joined #opentreeoflife
22:07 mtholder those methods are all read-only, so it shouldn't be possible for ot11 to mess up the treemachine/taxomachine on ot12, though.
22:07 mtholder but certainly pointing to ot10 is fine with me.
22:08 towodo yep. I have to keep convincing myself of that
22:08 towodo we need to be able to make tree/taxo changes for development, without putting released api at risk
22:08 mtholder I certainly agree that as we change tree/taxo we'll want to deploy that and test on ot10.
22:09 mtholder (and hence get ot11 pointing to ot10 for those services).
22:10 mtholder gotta run...
22:39 * towodo heading out for a bit, back within an hour
22:55 mtholder joined #opentreeoflife
23:31 * towodo back
23:52 mtholder joined #opentreeoflife
23:56 towodo devapi now points to ot10
23:57 towodo I need an api test suite.
23:58 towodo anyone happen to know where the latest api documenation lives?
23:59 towodo https://github.com/OpenTreeOfLife/opentree/wiki/Open-Tree-of-Life-APIs

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