Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2015-05-02

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

All times shown according to UTC.

Time Nick Message
00:00 jar286 the timing here is a bit of a problem. i’m ready to update ot10 to jessie but if it’s broken i won’t know if i broke it or not. not sure what to do
00:02 josephwb db extraction just finished. neo4j is restarting
00:02 josephwb should be free in a minute
00:03 jar286 but will the front end work with the new db, if the tree name has changed?
00:03 josephwb there it goes. it is all yours.
00:03 josephwb front end?
00:03 jar286 the webapp
00:03 josephwb jim's stuff?
00:03 josephwb oh
00:03 jar286 yes
00:03 jar286 let me try now
00:04 josephwb i hope so, but that is sort of the point: to troubleshoot the change
00:04 josephwb i am done; do what you need
00:04 jar286 seems fine.
00:04 josephwb ok
00:04 josephwb good luck
00:04 jar286 it still says otol.draft.22
00:04 josephwb huh?
00:04 jar286 in the URL
00:04 josephwb the webapp
00:04 josephwb yes
00:04 josephwb but not the db
00:05 jar286 ok
00:05 jar286 i think i will put off my update until the morning. doing backups now
00:05 josephwb oh, wait, yeah, it uses some hardcoded constants;
00:05 josephwb but if jim supplies the correct name it should work
00:05 josephwb will check later.
00:05 josephwb need to get my bus for real!
00:06 jar286 ok, i think things are fine
00:21 kcranstn joined #opentreeoflife
00:54 jimallman thanks for the updates, folks. i’ll take a look from here and see what i can see.
00:57 josephwb joined #opentreeoflife
01:00 josephwb hey jimallman and jar 286. anything clearly wrong with dev?
01:00 jimallman hi, just going there now to take a look…
01:01 josephwb should be same db with with different synth tree name
01:01 jimallman it seems to be working normally, but URL and page source still show ‘otol.draft.22’
01:02 jimallman how can i recognize the latest tree (vs. previous)?
01:02 josephwb but you see a tree?
01:02 josephwb tree is the same
01:02 josephwb just different name for testing
01:02 jimallman yes, the tree shows up normally. lemme see if this arguson is cached...
01:03 josephwb can you do the curl call with the new name?
01:04 josephwb [i guess i can do it just as well]
01:04 jimallman tinkering with this now… removed ‘cached/‘ to make sure response was fresh (still with ‘otol.draft.22’) and output looks normal. which is odd, yes?
01:05 jimallman i know you uploaded the database, are you sure it’s been updated and restarted in neo4j?
01:05 josephwb yes
01:05 josephwb no
01:05 josephwb yes: weird. know if restarted: no
01:06 jimallman ok, i’ll backtrack and see if we followed all the required steps…
01:07 josephwb yeah, i give it a fake name and it return the same
01:07 josephwb erg
01:07 jimallman funny, i see that calling getSyntheticTree with a completely bogus tree id actually “works”, at least it returns the target taxon as an isolated node (no surrounding tree).
01:07 josephwb i need to update the webservices
01:07 josephwb yeah, i cannot update getSyntheticTree with java 8
01:07 jimallman note that a curl fetch will state the tree id twice: once in the url, and again in the —data (JSON) arguments
01:08 jimallman i suspect that only one of these affects the response.
01:08 josephwb weird that it gets anything, tho
01:08 jimallman it’s the JSON argument that matters (supercedes the tree id in URL)
01:08 josephwb yeah
01:09 josephwb cannot push the update
01:09 josephwb devploy, that is
01:09 jimallman which update? code for treemachine?
01:09 josephwb deploy
01:09 jimallman or db?
01:09 josephwb yes
01:09 jimallman which?
01:09 josephwb the plugins
01:09 jimallman ok
01:09 jimallman and that’s because we haven’t put java8 on that machine?
01:09 josephwb yes
01:10 jimallman (you suggsted above that it’s not really a requirement, but maven thinks it is… is that the case?)
01:10 jimallman suggested
01:10 josephwb yes
01:10 jimallman ok, lemme see if i can talk maven off the ledge.
01:10 josephwb there must be a weird synonym mapping in the db for the tree name such that the hardcoded one still works
01:10 josephwb it worked correctly on my machine
01:11 jimallman it doesn’t seem that odd to me for a node id to work, if only by coincidence.
01:12 josephwb it is the same db
01:12 jimallman note that the solitary node returned is not the same as we get with the proper tree (otol.draft.22)
01:12 jimallman it seems the node id happens to exist, but assigned to another taxon by default.
01:12 josephwb same db; i just went in and changed the name of the tree
01:12 josephwb so, maybe it is working?
01:12 josephwb huh
01:13 josephwb should be the exact same as before
01:13 josephwb didn't re-run anything
01:13 jimallman “same db” meaning same as before? meaning your updated db is not being shown?
01:14 josephwb i don't know anything anymore
01:14 josephwb the latest db, i just changed the name and pushed it again to dev
01:14 jimallman ok. sit tight and let me take a closer look on the server. i’ll report here as i find things.
01:15 jimallman can you paste the push command here?
01:15 josephwb just a sec, but it is the same as the one i used yesterday
01:16 josephwb ./push.sh -c ../../deployed-systems/development/devapi.config pushdb newlocaldb.db.tgz treemachine
01:16 josephwb i literally just did the up-arrow unix thing
01:16 josephwb no typos!
01:17 josephwb i need a break from this. i will log into my office machine in a bit and see what is what
01:17 jimallman sounds good. i need some alone time here as well. :)
01:17 josephwb i know the arguson call does use the hardcoded (old) name
01:17 josephwb i am just surprised it hits anythign
01:18 josephwb the name must be stored in more places in the db than i realized
01:19 josephwb i was able to use the new webservices on my machine (java 8); that must be it
01:19 josephwb still, weird that it works. i will make sure to break whatever that is
01:20 josephwb i.e. my machine does things as expeected
01:20 jimallman right
01:20 josephwb i will do some investigating and see if i can turn the hard-coded stuff off
01:21 josephwb ok, talk to you in a bit
02:21 josephwb joined #opentreeoflife
02:22 jimallman josephwb: fyi, i’m (re)running install-db to try to get the new db installed:
02:22 jimallman $ ./push.sh -c ../../deployed-systems/development/devapi.config install-db downloads/treemachine.db.tgz treemachine
02:23 jimallman your pushdb command should have already done this, so it’s just a sanity check.
02:23 josephwb did it not work?
02:23 josephwb oh, ok
02:23 jimallman it’s taking awhile, not done yet
02:24 jimallman based on top output, it looks like it’s still decompressing…
02:24 josephwb hard to troubleshoot db things when the code bases differ
02:24 josephwb i am just going to make a dummy treemachine branch with all of the new changes but none of the java 8 stuff so we can actually test things
02:25 jimallman ok. i’ll keep you posted as i get more notifications from the install-db in progress.
02:25 josephwb ok
02:25 josephwb thanks
02:26 josephwb i don't expect anything positive
02:27 jimallman that’s the spirit!
02:27 jimallman i’m curious to see if we’re hitting the wall on disk space or somesuch
02:27 josephwb maybe
02:28 josephwb that last push took quite a long time
02:32 jimallman ok, it’s doing the next steps now (neo4j shutdown, etc)
02:34 jimallman ok, it’s done (apparently). checking the site again now...
02:35 josephwb it looked like it completed last time
02:36 jimallman still shows otol.draft.22 in page source and getSyntheticTree calls… i have more detective work to do.
02:37 josephwb nah, the arguson code is not updated
02:37 josephwb i don't think we can do anything at this point
02:38 josephwb hopefully this dummy branch won't take too long
02:38 josephwb right, ilbot3?
02:38 jimallman ok, i’ll hang back and wait for your dummy branch
02:38 jimallman (ok, maybe a little detective work in the meantime)
02:40 josephwb ack, this is a nightmare
02:40 josephwb cody went nuts with branch compatability and java versions
02:40 jimallman hrm
02:41 josephwb i mean, he had to do some things to get the work done, don't get me wrong
02:42 jimallman hasn’t jar worked out the kinks in java8? i thought he’s planning to update devapi soon.
02:42 jimallman https://github.com/OpenTreeOfLife/germinator/wiki/Debian-upgrade-notes:-jessie-and-openjdk-8
02:43 josephwb i need it now!
02:44 josephwb this dummy branch might be too much work. maybe i should just wait
02:44 jimallman yeah, i don’t know his timetable
02:45 jimallman maybe you should chime in on his thread in the OpenTree-software group, and see what he has in mind.
02:49 josephwb ok, i give up on the dummy branch (for now). so many cross dependencies my head is spinning
02:50 * jimallman is going to goof off for awhile, will return to this Saturday.
02:57 josephwb *is going away too, maybe never to return*
08:33 mtholder joined #opentreeoflife
11:49 mtholder joined #opentreeoflife
12:57 jar286 joined #opentreeoflife
13:35 jar286 joined #opentreeoflife
13:45 jar286 joined #opentreeoflife
15:07 jar286 jimallman, i’m having some ssl trouble after upgrading devapi to jessie… apache won’t start
15:12 jar286 for the time being i’ve disabled ssl on devapi
15:16 jar286 joined #opentreeoflife
15:43 jar286 joined #opentreeoflife
15:54 jimallman jar286: does apache2ctl point to any specific config problem?
15:54 jimallman re: ssl, i mean
15:55 jar286 hang on, looking
15:56 jar286 [Sat May 02 15:54:25.351293 2015] [ssl:warn] [pid 22046:tid 139910036371328] AH01909: ashby.csail.mit.edu:443:0 server certificate does NOT include an ID which matches the server name
15:56 jar286 [Sat May 02 15:54:25.351989 2015] [ssl:error] [pid 22046:tid 139910036371328] AH02579: Init: Private key not found
15:57 jar286 [Sat May 02 15:54:25.352025 2015] [ssl:error] [pid 22046:tid 139910036371328] SSL Library Error: error:0D0680A8:asn1 encoding routines:ASN1_CHECK_TLEN:wrong tag
15:57 jimallman looks like the wheezy => jessie upgrade pushes apache from 2.2 to 2.4, with some known issues
15:57 jar286 and a bunch of others like the last of the 3
15:57 jimallman ok, will dig for answers
15:57 jar286 yes
15:57 jar286 well maybe i can do some of that digging…
15:57 jar286 and you need to be aware of some changes I made
15:58 jar286 I’m keeping the 2.2 and 2.4 vhost conf files in parallel, so we can deploy to either kind of system
15:59 jimallman sounds good! it looks like the default set of apache modules has changed in 2.4, and auth and ident modules can suffer
15:59 jar286 but where would you start digging?
16:00 jimallman i just start googling like a frustrated person, eg “wheezy to jessie ssl breaking”
16:00 jimallman i’ve been known to google things like “jessie apache wtf” with good results  :)
16:01 jar286 right, ok.  I have a transcript from the installation, maybe that will help too. and http://httpd.apache.org/docs/2.4/upgrading.html . etc etc
16:02 jimallman sounds good. i’m going to try apache2ctl testconfig or configtest, or whatever it is…
16:02 jar286 did that already…
16:03 jimallman here’s a bit more detail on the problems with changed default apache modules
16:03 jimallman https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=752922
16:04 jar286 I just did a redeploy on devapi, checking to see whether everything works in absence of the https vhost
16:04 jar286 (remove from sites-enabled)
16:04 * jimallman nods
16:06 jar286 no, that loses too… there’s a web2py error… funny, since asterales works. now I have two errors to deal with… time for lunch
16:06 jar286 i don’t want us to work redundantly, so maybe hold off on working on these
16:08 jar286 oh wait.  problem #2 seems to be that the URL gets switched from http: to https:
16:08 jar286 is that something the browser is doing, or opentree?
16:08 jar286 both chrome and safari
16:09 jar286 so I will work on #1 after lunch.  thanks
16:14 jimallman ah yes, we do force many URLs to HTTPS (login of course, and anything that requires an authenticated user)
16:14 jimallman i’ll focus elsewhere until i hear more from you on this
16:53 jar286 fixed problems in mod_ssl config file…
16:58 jar286 now connections to port 443 are answered with http, not ssl
17:01 jar286 looking at new default-ssl.conf … I think the vhost config file just has to be modernized
18:06 josephwb joined #opentreeoflife
18:26 scrollback joined #opentreeoflife
18:31 josephwb joined #opentreeoflife
19:23 jar286 joined #opentreeoflife
20:03 jimallman joined #opentreeoflife
20:05 jar286 jimallman, first problem is that now ServerName is required in the ssl vhost config…
20:05 jar286 that lets the server start up, but now I’m getting an ‘internal server error’ which I need to track down. so progress
20:06 jar286 this will require a bit of sed scripting in the push.sh apache setup
20:24 jar286 need to tweak private key permissions…
20:24 jar286 now I’m getting [Sat May 02 20:19:33.650882 2015] [wsgi:error] [pid 26093:tid 140631106705152] [client 216.216.68.69:42832] Daemon process called 'web2py' cannot be accessed by this WSGI application: /home/opentree/web2py/wsgihandler.py
20:25 jar286 problem only happens with https, not http
20:25 jar286 ok, signing off for now… thanks for listening
20:56 blackrim joined #opentreeoflife
20:56 blackrim left #opentreeoflife
21:43 jar286 joined #opentreeoflife
22:27 jimallman_ joined #opentreeoflife
22:27 kcranstn joined #opentreeoflife
22:30 jar286 joined #opentreeoflife
23:07 jar286 jessie + apache 2.4 works!  now on to java 8…

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