Time
S
Nick
Message
00:57 axfelix joined #dataverse
01:26 garnett joined #dataverse
02:18 shauna_ joined #dataverse
02:22 cipherrot joined #dataverse
03:27 garnett joined #dataverse
03:42 xvx joined #dataverse
05:01 axfelix joined #dataverse
05:06 garnett joined #dataverse
05:55 garnett joined #dataverse
08:59 jri joined #dataverse
09:19 xvx joined #dataverse
09:36 bencomp joined #dataverse
09:40 bencomp
pdurbin: I saw :)
11:08 pdurbin
bencomp: oh? *Before* I linked to it? It was news to me.
11:15 bencomp
pdurbin: I saw a notification that you had linked to it ;)
11:15 bencomp
I hadn't seen it before, although I didn't think it had to be very hard to do it
11:16 pdurbin
it's cool that they mention Kubernetes
11:17 pdurbin
"Dataverse under Docker": https://docs.google.com/document/d/1sUf8ReWJv2rwjSmhYF7E3CXNwiemIm9P7AXNUaBA2zY/edit?usp=sharing
11:18 pdurbin
anyway, I linked to these IRC logs from that doc so maybe they'll come chat with us :)
11:29 pdurbin
on a related note, http://www.se-radio.net/2016/01/se-radio-show-246-john-wilkes-on-borg-and-kubernetes/ was great
13:13 bmckinney joined #dataverse
14:09 bencomp joined #dataverse
14:18 pdurbin
bencomp: not sure if you've listened to that episode yet
14:22 pdurbin
bmckinney: have I told you about http://phoenix.dataverse.org ?
14:26 bmckinney
pdurbin: no, interesting
14:28 pdurbin
I mention it because I noticed that some pull requests got merged into develop.
14:28 pdurbin
So I just kicked off https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-build-develop/9/
14:28 pdurbin
which triggered https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-deploy-develop/9/
14:29 pdurbin
which triggered https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-apitest-develop/
14:29 pdurbin
and all three passed (build, deploy, apitest), which is good, what we want and expect
14:30 pdurbin
anyway, I thought you might find it interesting
14:34 bmckinney
yeah, at some point i'll want to touch base on continuous integration, etc.
14:36 pdurbin
bmckinney: please check out the slides I posted at https://groups.google.com/d/msg/dataverse-community/X8OrRWbPimA/_1k64l7mDAAJ
14:40 bmckinney
cool, thanks!
14:42 pdurbin
oh sure
14:43 pdurbin
bmckinney: oh, did you see my comment the other day at https://github.com/IQSS/dataverse/issues/2249#issuecomment-190798544 ?
14:50 bencomp
pdurbin: I did listen to that episode of SE Radio, but it's been a while
14:50 bencomp
and I didn't do anything with Kubernetes
14:50 pdurbin
bencomp: but doesn't it sound cool?
14:51 bencomp
I've learned to look beyond cool ;)
14:51 bencomp
but yes :D
14:53 pdurbin
well, you might also like Google Cloud Platform Podcast - https://www.gcppodcast.com ... I just finished episode 5.
14:57 bmckinney
pdurbin: where do i assign myself the issue? not sure i have rights
14:59 pdurbin
bmckinney: this is how it looks for me: http://i.imgur.com/tX26d8a.png (Assignee on the right)
15:00 bmckinney
pdurbin: yeah, it's a rights thing, i simply see "no one assigned"
15:01 pdurbin
bmckinney: it might be time for me to give you push access then. Any objections?
15:05 bmckinney
pdurbin: no, that's fine (i won't abuse my power)
15:06 * pdurbin
eyes bmckinney
15:10 pdurbin
bmckinney: can you please refresh and let me know if it looks different?
15:12 pameyer joined #dataverse
15:13 bmckinney
pdurbin: all set, thanks!
15:13 pdurbin
cool. I actually created a new team: https://github.com/orgs/IQSS/teams/dataverse-write
15:14 romainM
hi people :)
15:14 pdurbin
mental note to check out https://help.github.com/articles/migrating-your-previous-admin-teams-to-the-improved-organization-permissions/
15:14 pdurbin
romainM: mornin'
15:14 romainM
depend for who :D
15:15 romainM
well, I got a little question
15:16 romainM
if I want to add a new identifier scheme to dataverse, should I make a request for it and where ?
15:16 pdurbin
romainM: for users?
15:16 romainM
for authors of datasets
15:17 pdurbin
right now we support ORCID, ISNI, LCNA per http://irclog.iq.harvard.edu/dataverse/2016-03-01#i_31822
15:17 romainM
in France, it seems our archivists use a system called "Idref"
15:18 romainM
(I need a little more info from my people to have a clear understanding of this)
15:18 pdurbin
romainM: is there a website explaining Idref?
15:18 romainM
but I wanted to know "if" it was possible to add a new identifier scheme, if it was needed, just in case
15:18 romainM
I asked more info, still waiting
15:18 romainM
but I got this, for example
15:19 romainM
http://www.idref.fr/autorites/autorites.html
15:19 romainM
in ... french
15:19 romainM
:D
15:20 romainM
from what I understood, it regroups multiples authorities about university libraries in France
15:20 pdurbin
romainM: as luck would have it the person who works on https://github.com/IQSS/dataverse/blob/v4.2.4/scripts/api/data/metadatablocks/citation.tsv speaks French. Please search there for authorIdentifierScheme
15:22 pdurbin
pameyer: speaking of monitoring, we monitor https://dataverse.harvard.edu at https://opsview.hmdc.harvard.edu/viewport/www
15:24 pameyer
I didn't think opsview gave you historical info about things like load and memory usage - so maybe monitoring wasn't the right term
15:25 pameyer
btw - looks like that opsview has the same "no results recieved, assuming it's all good" that sometimes crops up
15:25 pameyer
https://opsview.hmdc.harvard.edu/viewport/www?style=group_by_host&host=dataverse.org
15:25 romainM
ok, ty but ... what should I search ? I don't fully understand
15:25 pdurbin
pameyer: Opsview does collect "performance data" but I think you have to be logged in to see it.
15:26 pdurbin
romainM: line 130: https://github.com/IQSS/dataverse/blob/v4.2.4/scripts/api/data/metadatablocks/citation.tsv#L130
15:27 romainM
yes, I see the values but ... didn't you gave me the advice to not change that file last time ?
15:27 romainM
(or it was just about booleans ?)
15:27 romainM
we had a discussion about something like that
15:27 pdurbin
romainM: yes. please don't just change that file. please open an issue with a link to a website describing what "Idref" is.
15:28 romainM
if I can simply change the values, I got no problems with that, just wondering if it was an identifier to "ask" for the official dataverse
15:28 romainM
ok, ty and sorry for disturbing
15:28 pdurbin
no worries
15:28 romainM
I'll open that issue with more info later
15:29 pdurbin
I mean, on some level, you can just do what you want, of course... Dataverse is open source. But we'd like to keep all the code together instead of having a lot of forks out there.
15:29 pdurbin
pameyer: I only have a foggy memory of what apache-configs-locked might even mean
15:34 pameyer
foggy memory is better than my understanding of it - but opsview can be ideosyncratic sometimes; probably more of a sign of that than a real problem
15:36 pdurbin
pameyer: this is the monitoring system I'd like to have some day: https://prometheus.io
15:36 pdurbin
romainM: no worries. please paste a link in here after you've created the issue
15:39 pameyer
pdurbin: looks prometheus looks interesting
15:40 pdurbin
too bad I missed http://monitorama.com/2013/boston.html
16:18 axfelix joined #dataverse
16:26 metamattj joined #dataverse
16:28 pdurbin
I love it when there are so many people in here people start to scroll off my screen (sorry yoh) :)
16:38 xvx joined #dataverse
17:24 yoh
yoh: why you are sorry for me? ;)
17:32 pdurbin
yoh: just saying you're scrolling off the list, that's all: http://i.imgur.com/l19wMN1.png . It's a good thing. :)
17:34 yoh
pdurbin: ah ... yeah -- a good thing I can contribute to ;)
17:36 pdurbin
heh
18:01 bmckinney
deploy error: org.postgresql.util.PSQLException: ERROR: could not open file "base/16384/16510": No such file or directory
18:01 bmckinney
something simple i'm missing?
18:18 pdurbin
base? huh
18:19 pdurbin
bmckinney: how did you install postgres?
18:21 pdurbin
'all the files in the data directory (the one that contains the "base", "pg_xlog", "pg_clog" etc folders)' http://wiki.postgresql.org/wiki/Corruption
18:22 pdurbin
base is "Subdirectory containing per-database subdirectories" according to http://www.postgresql.org/docs/9.1/static/storage-file-layout.html
18:22 bmckinney
ahh, ok
18:22 bmckinney
i'm created a mac-ansible playbook for postgres
18:23 bmckinney
i'll double-check locations and ownership, etc.
18:23 pdurbin
cool
18:23 pdurbin
bmckinney: do you have to run ansible as root?
18:23 bmckinney
how does glassfish resolve "/base..." ?
18:24 bmckinney
pdurbin: no, that was part of the exercise
18:24 pdurbin
great!
18:24 pdurbin
so maybe it *will* be a fix for https://github.com/IQSS/dataverse/issues/2443
18:24 bmckinney
i hope, but still banging my head a bit
18:25 bmckinney
i'll send you the playbook snippet separately so that you can flag anything
18:28 pdurbin
bmckinney: I would need instructions on how to run it as well
18:29 pdurbin
bmckinney: interesting, so you're using homebrew, like I am
18:29 bmckinney
pdurbin: so, i'm curious how glassfish would resolve "base/16384/16510" to "/usr/local/var/postgres/base/16384/16510"
18:30 bmckinney
that seems to be the issue
18:30 pdurbin
yeah, I dunno
18:30 bmckinney
when i ping the database pool it seems happy
18:30 pdurbin
works fine when I install postgres via homebrew manually, without ansible
18:31 pdurbin
psql (PostgreSQL) 9.2.2 ... installed years ago
18:34 pameyer
are you templating the postgres configs?
18:35 pdurbin
I certainly am not. :)
18:35 pameyer
I'd think glassfish wouldn't be seeing the path directly - assuming ports
18:38 pdurbin
there are a number of options at http://www.postgresql.org/download/macosx/ but homebrew has worked out ok for me. I think I'm the only one using homebrew though
18:38 pdurbin
pameyer: you have a dev environment set up?
18:40 pameyer
pdurbin: for me, yes and no
18:40 pameyer
enough to build and read code on; but no local postgres setup
18:40 pdurbin
ah, ok. just enough to build the war file
18:41 pameyer
I was assuming that a dev setup would have glassfish talking to postgres over sockets rather than files; asumption = don't know for sure...
18:41 pameyer
yup
18:41 pdurbin
buh. it's either sockets or ports. not files
18:42 cnk joined #dataverse
18:42 pameyer
right - but 'base/16384/16510' that bmckinney ran into made me think something was trying to use files
18:42 jri joined #dataverse
18:43 pdurbin
bmckinney: I think postgres is trying to bubble up to glassfish an error it thinks is helpful. I'm surprised you see stuff about "base" in server.log
18:43 pdurbin
seems quite low level
19:45 pdurbin
pameyer: oh, you can add "for testing only" like at https://demo.dataverse.org to the top of your installation if you want
19:46 pdurbin
and when you click it you get a popup with more detail
19:58 pameyer
pdurbin: good suggestion - thanks
19:58 bmckinney
when i tail the postgres server log during a deployment:
19:58 bmckinney
LOG: autovacuum launcher started LOG: database system is ready to accept connections ERROR: relation "sequence" does not exist at character 22 STATEMENT: SELECT SEQ_NAME FROM SEQUENCE WHERE SEQ_NAME <> SEQ_NAME ERROR: column "fk_dataverse_id" does not exist STATEMENT: CREATE INDEX INDEX_DATAVERSE_fk_dataverse_id ON DATAVERSE (fk_dataverse_id)
20:05 pdurbin
bmckinney: honestly, some weird stuff like that in your logs *might* be normal... Were you able to deploy?
20:06 pdurbin
I'm seeing INDEX_DATAVERSE_fk_dataverse_id at https://github.com/IQSS/dataverse/issues/1880 ... not sure what it's about.
20:10 bmckinney
pdurbin: it does deploy, but with lots of severe errors and warnings
20:10 bmckinney
i end up with this: "javax.persistence.NoResultException: getSingleResult() did not retrieve any entities."
20:11 pdurbin
check out https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-deploy-develop/9/consoleText (from today) about INDEX_DATAVERSE_fk_dataverse_id ... seems normal (but probably a bug).
20:12 pdurbin
bmckinney: if you feel like creating (and working on) that INDEX_DATAVERSE_fk_dataverse_id issue please go ahead.
20:13 bmckinney
right, not sure how much time i should spend on a mac-developer-install at this point
20:13 bmckinney
might just have to punt for now and go with vagrant
20:13 pdurbin
well, the point from that consoleText log is that we are seeing that INDEX_DATAVERSE_fk_dataverse_id error on Linux as well
20:14 bmckinney
feels like a race condition to me
20:14 bmckinney
the table is there and the error disappears after a redeploy
20:14 pdurbin
huh. dunno
20:16 pdurbin
"javax.persistence.NoResultException: getSingleResult() did not retrieve any entities" is "normal" for when you have deployed the war file but not yet run the other stuff like setup-all, reference data: https://github.com/IQSS/dataverse/blob/v4.2.4/scripts/deploy/phoenix.dataverse.org/post
20:17 bmckinney
ahh, good
20:17 pdurbin
(you can ignore those tests I run on phoenix)
20:28 bencomp joined #dataverse
20:35 bmckinney
pdurbin: setup-all did the trick, thanks!
20:39 pdurbin
bmckinney: sure thing. did you also run the reference data?
20:39 pdurbin
add* the reference data
20:40 bmckinney
pdurbin: yes, that was already part of the ansible playbook
20:42 pdurbin
ah. ok
21:01 pameyer joined #dataverse
