Time
S
Nick
Message
00:21 searchbot joined #dataverse
00:27 pdurbin joined #dataverse
08:20 bencomp joined #dataverse
08:48 jri joined #dataverse
08:48 jri left #dataverse
09:06 romainM joined #dataverse
09:06 romainM
hi
09:25 bencomp joined #dataverse
11:58 pdurbin
romainM: good morning
12:58 romainM
afternoon, actually
12:58 romainM
:D
12:58 romainM
but good morning to you ;)
13:45 pdurbin
romainM: you said hi at 4 in the morning :)
13:46 romainM
pdurbin: ya ya, I know it was a little bit early :D
13:47 romainM
pdurbin: just jocking here ;)
13:47 pdurbin
:)
13:48 pdurbin
romainM: can your server reach the outside world?
13:48 romainM
our dataverse server ? Yes, but there is a log entry to do so
13:49 pdurbin
ok. I was wondering about "edu.ucsb.nceas.ezid.EZIDException: bad request - invalid target URL " from your logs
13:53 romainM
maybe we should try an open server
13:54 romainM
but the thing is that we had no problems before, with a close server so ...
13:54 romainM
dunno what happened :S
13:54 pdurbin
romainM: ok, that's something else I was wondering... so it used to work
13:54 pdurbin
romainM: what changed?
13:54 romainM
yup
13:55 romainM
upgrade to the newer version of dataverse
13:55 romainM
the thing is, I didn't tested the new dataverse environnement after that upgrade
13:55 romainM
so I can't say if it's due to this or not
13:56 romainM
cause just after this update, we opened the server with a log entry to access it
13:58 pdurbin
I'm not sure what that means. "we opened the server with a log entry to access it". Do you mean you can see server.log for the first time?
14:00 romainM
ah, no
14:00 romainM
there is a login/password system to access our server from outside
14:01 romainM
I mean, to acess the dataverse url
14:06 pdurbin
ah. ok
14:16 jri joined #dataverse
14:16 romainM
re jri
14:17 romainM
pdurbin: jri is our system admin
14:17 pdurbin
jri: hi! welcome!
14:20 jri
Hi :)
14:22 romainM
jri is more suitable than me if you have questions related to the serve conf, etc
14:22 romainM
*server
14:23 pdurbin
jri: I opened this issue yesterday because it's hard for me to know what's going wrong: https://github.com/IQSS/dataverse/issues/2780
14:25 pdurbin
jri: you could try increasing debugging on a per class (or package, I believe) basis: http://guides.dataverse.org/en/4.2.2/developers/debugging.html
14:31 sekmiller joined #dataverse
14:31 pdurbin
sekmiller: does "edu.ucsb.nceas.ezid.EZIDException: bad request - invalid target URL " mean anything to you? sounds bad :)
14:32 romainM
hi sekmiller
14:32 sekmiller
is it in your local environment?
14:32 pdurbin
sekmiller: now, I'm able to publish datasets just fine. romainM and jri are having trouble
14:32 pdurbin
no*
14:34 pdurbin
sekmiller: it used to work for them. we aren't sure why it stopped working
14:34 jri
Ok edu.harvard.iq.dataverse.api.Datasets is now to "FINE" debug level, but it don't seems to provide more info
14:35 pdurbin
jri: well, that class is only exercised when you use the "datasets" API endpoint
14:35 jri
Here are the last logs: http://pastebin.com/MYvWzGjY
14:35 jri
Arf
14:35 pdurbin
jri: maybe you could try turning up debugging on edu.ucsb.nceas.ezid
14:35 jri
ok
14:35 sekmiller
do you have a system property set for dataverse.siteUrl?
14:38 jri
pdurbin: no more info with edu.ucsb.nceas.ezid... Could be useful to do it on edu.harvard.iq.dvn.core.index.DOIEZIdServiceBean ?
14:38 jri
or edu.harvard.iq.dataverse.DatasetPage ?
14:39 pdurbin
jri: you should listen to what sekmiller is saying. he wants you to check a JVM option
14:39 sekmiller
The doi service bean is where the target url is created.
14:40 sekmiller
in the method getSiteUrl it forst looks for the system setting of dataverse.siteUrl
14:40 pdurbin
and failing that, it looks for dataverse.fqdn it seems
14:40 jri
dataverse.siteUrl should be JVM option ?
14:42 pdurbin
sekmiller: it looks to me that either dataverse.fqdn or dataverse.siteUrl needs to be configured. I guess it shouldn't hurt to configure both...
14:42 sekmiller
yes
14:42 pdurbin
jri: do you have either one configured?
14:43 sekmiller
right.
14:45 jri
I only have dataverse.fqdn in my domain.xml
14:45 jri
but, it's not the fqdn
14:45 jri
only the hostname of my server
14:46 pdurbin
jri: can you please try making dataverse.fqdn a fqdn?
14:46 jri
yep
14:51 jri
Ok it's seems to work now \o/
14:51 romainM
woohoo
14:51 pdurbin
\o/
14:51 pdurbin
sekmiller: thanks!!
14:52 romainM
so ... the problem was the "fqdn" thing ?
14:52 sekmiller
you're welcome. Soemthing for the developers' guide?
14:52 jri
Is it the install script that automatically put this parameter or me? don't remember...
14:53 pdurbin
jri: the install script adds it
14:53 pdurbin
sekmiller: jri and romainM are installing Dataverse. it should go in the Installation Guide
14:53 sekmiller
OK
14:54 pdurbin
jri and romainM would you be able to read and comment on https://docs.google.com/document/d/1GLky44JbtCNhGuavLXGEP8WPNdjbyHNxalEGjR78MaA/edit?usp=sharing ? I can even give you edit access if you want. We'd like to improve the Installation Guide.
14:54 romainM
just a sec, not sure it works
14:56 pdurbin
I just added a bit under JVM options
14:56 pdurbin
sekmiller: maybe think about how the error messages could be improved
14:57 pdurbin
and maybe the install script should force you to enter an fqdn? not sure
14:57 pdurbin
romainM: you can't get into that Google Doc?
14:57 pdurbin
it should be public and allow comments
14:58 romainM
wasn't talking about that, but the "previous" bug with dataverse
14:58 romainM
I can't publish actually
14:58 romainM
both on internal and external
14:58 romainM
:S
14:58 pdurbin
then why did jri say it works now? :)
14:59 romainM
it seems he did publish one thing
14:59 romainM
aaaaaand, it's gone
15:00 pdurbin
romainM: the problem is gone?
15:00 romainM
the problem supposed to be gone is back
15:00 pdurbin
d'oh!
15:00 romainM
dunno how jri could publish one thing, when now I can't
15:00 romainM
he can't too it seems
15:01 romainM
I ... don't understand
15:02 pdurbin
sekmiller: any ideas? we're tracking this issue at https://help.hmdc.harvard.edu/Ticket/Display.html?id=229931 so we can follow up there later if need be
15:02 pdurbin
jri: the nuclear option is to reinstall Dataverse, I guess
15:03 sekmiller
That one dataset that you published, can you clcik on the link in the citation?
15:03 romainM
yup
15:03 romainM
and it's not like the unpublished ones
15:04 romainM
it redirect to the dataset
15:04 romainM
the only one that has been published
15:04 romainM
during the brief moment jri could publish
15:05 romainM
as opposed to the unpublished ones that redirect here : https://www.datacite.org/testprefix
15:05 romainM
when clicking on the link
15:06 romainM
(the unpublished has DVN in their url, FK2 the published one)
15:06 pdurbin
are you still seeing "invalid target URL " in server.log?
15:07 romainM
ok, it seems it's solved
15:08 pdurbin
good!
15:08 pdurbin
what was it?
15:08 romainM
the fqdn solved a part of the prob
15:09 pdurbin
ok
15:09 romainM
it's just a change in the server conf that occured this morning, when we tried something from the answer of my mail to your support
15:09 romainM
update setting set content='10.5072/FK2' where name like '%Authority%';
15:09 romainM
that contained DVN instead of FK2
15:09 pdurbin
oh. yeah
15:09 pdurbin
FK2 is for testing
15:09 romainM
the published dataset stil had the good thing, so it passes ty to the fqdn
15:10 romainM
(correct me if I'm bullshiting here jri)
15:10 pdurbin
you should be using FK2 if you're just testing. that's what we use on our demo site. here for example: https://demo.dataverse.org/dataset.xhtml?persistentId=doi:10.5072/FK2/WPYBFR
15:12 jri
Should we have to buy EZID subscription for production environment ? :/
15:13 pdurbin
jri: I think you have to. sekmiller would know
15:13 romainM
oh
15:13 romainM
that's something we didn't knew
15:14 sekmiller
for a production environment, yes you'd need to make an arrangement with EZID.
15:17 jri
Well, we have to announce that to the project management team :D
15:19 romainM
we may have an other doi provider
15:20 romainM
dunno how it works, if we can use it, etc
15:21 romainM
but anyway, thanks a lot for your help
15:22 romainM
was very odd to say "hey, we installed dataverse, try it!" and then say "hey ! you can't publish"
15:22 romainM
^^
15:22 romainM
at least it's working perfectly now
15:23 pdurbin
romainM: I hear you. You probably don't want to publish fake/test DOIs though...
15:24 romainM
given that it's a testing dataverse, that's not a prob
15:24 romainM
now that I think about it
15:24 romainM
I have to find the way to delete datasets in db ...
15:28 pdurbin
romainM: there's a "destroy" endpoint in the datasets API
15:30 romainM
the delete dataset command doesn't test if the dataset is draft or not ?
15:31 pdurbin
romainM: it checks if it's published
15:32 romainM
destroy is an other command then ? didn't find it in the doc yet
15:32 pdurbin
romainM: it's API only
15:32 pdurbin
superuser only
15:32 romainM
ah
15:33 jri
pdurbin or sekmiller: just to be sure, for production I need to have "dataverse.fqdn" as my real machine FQDN and "dataverse.siteUrl" as my public URL ? Our Dataverse will be behing reverse proxys and the machine name is not the public domain
15:35 pdurbin
jri: those JVM options do almost the same thing. in both cases, it's for the hostname or URL you want to promote to the outside world, not the machine name
15:37 jri
Ok I see... Configuring dataverse.fqdn with the public domain (that I want to promote) don't degrade other functionalities ?
15:38 pdurbin
jri: nope
15:38 pdurbin
go for it
15:38 jri
Cool :) Thanks!
15:39 pdurbin
oh sure
15:42 donsizemore joined #dataverse
15:48 pdurbin
jri: were you able to get into the Google Doc?
15:48 pdurbin
donsizemore: oh hi
15:49 pdurbin
donsizemore: you might be interested in this: https://github.com/IQSS/dataverse/issues/952#issuecomment-161002345
15:50 donsizemore
excellent =)
15:50 pdurbin
bricas: it turns out fixing `vagrant up` helped another guy so thanks again
15:52 donsizemore
i'm working on an ansible playbook... wondering whether it's better to stick with some minor tweaks to the installation scripts included in the release zip file, or crib from the vagrant stuff.
15:54 pdurbin
donsizemore: you might want to crib from https://github.com/IQSS/dataverse-puppet
15:54 donsizemore
i really like your vagrant scripts, but trying to trace where it's getting /dataverse/scripts/installer
15:55 pdurbin
donsizemore: we'll, the git repo gets mounted at /dataverse
15:55 pdurbin
if that helps
15:56 bencomp
donsizemore: I'm interested to see your playbook(s)!
15:58 pdurbin
I think https://github.com/sbgrid is using Ansible too.
15:58 donsizemore
yes =) bencomp: it's pretty much spaghetti ansible from the published installation guide. i toddled along fine until i hit the "asadmin login" bit and am trying to find the right combination of cookies and ice cream to get it to store the login info
15:59 donsizemore
everything works until the install script wants to restart the domain
15:59 bencomp
donsizemore: is the playbook calling the shell script?
16:00 bencomp
as I understand Ansible, restarting things would be a good use for handlers
16:03 donsizemore
how often do your install scripts change? plan A was to have Ansible unzip dvinstall.zip and call a slightly modified script for forwards compatibility. not what you all want long-term, i know.
16:04 jri
pdurbin: Yep I can access to it, I'll read and try to help when I have more time
16:11 bencomp
donsizemore: if we ever make it to go 'fully Ansible' for DataverseNL, I would like to have all configuration info in our local repository and use the generic playbooks to update the service on the server(s)
16:12 donsizemore
agreed. this is my first stab it at, so as you asked above i was trying to call the install script from ansible and just "make it work"
16:12 donsizemore
i must run, back in a bit. thank you both for your help =)
16:12 donsizemore left #dataverse
16:13 bencomp
maybe we should collaborate and ask pdurbin for the repo to hold the generic Dataverse playbooks
16:22 pdurbin
jri: thanks
16:23 pdurbin
bencomp: I would be more than happy to create an equivalent of https://github.com/IQSS/dataverse-puppet for Ansible and give you and others access
16:23 pdurbin
may the best configuration management framework win ;)
16:25 pdurbin
it would be awesome to set up tests for dataverse-puppet or dataverse-ansible: https://github.com/IQSS/dataverse-puppet/issues/10
16:26 bencomp
pdurbin: I think it's time to prevent people doing the same work and create the dataverse-ansible repo. could you please create it?
16:29 pdurbin
bencomp: done! https://github.com/IQSS/dataverse-ansible
16:29 bencomp
pdurbin: starred! ;)
16:30 pdurbin
now someone just needs to do this: git remote add origin git
16:30 pdurbin
:)
17:48 pdurbin
shauna: do you feel like helping me edit that job ad I showed you? http://irclog.iq.harvard.edu/dataverse/2015-11-05#i_26832
18:00 donsizemore joined #dataverse
18:00 donsizemore
pdurbin: thanks for the invite =) let me get what i have a little more presentable, and i'll send you a link? could be a start if nothing else.
18:02 pdurbin
donsizemore: you can just go head and push to that repo
18:03 pdurbin
git remote add origin git
18:04 metamattj joined #dataverse
18:11 donsizemore
pdurbin: i certainly will, but i wrote it off the top of my head over thanksgiving and just started testing it yesterday afternoon, etc. it needs a lot of work.
18:23 garnett joined #dataverse
18:23 pdurbin
donsizemore: oh sure, no rush
18:24 garnett joined #dataverse
18:24 pdurbin
donsizemore: so check this out: https://github.com/IQSS/dataverse/blob/4.2.2/scripts/deploy/phoenix.dataverse.org/deploy
19:03 donsizemore joined #dataverse
19:16 donsizemore
pdurbin: take a peek at https://github.com/IQSS/dataverse-ansible - far from perfect, but works until mid-way through the installer script. you'd want to break that out into ansible tasks anyway, but there it is.
19:25 bencomp joined #dataverse
19:33 dataverse-user joined #dataverse
19:42 pdurbin
donsizemore: nice!!! Hey, any familiarity with Travis? Can I open an issue similar to https://github.com/IQSS/dataverse-puppet/issues/10 ? :)
20:15 donsizemore joined #dataverse
20:42 axfelix joined #dataverse
20:54 bencomp
donsizemore: great work on dataverse-ansible
20:55 bencomp
do you keep the variables locally? and where is the role 'dataverse' defined?
21:10 jri joined #dataverse
21:18 donsizemore
we'd want to add an example group_vars file (mine is outside the git tree)
21:18 donsizemore
ansible will automatically select tasks/main.yaml; i used includes to split up the prereq and install bit.
21:19 donsizemore
phil sent me his phoenix stuff so i'm reading through it and matching up variables
21:20 donsizemore
mine is also centos7-specific (i was writing it to cobble it into the iPlant Discovery Environment playbooks) so we'll need to add distro checks to play nice with Debian
21:23 pdurbin
donsizemore: ah. so maybe your Ansible stuff wouldn't play nicely with Travis since you're targeting centos (which makes sense) and Travis uses ubuntu.
21:28 donsizemore
pdurbin: it could pretty easily be made OS -agnostic with ansible_distribution checks
21:28 bencomp
donsizemore: I've been thinking a bit, and I'm not sure yet, that perhaps this repo should be a 'pure' role that can be included in a site.yml and galaxy. it could have a defaults folder
21:29 pdurbin
donsizemore: cool. what I want (maybe not explained well enough in that issue) is automated integration tests on Travis like this: https://travis-ci.org/openscholar/openscholar/jobs/94158615
21:33 pdurbin
basically, Travis spins up an Ubuntu VM , installs Apache, Solr, etc.
21:33 pdurbin
Ansible playbook for Dataverse - Google Groups - https://groups.google.com/forum/#!topic/dataverse-community/VsoDw71PfBI
21:33 pdurbin
:)
21:47 donsizemore
bencomp: definitely. i just offered what i had as a starting point
21:54 bencomp
donsizemore: do you mind if I create a short readme?
22:08 pdurbin
donsizemore bencomp: hey I just added https://github.com/pameyer (from https://github.com/sbgrid ) to https://github.com/orgs/IQSS/teams/dataverse-ansible
22:17 bencomp
great. I just added a README :)
23:29 metamattj joined #dataverse
23:55 pdurbin
+1
