IQSS logo

IRC log for #dataverse, 2018-12-20

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
08:22 jri joined #dataverse
10:16 juancorr90 joined #dataverse
10:16 juancorr90 left #dataverse
10:17 juancorr joined #dataverse
13:30 donsiemore joined #dataverse
15:27 andrewSC joined #dataverse
15:41 jri joined #dataverse
15:42 pameyer joined #dataverse
15:46 pdurbin pameyer: morning. I hope you don't mind but I just took you off https://github.com/IQSS/dataverse/issues/5413 since I just approve the associated pull request
15:47 pameyer pdurbin: no problem
15:47 pdurbin cool
15:47 pameyer I'm not planning to stop taking a look at it though
15:47 pdurbin you are still very welcome to do whatever, leave a review, or course
15:47 pdurbin yes, good
15:52 pameyer pdurbin: did you run ITs on 5413?
15:54 pdurbin nope
16:03 pdurbin pameyer: just saw your comment. Thanks for running them!
16:03 jri joined #dataverse
16:03 pameyer pdurbin: no problem
16:09 pameyer shibboleth wizards - any tips for the best way to find out if ADFS 3.0 SAML works with dataverse shibboleth?
16:11 pdurbin pameyer: the information I have on this is 5 years old: https://github.com/IQSS/dvn/wiki/configuring-ADFS-relying-party
16:12 pameyer pdurbin: thanks
16:13 pdurbin and the guy I was working with left HMS in 2016
16:14 pameyer huh - just noticed that page was instructions for doing a demo install on windows 7
16:16 pdurbin I know about the wiki page you're talking about (this was the DVN 3.x days), but I'm not quite following.
16:16 pameyer https://github.com/IQSS/dvn/wiki/DVN-Install-Demo-on-Windows-7
16:17 pameyer not going down that trail; just surprised to see windows there
16:22 donsiemore @pameyer I've got one half of an ADFS 2.0 config, FWTW
16:26 pameyer @donsiemore I'm trying to figure out which versions of what play with what else.  and map the acronyms to things that make sense in my head...
16:53 donsiemore @pameyer as long as the chosen names are unfairly cruel to Active Directory, I'm all in!
16:53 pameyer :)
17:07 pdurbin pameyer: People in the Dataverse community have gotten "Institutional Login" to work with non-Shibboleth systems. Dataverse supports SAML. Shibboleth implements SAML. When you click "Login" at https://data.qdr.syr.edu you're auth'ing with SAML to a Drupal site, for example.
17:10 pameyer pdurbin: thanks.  my current thinking is that it would be connecting having adfs work as an idp for shibboleth
17:11 pameyer sometimes things don't stick in my head until I've gone a few rounds with them
17:11 pdurbin Yep. You're using the acronyms in the right places. :)
17:12 pameyer I'll count that as progress ;)
17:13 pdurbin You want ADFS to be an IDP that implements SAML in a configuration supported by Dataverse, especially with regard to passing required attributes such as "eppn" and "mail" in the assertion.
17:15 pameyer sounds like I shouldn't assume that the saml spec not having been changed for a while means that all saml implementations do it the same way
17:15 pameyer ^ doesn't mean that they all do it the same way
17:16 jri joined #dataverse
17:24 pdurbin Well, the nice thing about Shibboleth is that you can configure which attributes (eppn, mail, etc.) are released.
17:25 pdurbin Some systems use NameID instead of eppn so you have to map the former to the latter for Dataverse login to work.
17:25 pdurbin So you might need to fiddle with mappings a bit.
17:26 pdurbin Longer term we should make Dataverse's SAML support more like our OAuth support. We should not require so many attributes.
17:31 pameyer the optional email bit from OAuth is a nice way to handle it
17:37 pdurbin yeah, exactly, more like that
18:09 jri joined #dataverse
19:35 donsizemore joined #dataverse
20:24 donsizemore hey @pameyer pester pester?
20:24 pameyer hey @donsizemore - sure
20:24 pameyer what's on your mind?
20:26 donsizemore @pameyer akio is noting similarities in the DLM diagram presented in #3403 with the TRSA he's working on
20:27 donsizemore @pameyer he's asking if there is much source code up he could refer to. he's not finding much on GitHub
20:27 * pameyer look to refresh about which diagram that is
20:28 donsizemore @pameyer (mostly whether it's all in the sbgrid/DCM, or if there's more)
20:29 pameyer re-checking now, but I don't think that repository made it to github
20:29 pameyer and looks like it hasn't made it out into the sunlight yet
20:29 donsizemore @pameyer how may we bribe you to share a tarball?
20:30 pameyer @donsizemore - no bribes necessary, but I'll have to ask upstairs
20:31 pameyer there are conceptual similarities between DLM and TRSA; but DLM itself is pretty low complexity
20:31 pameyer the first iteration took a list of "storage sites" and list of datasets, and used globus to make sure that the datasets were at the correct sites (using a "primary" site as source of truth
20:32 pameyer ).  datasets were "everything in the data directory", and it never got to file level granularity
20:32 pameyer the second iteration was much the same, except that the list of datasets came from an OAI-PMH feed
20:33 pameyer the plan was/is to use the feed from dataverse, and provide an interface for reporting to dataverse the status of datasets at the other sites
20:33 pameyer but things have not gotten that far; other things took longer than expected :(
20:35 pdurbin pameyer: curious. Do you make use of `curl http://localhost:8080/api/admin/storageSites`?
20:36 donsizemore @pameyer how did you handle transferring large files between sites? OAI is metadata only?
20:36 pameyer pdurbin: not yet - that's the planned interface, but it didn't exist then
20:37 pameyer @donsizemore sometimes I miss obvious things.  globus was used for data transfer, OAI was just used for the identifier list
20:37 poiki-at-home joined #dataverse
20:37 pameyer designed for multiple protocols, but at the moment globus only
20:38 pdurbin pameyer: ok. Thanks. I'm just wondering if would make sense for donsizemore and Akio to also use the "storagesite" table.
20:39 donsizemore @pdurbin as long as we use irsync i'll be happy
20:39 donsizemore @pameyer and the DAA is a sensitive datastore?
20:39 pdurbin donsizemore: please see add-storage-site.json at http://guides.dataverse.org/en/4.9.4/developers/big-data-support.html
20:40 donsizemore @pdurbin i was just remembering this from setting up the DCM as a test last year
20:41 pdurbin poiki-at-home: not to distract donsizemore and pameyer but we were talking about you in #glassfish today. All good things! :) https://javabot.evanchooly.com/logs/%23glassfish/2018-12-20
20:42 poiki-at-home Hey guys, just read up the logs...
20:42 poiki-at-home Sorry I have not been around today - my son is ill and I need to look after him.
20:42 poiki-at-home Most certainly I wont be around tomorrow, too
20:43 poiki-at-home But I did not want to miss the chance to write to you guys
20:43 pdurbin hope he feels better!
20:43 poiki-at-home A big and warm thank you to all of you, all the others at IQSS for some very interesting months so far.
20:44 poiki-at-home As the end of the year approaches, some things start to settle for the beginning of the next and I hope this journey can continue as pleasant as it has started
20:45 poiki-at-home I wish you all and your families a Merry Christmas and "Guten Rutsch!" (Happy new year!)
20:46 poiki-at-home pdurbin: Yeah, he is ok. Not allowed to be in kindergarten tomorrow, so my job to entertain him... ;-)
20:47 pameyer @donsizemore DAA is a little orthogonal to sensitivity.  dataverse doesn't need direct access to DAA sites, the access restrictions are up to the sites
20:48 pameyer sorry - was interrupted, and will likely be interrupted again in a few minutes (hence the lag)
20:51 pdurbin poiki-at-home: well, maybe when he falls asleep you can catch up on the Docker stuff in #glassfish :)
20:52 poiki-at-home pdurbin: just read through it
20:53 poiki-at-home seems like basing on adams stuff could backfire...
20:53 poiki-at-home (although I had no plan to do so...)
20:54 pameyer @poiki-at-home did you see the google group post about a user trying for POC with openshift dataverse?
20:57 poikilotherm joined #dataverse
20:59 pdurbin poiki-at-home: I'm just looking to follow someone's lead. Your lead. Adam's lead. Whoever. :)
21:01 pdurbin donsizemore: I hope I didn't derail your TRSA/DLM stuff. I'm glad you got something out of the diagram from pameyer that I threw into that issue. :)
21:04 poiki-at-home pameyer: nope, not yet? could you share a link?
21:04 poiki-at-home pdurbin: we'll see ;-)
21:06 poiki-at-home pdurbin: I just left a comment on https://github.com/IQSS/dataverse/issues/5412#issuecomment-449134483 ... There IS a solution (or multiple), but I fear it will not be possible with no code changes at all
21:07 pdurbin poiki-at-home: here's the openshift thread pameyer mentioned: https://groups.google.com/d/msg/dataverse-community/kERoDtt95PE/bfXbMPAxBwAJ ... please note that there was also a thread from Jurn on openshift in January.
21:08 poiki-at-home Thx!
21:14 pdurbin poiki-at-home: here's the earlier openshift thread where I was asking him what he's up to: https://groups.google.com/d/msg/dataverse-community/IOeQ6umnbO4/kDmsGW9VAAAJ
21:17 poiki-at-home I will point him to the community issue 5373 tomorrow. Need to get some other stuff done before going to bed ;-)
21:17 pdurbin ok
21:18 poiki-at-home Have a good night everyone :-)
21:18 pdurbin you too
21:18 poiki-at-home left #dataverse
21:19 donsizemore i'm off. everybody have a wonderful new year!
21:20 pdurbin bye!
21:20 pameyer @donsizemore you too
21:28 pdurbin Dataverse 4.10 is out: https://github.com/IQSS/dataverse/releases/tag/v4.10
21:57 andrewSC joined #dataverse
22:00 jri joined #dataverse
22:14 andrewSC joined #dataverse

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.