IQSS logo

IRC log for #dataverse, 2016-02-25

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
08:28 bencomp joined #dataverse
09:31 majest1c joined #dataverse
10:55 xvx joined #dataverse
12:47 bencomp joined #dataverse
13:35 mckinney joined #dataverse
14:20 pdurbin bencomp: hi. Do you plan to use more than one Glassfish server with Dataverse 4? If so, I have a question related to Shibboleth. bricas how about you? mckinney do we even know where the HMS instance will be hosted for sure, how many servers, and if Shibboleth will be used? :)
14:22 bencomp pdurbin: no, we don't plan to. But please ask :)
14:24 pdurbin actually, hmm
14:24 pdurbin maybe it's fine
14:25 pdurbin maybe I'll throw away this email I've been writing
14:25 pdurbin :)
14:28 pdurbin my observation is that if I curl https://dataverse.harvard.edu/Shibboleth.sso/Metadata I get Location="https://dataverse.harvard.edu/Shibboleth.sso/SAML2/POST" (for example), which is good. I was surprised that if I curl the *actual* "machine name" (or whatever you call it) for each of the three glassfish servers the "Location" changes. So I guess it's dynamic. Probably fine. We'll just always give out the
14:28 pdurbin production URL when exchanging metadata.
14:29 bencomp my little experience is that Shibboleth is pretty good with proxies
14:37 pdurbin bencomp: any thoughts on how imporant it is to back up /etc/shibboleth/sp-cert.pem and /etc/shibboleth/sp-key.pem if you've already shared metadata with InCommon or some other federation?
14:37 pdurbin heh. I guess I wrote that it's important to back up these files: http://guides.dataverse.org/en/4.2.4/installation/shibboleth.html#backups :)
14:37 pdurbin but I don't explain why
14:38 pdurbin it's probably not just me being paranoid
14:38 pdurbin I'm pretty sure the public key shows up in the metadata
14:40 bencomp I don't know if you can recreate a .pem  or both .pems from the metadata
14:41 bencomp it might be easier to restore the files directly from a backup
14:43 pdurbin yeah
14:43 pdurbin backups++
14:43 pdurbin bencomp: very unlikely that the metadata contains the *private* key
14:47 * pdurbin adds a reminder to https://github.com/IQSS/dataverse/issues/2953
15:33 bencomp1 joined #dataverse
15:39 pameyer joined #dataverse
15:45 pdurbin mckinney: how do you feel about me assinging https://github.com/IQSS/dataverse/issues/2728 to you too? to think about. /cc pameyer
15:59 mckinney pdurbin: yes, i'll take that one - it seems highly relevant to our requirements & use cases
16:00 pdurbin mckinney: please see if you have permissions to assign it to yourself
16:00 pameyer I don't know if object storage fits into that one
16:01 mckinney pdurbin: pameyer can fill you in on HMS hosting plans - not sure yet if shibboleth will come into play
16:02 pameyer if shibboleth is that least-complexity path to getting ldap auth working, then possibly
16:02 pameyer that route *could* be through having ldap produce something that shibboleth can understand (which might be on the table for other stuff)
16:13 mckinney joined #dataverse
16:21 pdurbin pameyer: we really aren't planning on supporting LDAP. Can you use Shibboleth instead? We can talk about this in person tomorrow. :)
16:26 pameyer pdurbin: sounds like a plan
16:27 pdurbin cool
16:27 pdurbin pameyer: also, this is the issue about object storage: https://github.com/IQSS/dataverse/issues/1347
16:27 pdurbin pameyer: S3 support is how I think of it. But I don't know if you're interested in that or not.
16:28 pameyer we're interested in S3 for uploads at least
16:28 pdurbin pameyer: interesting. ok
16:29 pdurbin let me find a branch. hold please
16:29 pameyer the reason I wasn't sure if object storage fit into 2728 was fopen/cloud api call
16:29 pameyer but it could go either way
16:29 pameyer no hurry :)
16:30 pdurbin https://github.com/IQSS/dataverse/commits/2909-MOC-collab-swift but more useful to read is probably https://github.com/IQSS/dataverse/issues/2909 . Note that Swift basically S3.
16:31 pameyer cool - hadn't known swift ~= s3
16:32 pameyer both are on the list of stuff to learn more about, but as 2 things...
16:36 pdurbin yeah, this is interesting. good stuff: https://github.com/IQSS/dataverse/blob/e67238071b1f1d39e90993280f671e55cf2bd9cd/src/main/java/edu/harvard/iq/dataverse/dataaccess/SwiftAccessIO.java
16:39 pdurbin pameyer: I'm thinking about adding you to https://github.com/orgs/IQSS/teams/dataverse-readonly (like I did for mckinney) so I can @mention you in issues.
16:40 pameyer public InputStream openSwiftFileAsInputStream :)
16:40 pameyer pdurbin: I thought you'd bee @mentioning me in issues already
16:42 pdurbin pameyer: I have been but it doesn't autocomplete and I'm lazy. Do you mind if I add you?
16:43 pameyer nope
16:44 pdurbin pameyer: added but you have to accept
16:50 pameyer pdurbin: looking for the accept option ...
16:58 pameyer ...not seeing it
17:03 pdurbin pameyer: you seem to be in. thanks
18:52 mckinney joined #dataverse
19:52 pdurbin nice list of Shibboleth/SAML federations (including InCommon): http://www.protectnetwork.org/support/faq/identity-federations
20:29 bencomp joined #dataverse
21:18 pdurbin pameyer mckinney: shoot. I gotta run soon but I wanted to touch base with you about the sprint meetings. Specifically, are you ok with me creating a folder in Google Drive that's public with a doc for each meeting? The docs would allow public comments.
21:22 pdurbin or we could just take notes right here in this channel :)
21:22 mckinney either way is fine by me
21:22 pdurbin might get a little messy/annoying though
21:23 mckinney i plan to add them to a trello board that would be available to anyone on the team who is interested
21:23 pdurbin oh!
21:24 mckinney we can discuss what might work best tomorrow
21:25 mckinney have a good night
21:27 pdurbin mckinney pameyer: here, for now, please add agenda items here: https://docs.google.com/document/d/1gZ1AkqS3DAwBHLCm6rkzROSKPXHGRCD7aK5mlTw9XAg/edit?usp=sharing
21:29 pdurbin can you both edit that doc?
21:31 pdurbin bencomp: is DANS part of one of these? http://www.protectnetwork.org/support/faq/identity-federations
21:32 axfelix joined #dataverse
21:33 bencomp pdurbin: no
21:33 pdurbin oh. I forget what yours is called
21:33 pdurbin or maybe yours isn't a federation
21:34 bencomp SURFconext is "our" federation
21:35 pdurbin searchbot`: lucky SURFconext
21:35 searchbot` pdurbin: https://www.surf.nl/en/services-and-products/surfconext/index.html
21:35 pdurbin why isn't it listed on that other page then? where is the best list of all the federations?
21:35 bencomp edugain?
21:36 pdurbin searchbot`: lucky edugain
21:36 searchbot` pdurbin: http://services.geant.net/edugain/Pages/Home.aspx
21:36 pdurbin hmm, can't find a list there. anyway, gotta go pick up the kids
21:37 bencomp o/
21:37 pdurbin have a good night, #dataverse
21:38 pdurbin xvx: I hear https://trello.com/c/7AQ5V9aJ/28-about-page-replicate-others-work-vs-replicate-others-work is fixed but it's probably cached for a bit. Please check later and thanks again.
21:39 xvx jolly good, fix wasn't too long of a wait
21:44 pdurbin yeah, cache must be shorter than I realized
21:44 pdurbin ok, really going. o/
22:20 xvx joined #dataverse
22:25 pameyer pdurbin: is that agenda items for tomorrow or monday?
22:45 pameyer and I just saw pdurbin's "really going"...
22:46 pdurbin pameyer: tomorrow

| 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.