Time
S
Nick
Message
00:54 shauna joined #dataverse
15:12 iqlogbot joined #dataverse
15:12 Topic for #dataverse is now Dataverse: http://dataverse.org | logs at http://irclog.iq.harvard.edu/dataverse/today
15:12 pdurbin
whoops, iqlogbot was MIA. let me say that again
15:12 pdurbin
donsizemore: I appreciate you helping that person trying to install Dataverse yesterday.
15:13 pdurbin
thank goodness for up arrow :)
15:13 donsizemore
pdurbin: our ID mgmt guy writes "I notice that the registration form does not fill in the Affiliation, though that is one of the attributes (eduPersonScopedAffiliation) that the IdP sends to the SP. The attribute is multi-valued; the values for me (from the SP's built-in Session handler) are: affiliation: alum
15:17 pdurbin
donsizemore: that's great that you'll be able to get those. I think we do too. Do you want to use them for anything?
15:18 donsizemore
pdurbin: i'm not sure if it will matter to us; just didn't know if you all had encountered that
15:19 pdurbin
donsizemore: we aren't doing anything with that attribute yet
15:19 pdurbin
but I'd be interested in hearing about use cases
15:20 pdurbin
some day I'd like to work on this: Shibboleth Groups: arbitrary attributes and regex support · Issue #1515 · IQSS/dataverse - https://github.com/IQSS/dataverse/issues/1515
15:26 pdurbin
donsizemore: Jon had asked for that. It was deferred: https://docs.google.com/document/d/12Qru8Gjq4oDUiodI00oObHJog65S7QzFfFZuPU3n8aU/edit?usp=sharing
15:29 axfelix joined #dataverse
15:31 pdurbin
michbarsinai: you around?
15:31 michbarsinai
pdurbin: Yep
15:31 axfelix joined #dataverse
15:32 pdurbin
michbarsinai: async/threading issue: https://github.com/IQSS/dataverse/issues/2453
15:36 michbarsinai
pdurbin: woah... this does not read nice.
15:39 pdurbin
nope
15:39 pdurbin
michbarsinai: linked from that issue is a google doc from which you can download the tiny database dump I can reproduce the error with
15:40 pdurbin
it's only 826 KB
15:43 pdurbin
wow. hmm. I can actually reproduce it with less data
15:43 michbarsinai
Are you doing anything with the result of the call to indexing?
15:43 pdurbin
nope
15:44 michbarsinai
Do you return Future<X> or void?
15:44 pdurbin
Future
15:44 michbarsinai
Try to return void. This way the calling code can terminate rather than wait for completion.
15:45 michbarsinai
It's probably a deadlock where both threads wait for each other to give up locks. A got x and waits for y, B got y and waits for X.
15:45 michbarsinai
Or the like
15:45 michbarsinai
*x, not X
15:50 pdurbin
oh wow. with the system in a bad state like this I can't even log in
15:50 pdurbin
michbarsinai: I hear you about void. Let me upload an even smaller database that can be public.
15:51 michbarsinai
But you can logout any time you want :-)
15:52 pdurbin
let's just say I pushed a helpful script: https://github.com/IQSS/dataverse/blob/a21607884d55141e5e48ebee4b061cc5afffad23/scripts/database/homebrew/kill9glassfish
16:44 cnk joined #dataverse
16:55 pdurbin
michbarsinai: do you mean I should "find usages" for all Future and change to void? There are quite a few.
16:57 michbarsinai
Nah. Change the return type, and let the compiler do that work for you. Better yet - try NetBeans refactoring "change method parameters" to change the retrun type. Might be able to do so automatically.
16:58 pdurbin
sure, that's what I'm trying but...
16:58 pdurbin
michbarsinai: I'm still wondering the best way to communicate back the status of indexing
16:59 michbarsinai
If you need to communicate it back, it can't be async
16:59 pdurbin
yeah
16:59 pdurbin
michbarsinai: should we be using RabbitMQ or something?
17:00 michbarsinai
You can place some callbacks, but if you're waiting on those callbacks it still wouldn't be async...
17:01 michbarsinai
pdurbin: any additional infrastructure makes this app a harder sale. What are you planning to do with the result?
17:01 pdurbin
michbarsinai: well, at a high level, it's quite a bit of effort to keep Solr in sync with the database.
17:04 michbarsinai
Which is why you should'nt do it. You can't really... this is why they came up with the "eventually consistent" relaxation.
17:05 pdurbin
I can't? I'm required to.
17:05 michbarsinai
We're OK if Solr is updated 1-2 minutes later, I think
17:06 michbarsinai
I don't think anyone envisoined you doing distributed transactions between Solr and PostgreSQL.
17:06 pdurbin
michbarsinai: oh, sure, but that's not what I'm talking about.
17:15 axfelix joined #dataverse
18:21 michbarsinai joined #dataverse
18:37 michbarsinai joined #dataverse
19:12 donsizemore joined #dataverse
19:38 axfelix joined #dataverse
19:49 buckwheat joined #dataverse
19:54 buckwheat
Hello All! Having a few troubles getting this up and running but getting close. When I deploy the dataverse app I get a "javax.persistence.NoResultException: getSingleResult() did not retrieve any entities." error. Mean anything to anyone?
19:59 pdurbin
buckwheat: hooray! beleive it or not, you are so so close now :)
20:00 buckwheat
told you i dont give up easily...not sure why this has been such a difficult install since many seem to have had no problems. Can you guide me a little on where this may be coming from?
20:00 pdurbin
buckwheat: you need to create your root dataverse with the "setup all" script
20:00 pdurbin
the installer (written in perl) should call the "setup all" script for you but you can run it manually
20:01 buckwheat
i did that. deleted the db and ran it again. noticed that the curl call to set root dataverse failed. do i need to delete all again or can i just run that section of script...many env vars being set so just running a "section" may not work
20:02 pdurbin
set or create?
20:03 pdurbin
"curl call to set root dataverse failed". do you mean create?
20:04 buckwheat
this is the line that failed : curl -s -H "Content-type:application/json" -X POST -d @data/dv-root.json "$SERVER/dataverses/?key=$adminKey"
20:05 buckwheat
not sure if you call that set or create
20:05 pdurbin
create
20:05 pdurbin
buckwheat: what was the error, please?
20:06 buckwheat
it was a curl(52) Empty response from server...not always an error per se, but seems like it did not take
20:06 pdurbin
huh. weird
20:06 pdurbin
buckwheat: anything in server.log?
20:08 buckwheat
oddly enough, nothing helpful. Im going to burn it down one more time and run thru it again. Quick question, why did you guys run solr in jetty when you have glassfish?
20:09 pdurbin
buckwheat: good question. Is it easy to?
20:10 buckwheat
dont know. running solr in tomcat is a breeze...new to glassfish so cant say
20:10 pdurbin
wow, this is totally new for Solr 5 vs. Solr 4: http://lucene.apache.org/solr/quickstart.html
20:11 pdurbin
we probably want SolrCloud someday per https://github.com/IQSS/dataverse/issues/2322 but I'm not sure if we can run that in Glassfish
20:52 axfelix joined #dataverse
21:35 axfelix joined #dataverse
22:11 pdurbin
sweet! https://wiki.archivematica.org/Dataverse
22:32 axfelix joined #dataverse
23:10 axfelix joined #dataverse
