Time
S
Nick
Message
03:02 nightowl313 joined #dataverse
03:09 nightowl313
Hi...I've upgraded dataverse to 4.18...how do I add the microsoft login option?
03:12 nightowl313
I don't see any instructions to add the "Other Options" link
03:25 nightowl313 joined #dataverse
03:25 nightowl313
oh nevermind I just found the new 4.18 procedures!
08:15 jri_ joined #dataverse
08:34 poikilotherm joined #dataverse
08:35 jri joined #dataverse
08:52 juancorr joined #dataverse
10:15 dataverse-user joined #dataverse
10:18 dataverse-user
Good morning, my name is Flavio Castro, and I'm responsible for the data repository of the Federal University of São Paulo. We used Dataverse and would like our institution to appear on the map. How should I proceed ? our address is https://repositoriodedados.unifesp.br Thanks.
11:24 pdurbin joined #dataverse
11:25 pdurbin
Shoot. I missed Flavio. To get on the map, one should email support
11:25 pdurbin
Cool that there's interest in the new Microsoft login option. :)
11:34 * pdurbin
emails him
12:31 yoh joined #dataverse
12:59 jri joined #dataverse
13:29 donsizemore joined #dataverse
13:36 stefankasberger joined #dataverse
14:09 jri_ joined #dataverse
14:45 Youssef_Ouahalou joined #dataverse
14:46 Youssef_Ouahalou
[root
14:46 Youssef_Ouahalou
sorry
14:46 Youssef_Ouahalou
Hello everyone
14:46 Youssef_Ouahalou
i have a error when i do a update of dataverse
14:47 Youssef_Ouahalou
Detail: Key (seq_name)=(SEQ_GEN) already exists. Command deploy failed.
14:47 Youssef_Ouahalou
Do you know where it comes from?
14:47 pdurbin
Youssef_Ouahalou: be careful when you're logged in as root. :)
14:48 pdurbin
I'm just kidding.
14:49 pdurbin
Please email your server.log file to support
14:49 Youssef_Ouahalou
you think the mistake comes from there? From root
14:50 pdurbin
No, no. I'm just reacting to your copy paste and "sorry". :)
14:50 Youssef_Ouahalou
ok nice thank you :)
14:50 pdurbin
There's probably an error in server.log that will reveal the problem.
14:57 Youssef_Ouahalou
Is my ticket #283260
14:57 pdurbin
Thanks. I see it. It looks like you're upgrading from 4.14 to 4.15.
14:58 Youssef_Ouahalou
yes is correct
14:58 pdurbin
PER01000: Got SQLException executing statement "INSERT INTO SEQUENCE(SEQ_NAME, SEQ_COUNT) values ('SEQ_GEN', 0)": org.postgresql.util.PSQLException: ERROR: duplicate key value violates unique constraint "sequence_pkey"
15:00 pdurbin
Hmm, but "variablemetadata" is the only table we altered in that release: https://github.com/IQSS/dataverse/blob/v4.15/src/main/resources/db/migration/V4.14.0.1__5822-export-var-meta.sql
15:01 xarthisius
pdurbin: about your RO-Crate question: yeah, we're pondering adopting it but I'm not entirely convinced
15:01 pdurbin
Youssef_Ouahalou: you send the output from the asadmin "deploy" command, right? I don't see your server.log attached. Did you attach it?
15:01 pdurbin
xarthisius: cool. I'm hoping we can have a call with Whole Tale soon. :)
15:02 Youssef_Ouahalou
Where is the server.log?
15:02 xarthisius
https://github.com/ThomasThelen/RO-Crate-Example
15:02 xarthisius
AFAIR Tommy was even added as an author to the spec, he's WT dev
15:02 pdurbin
Youssef_Ouahalou: by default, it's at /usr/local/glassfish4/glassfish/domains/domain1/logs/server.log
15:03 pdurbin
Sure, I remember Tommy from the workshop in Chicago.
15:04 pdurbin
at least I think I do
15:05 pdurbin
Youssef_Ouahalou: any luck finding server.log?
15:05 Youssef_Ouahalou
I put it here?
15:05 pdurbin
Youssef_Ouahalou: it's probably pretty long so it would be best to email it to support
15:05 Youssef_Ouahalou
ok fine :)
15:06 pdurbin
xarthisius: this repo is less than a month old. Nice. :)
15:12 Youssef_Ouahalou
Have you seen the server.log that I send by email?
15:13 pdurbin
not yet, did you get another ticket number?
15:15 Youssef_Ouahalou
No, I answered the other mail. Would you prefer a new mail?
15:18 pdurbin
Youssef_Ouahalou: that's ok. I see it now. But this is not the whole server.log file, right? Im looking for more detail.
15:22 pdurbin
I haven't seen anything yet that jumps out at me as causing the deployment problem but maybe I could add donsizemore if he's available to give a second opinion. :)
15:24 Youssef_Ouahalou
I sent you another mail
15:27 pdurbin
Youssef_Ouahalou: thanks. How many lines does server.log have according to `wc`? server.log is rotated often so the information I'm looking for might be in an already rotated file.
15:34 Youssef_Ouahalou
Can I send you all the server.log file by mail?
15:35 pdurbin
Youssef_Ouahalou: sure, please send it to philip_durbin
15:35 Youssef_Ouahalou
ok fine thank you soo much
15:42 pdurbin
Youssef_Ouahalou: ok, I have your server.log and I think I see the problem.
15:42 pdurbin
Migration V4.13.0.1__3575-usernames.sql failed
15:43 pdurbin
Message : ERROR: could not create unique index "index_authenticateduser_lower_useridentifier"
15:43 Youssef_Ouahalou
I have to delete it?
15:44 donsizemore
@pdurbin i harvested from flavio at https://dataverse-test.irss.unc.edu/dataverse/dlstest and sent it to him personally
15:45 pdurbin
Youssef_Ouahalou: I think you need to do some clean up of duplicate usernames. Please see "In an effort to prevent accidental duplicate accounts, user spoofing, or other username-based confusion, this release introduces a database constraint that no longer allows usernames that are exactly the same but use different capitalization, e.g. Bob11 vs. bob11. You may need to do some cleanup before upgrading to
15:45 pdurbin
deal with existing usernames like this." at https://github.com/IQSS/dataverse/releases/tag/v4.15
15:45 pdurbin
donsizemore: oh good, thanks
15:45 pdurbin
Youssef_Ouahalou: "Once you identify the usernames that need cleaning up, you should use either Merge User Accounts (if it’s the same person) or Change User Identifier (if they are different people). After the cleanup you can safely upgrade without issue."
15:47 pdurbin
Youssef_Ouahalou: does that make sense? I can post the username here if you want. It starts with an "e". The line in server.log says "Detail: Key (lower(useridentifier::text))=(e******) is duplicated."
15:48 Youssef_Ouahalou
Thanks a lot for your help. So if I only delete their name in the dataverse the error will not happen again?
15:49 Youssef_Ouahalou
Yes, thank you very much, I think that's it too.
15:50 pdurbin
Youssef_Ouahalou: well, please don't delete anything directly in the database. :)
15:50 pdurbin
It's better to manage users via the Dashboard or APIs.
15:50 Youssef_Ouahalou
Yes, I'll do this :)
15:51 pdurbin
great, those release notes link to these APIs:
15:51 pdurbin
- http://guides.dataverse.org/en/latest/api/native-api.html#merge-user-accounts
15:51 pdurbin
- http://guides.dataverse.org/en/latest/api/native-api.html#change-user-identifier
15:51 pdurbin
We had to do a lot of merging and renaming for Harvard Dataverse.
15:56 pdurbin
donsizemore: it's funny, on Friday as I was heading home I ran ec2-list-all.sh and say 16 EC2 instances started by Jenkins. Now they've all been terminated by the magic of your scripts. :)
16:06 pdurbin
saw*
16:11 Youssef_Ouahalou
I managed to update at 4.15 infinitely thanks
16:15 pdurbin
Youssef_Ouahalou: great! Also, you saw that I opened https://github.com/IQSS/dataverse-installations/issues/30 , right?
16:18 Youssef_Ouahalou
Yes Thank you very much I sent them all the information needed last Thursday. :)
16:27 pdurbin
Youssef_Ouahalou: ok. I was curious about the status because in https://help.hmdc.harvard.edu/Ticket/Display.html?id=283158 I see "Okay, it's ready to be added to the map. I'll wait for your email."
16:28 Youssef_Ouahalou
Yes because we are only in test version there is a lot of draft dataset
16:30 Youssef_Ouahalou
thanks a lot for your help :)
16:50 nightowl313 joined #dataverse
16:53 Youssef_Ouahalou
And if not for WYSIWYG do you have something new?
17:35 pdurbin
ah, only in test, ok
18:42 donsizemore
@pdurbin knock knock?
18:42 pdurbin
who's there
18:42 donsizemore
orange
18:42 pdurbin
orange who?
18:42 Youssef_ouahalou joined #dataverse
18:42 pdurbin
so glad you didn't say banana
18:43 donsizemore
well, i was going to say "orange you glad it's not cold and rainy up there" then i checked the cambridge weather
18:43 donsizemore
it's 5°F cooler here than there
18:43 pdurbin
it's awful out
18:43 donsizemore
^there than here
18:43 Youssef_ouahalou
you think it will be in the future version?
18:43 pdurbin
Youssef_ouahalou: a WYSIWYG editor?
18:43 donsizemore
@pdurbin I think I have https://github.com/IQSS/dataverse-ansible/issues/133 working but a couple questions
18:44 pdurbin
nice!
18:45 donsizemore
@pdurbin question a) I just used the example TSV you pasted in the ticket, but it's only good for a specific sub-dataverse. okay to steal that file (or do you have another one you'd recommend) to pop into maybe the sample-data directory and allow for all Dataverses?
18:47 pdurbin
Huh. Only good for a specific sub-dataverse... I didn't know that was possible. Let me find some URLs of custom metadata blocks we use for Harvard Dataverse.
18:47 pdurbin
donsizemore: can you please try any of these that start with "custom"? https://github.com/IQSS/dataverse/tree/develop/scripts/api/data/metadatablocks
18:49 donsizemore
@pdurbin yes. if you specify a dataverse in the TSV, you may only enable it for the specified dataverse. and through the magic that is the API it is possible to enable a block on the _wrong_ dataverse and _really_ screw things up
18:49 donsizemore
@pdurbin (I make our archivists use the GUI for this reason)
18:50 pdurbin
interesting, I hope that behavior is documented :)
18:51 donsizemore
https://github.com/IQSS/dataverse/issues/5782
18:52 donsizemore
question b) did oliver's rewrite negate the need to pop all this stuff into solr buckets?
18:53 pdurbin
right, I forgot about this issue
18:54 pdurbin
thanks for the heads up about that codemeta custom metadata block tsv
18:54 pdurbin
I'm not sure what you mean by Solr buckets but Oliver's goal was to make the updating of the Solr schema scriptable. And I think it works.
18:55 donsizemore
because i can put the stuff into solr, but it'd just be the fieldnames, allowing customizable type would be a lot messier
18:55 pdurbin
Hmm, my thought was that we'd load the tsv and then run Oliver's script to update Solr.
18:56 donsizemore
well, that happens. and it looks like the copyfield sources aren't customized https://github.com/IQSS/dataverse/blob/develop/conf/solr/7.3.1/schema_dv_mdb_copies.xml
18:56 Youssef_ouahalou joined #dataverse
18:56 donsizemore
so we're just assuming that solr has the ability to write to its conf files
18:57 Youssef_ouahalou
@pdurbin yes for editor
18:57 pdurbin
Well, the idea is/was that schema_dv_mdb_copies.xml contains the fields that are enabled by default out of the box by the 6 or 7 blocks that are enabled.
18:58 pdurbin
but if you load up a few more TSVs and run that script, schema_dv_mdb_copies.xml should have the additional fields
18:59 pdurbin
schema_dv_mdb_fields.xml should have more fields at that point too
18:59 pdurbin
donsizemore: is this making sense?
19:00 pdurbin
Youssef_ouahalou: well, the future is a long time. It's forever, I guess. :) Features get added quicker (usually) when there is a pull request. :)
19:00 donsizemore
@pdurbin but the script just dumps those in /tmp
19:01 donsizemore
@pdurbin so in theory you'd have to run it on each solr launch?
19:01 Youssef_ouahalou
Ahhahaha ok thank you
19:01 pdurbin
I think you can configure where it dumps the files. Shouldn't have to be /tmp
19:02 pdurbin
And the idea is that you run the script as the Sorl user.
19:02 pdurbin
Youssef_ouahalou: honestly, instead of a WYSIWYG editor I think I would prefer Markdown support. Like we have in GitHub and Stackoverflow.
19:03 donsizemore
@pdurbin i've started one more run using a different TSV, I can make it dump to the solr conf dir
19:03 pdurbin
Solr* user
19:03 pdurbin
donsizemore: sounds good. Then Oliver's script is supposed to restart Solr at the end. So that the new fields are available.
20:15 pdurbin
donsizemore: you pull request looks great! I'm thinking we should disable the custom metadata blocks by default, though.
20:24 donsizemore
@pdurbin whoops meant to turn them back off
20:29 pdurbin
looks like it's ready to test, thanks
