Time
S
Nick
Message
01:42 jri joined #dataverse
07:58 jri joined #dataverse
12:01 jri joined #dataverse
12:15 donsizemore joined #dataverse
13:33 donsizemore joined #dataverse
14:42 Thalia_UM joined #dataverse
14:42 Thalia_UM
Hiii!!! Good morning! :)
14:43 pdurbin
Thalia_UM: have you had enough coffee? ;)
14:44 Thalia_UM
Hahahaha no :S
14:45 Thalia_UM
I can't drink coffee because I have gastritis
14:45 Thalia_UM
:-$
14:45 pdurbin
:(
14:45 pdurbin
I'll drink an extra up for you. :)
15:11 rebecabarros joined #dataverse
15:17 rebecabarros
pameyer: about DCM /hold setting, it's enough point out '/usr/local/glassfish4/glassfish/domains/domain1/files" or do I have to put everything user
15:43 pameyer joined #dataverse
15:50 pameyer
rebecabarros: for NFS, the UID/GID that the DCM sets the files to after moving them to `/hold` needs to match the UID/GID that glassfish is running as. I may not be understanding your question though (aka - where `user
15:55 rebecabarros
pameyer: https://pastebin.com/5vhNLZmh - is that right for main.yml?
15:57 pameyer
two questions: which user is glassfish running as, and are you trying to run DCM and DV on the same system?
15:58 rebecabarros
since I'm running locally I try to with: ansible-playbook -i local.hostlist dcm.yml but I've got an error that maybe is related with my ansible config - https://pastebin.com/b6yVAb8t
15:59 pameyer
that error is an ssh error from ansible
16:01 rebecabarros
they're running in different systems. And I'm afraid that glassfish is running as root
16:22 Thalia_UM
I have a question
16:22 Thalia_UM
Can i change dc-terms for dc???
16:23 Thalia_UM
at metadata file
16:23 Thalia_UM
???
16:56 pameyer joined #dataverse
17:13 pameyer
rebecabarros: glassfish running as root isn't a problem for this. do you have NFS setup between the DV and DCM systems?
17:13 pameyer
Thalia_UM: is this the metadata file for exports, or something else?
17:13 Thalia_UM
no
17:14 pdurbin
Thalia_UM: are you using SWORD?
17:14 pdurbin
I'm confused.
17:14 Thalia_UM
yes
17:15 Thalia_UM
It's to load the metadata and I want to know if I can attach email and other attributes
17:19 pdurbin
Thalia_UM: have you seen http://guides.dataverse.org/en/4.8/api/sword.html#dublin-core-terms-dc-terms-qualified-mapping-dataverse-db-element-crosswalk ?
17:20 rebecabarros joined #dataverse
17:22 Thalia_UM
Can I add other labels?
17:22 Thalia_UM
as <dcterms: format> <dcterms: size>
17:23 Thalia_UM
<dcterms:creator emali="">
17:23 pdurbin
no
17:23 rebecabarros
pameyer: is this NFS, right? https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/3/html/Reference_Guide/ch-nfs.html I think not because I did not do any change regarding this, I will check.
17:24 pdurbin
rebecabarros: yes, that's what NFS is
17:24 Thalia_UM
So, Can I only load the ones described in the example?
17:24 pameyer
it doesn't have to be NFS, but DCM and DV need to have a shared filesystem
17:25 pdurbin
Thalia_UM: that's correct. If you need more metadata fields, you will need to switch from SWORD (XML ) to the "native" API (JSON ).
17:27 pdurbin
pameyer: I was just thinking to myself what else one could use for a shared filesystem besides NFS. I guess NFS is the oldest and most popular.
17:27 pameyer
SMB *should* work, although in my experience it does odd things with permissions
17:29 pdurbin
yeah
17:35 pdurbin
Thalia_UM: you could try this: curl -s -X POST -H "Content-type:application/json" -d @scripts/search/tests/data/dataset-finch1.json "http://localhost:8080/api/dataverses/finches/datasets/?key=$FINCHKEY "
17:35 pdurbin
Thalia_UM: here's the JSON file https://github.com/IQSS/dataverse/blob/develop/scripts/search/tests/data/dataset-finch1.json
17:37 rebecabarros
in case of NFS, DCM is my NFS server and DV is my NFS client? how should I set the access options for /etc/exports?
17:40 Thalia_UM
Philip, I am confused me with respect to the subject, I would like you to explain to me what is the use of loading the XML file through SWORD and json with the native API
17:40 pameyer
rebecabarros: DCM as NFS server, DV as NFS client would work (so would the other way around, as would having them both be NFS clients of another fileserver).
17:41 pameyer
for glassfish as root, you'll need `no_root_squash` in /etc/exports
17:42 pdurbin
Thalia_UM: can you please email support
17:42 Thalia_UM
Yes philip
17:42 Thalia_UM
thank you
17:46 Thalia_UM
I just got confused me a coworker :-$
18:02 rebecabarros
pameyer: should I mount the shared NFS directory in my Dataverse files directory?
18:04 pameyer
rebecabarros: that should work (or create a symbolic link from your dataverse files directory)
18:05 Thalia_UM
Philip
18:06 Thalia_UM
the file when exporting the metadata, is equal to the metadata example for SWORD?
18:07 pdurbin
Thalia_UM: no, it is not equal, unfortunately :(
18:07 Thalia_UM
In the repository that has harvard as they load the metadata?
18:07 Thalia_UM
Ah ok
18:08 pdurbin
I think rebecabarros was asking a similar question once. Can you created a dataset using DDI XML , which Dataverse exports. The answer no. (but please feel free to open an issue about this, rebecabarros !)
18:12 jri joined #dataverse
18:21 jri joined #dataverse
18:21 donsizemore joined #dataverse
18:28 rebecabarros
yes, pdurbin and Thalia_UM. the metadada json that you get from http://$SERVER/api/datasets/$id/versions/$versionNumber?key=$apiKey is different from the one in api export option. For now you need to use the first json model to be able to update metadada using the api.
18:30 pdurbin
It would be more logical if data in matched data out. Sorry.
18:34 pameyer
huh - I hadn't know there was a way other than `http://$SERVER/api/datasets/$id/versions/$versionNumber`
18:34 pdurbin
pameyer: export was shipped along with harvesting
18:34 pameyer
ah - gotcha
18:35 pameyer
pdurbin: any ideas if what folks have been referring to as the "edit API " is the "replace all metadata items with this json" API, or if this is something that allows changing a single field?
18:36 pameyer
rebecabarros's mention of updating metadata reminded me
18:37 pdurbin
Edit via API is always a replace of the entire contents of the metadata. You can't simply change a single field such as the title of the dataset. You have to include all the metadata in either XML (SWORD) or JSON (native) format.
18:38 Thalia_UM
Can i write that? ---> <dcterms:creator Contact="thalia
18:39 rebecabarros
pameyer: all set regarding NFS. the list of dependencies, do I need to install theses packages before running the playbook or this action are included in playbook execution? I'm still getting the ssh error regarding ansible. Trying to solve this.
18:40 pdurbin
Thalia_UM: are you trying to fill in "Author" or "Contact"?
18:41 pameyer
rebecabarros: if I'm remembering correctly, the playbook doesn't attempt to configure NFS during install
18:42 Thalia_UM
Author with contact
18:43 rebecabarros
pameyer: not the NFS, but the packages: lighttpd, sshd, rsync...
18:43 pdurbin
Thalia_UM: are you able to fill in the email address of an author using the GUI ?
18:43 Thalia_UM
omplete Author
18:43 pameyer
pdurbin: guess I'm not understanding the difference between "create" and "edit" - replace-all seems like create to me (but I don't need to get un-confused on this today)
18:43 Thalia_UM
Complete Author
18:44 Thalia_UM
Can it be completed?
18:44 Thalia_UM
Can I complete the author's email on the <dcterms: creator> tag?
18:47 pdurbin
Thalia_UM: I'm asking you to look at the GUI because it is not possible via the GUI. It is also not possible via API .
18:47 Thalia_UM
Why?
18:48 pdurbin
No one has wanted to enter email addresses for authors. If you want this, please open a GitHub issue. Most people enter the email address under "Contact".
18:48 pameyer joined #dataverse
18:49 Thalia_UM
are there more tags than the GUI ? Another website that shows the metadata example?
18:50 pdurbin
The GUI is the best place to see all of the available metadata fields.
18:50 Thalia_UM
Who is the tag od Contact?
18:51 Thalia_UM
dcterms:contributor datasetContactEmai
18:51 Thalia_UM
What is the tag of Contact?
18:53 pdurbin
'“Contact E-mail” is automatically populated from dataset owners email.' http://guides.dataverse.org/en/4.8/api/sword.html Thalia_UM
18:58 pameyer joined #dataverse
18:59 Thalia_UM
Yes, I read that part
19:01 pameyer joined #dataverse
19:05 rebecabarros
pameyer: in dev.hostlist should I set the IP of DV machine?
19:07 pameyer
IP or hostname should work
19:12 rebecabarros
pameyer: ok. I've got an error in the task: rsync symlink
19:14 rebecabarros
I have to run the playbook with sudo?
19:15 pameyer
`ansible-playbook -u root` might help. is the error permission related?
19:17 rebecabarros
seems like: TASK [dcm : rsync symlink] ****************************************************************************************************************************************************************************************************************** fatal: [172.16.9.36]: FAILED! => {"changed": false, "failed": true, "gid": 0, "group": "root", "mode": "0755", "msg": "refusing to convert between file and link for /bin/rsync", "owner": "root",
19:23 pameyer
you found a point where different linux distributions are different - if rsync is at `/bin/rsync`, then that part of the playbook isn't needed (so you can comment it out)
19:32 rebecabarros
what should I comment it out? And where? config.yml?
19:39 rebecabarros
I will have to go. Thanks for all patience and help today pameyer and pdurbin. I guess that I almost there.
19:39 pameyer joined #dataverse
19:39 rebecabarros left #dataverse
20:04 axfelix joined #dataverse
20:05 pameyer joined #dataverse
20:20 pameyer
roles/dcm/tasks/config.yml for the symlink job
20:25 pdurbin left #dataverse
20:30 pameyer joined #dataverse
20:33 donsizemore joined #dataverse
20:34 donsizemore
@pdurbin just curious if each build still runs schemaspy per https://github.com/IQSS/dataverse/commit/184e687a33ac3ff99b8622ef858f563e61d1e393 ?
20:35 donsizemore
@pdurbin i ask because https://apitest.dataverse.org/guides/developers/database/schemaspy/relationships.html is dated june 2015
20:39 donsizemore joined #dataverse
20:48 pameyer
donsizemore: it looks to me like it should be running
20:49 pameyer
I'm wondering about the hostname though
21:15 donsizemore joined #dataverse
21:29 pameyer joined #dataverse
22:37 axfelix joined #dataverse
23:48 pdurbin joined #dataverse
23:49 pdurbin
The script is now at https://github.com/IQSS/dataverse/blob/v4.8.1/scripts/deploy/phoenix.dataverse.org/post#L15 and the schemaspy data is now here: http://phoenix.dataverse.org/schemaspy/latest/
