v5.14 : ReExportAll fails for some datasets

24 views
Skip to first unread message

jeyala...@yahoo.com

unread,
Nov 6, 2025, 8:45:12 AM (yesterday) Nov 6
to Dataverse Users Community
Hi!
     We upgraded to v5.14 and ran ReExportAll to update metadata exports.
The export failed for 28 datasets. Please see attached log for one dataset.

When we analyze these datasets, for all of them, metadata export of "DDI" and "DDI HTML Codebook" formats is throwing error.
  Few sample datasets are:


All datasets contain at least one TAB file and error is due to the content of the files.
Can someone please help on how to rectify this?

Thanks and regards,
Jeya



reexportall_1RDZ7Z.txt

Philip Durbin

unread,
Nov 6, 2025, 2:43:29 PM (yesterday) Nov 6
to dataverse...@googlegroups.com
Hi Jeya,

Thanks, this is helpful:

  <exception>
    <message>io.gdcc.spi.export.ExportException: Unknown runtime exception exporting metadata. </message>
    <frame>
      <class>edu.harvard.iq.dataverse.harvest.server.OAIRecordServiceBean</class>
      <method>exportAllFormatsInNewTransaction</method>
      <line>249</line>
    </frame>


Do you see a stacktrace in server.log? If not, are you able to modify the code to print a stacktrace just above that line, and then build and redeploy a war file? I know this is asking a lot, but it would help us understand what's going on.

Thanks,

Phil

 


--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dataverse-community/5b6ef430-5c85-4355-b9ed-5cc93c568844n%40googlegroups.com.


--

jeyala...@yahoo.com

unread,
2:55 AM (18 hours ago) 2:55 AM
to Dataverse Users Community
Dear Philip,
     Thank you very much for the prompt reply. Adding the code to print stacktrace may not be immediate. 
     Yesterday, for ReExportAll failures, nothing is logged in server.log files.
     But when we did the individual reExportDataset, the exception is logged into server.log. Please see attached logs for 2 dataset ids doi:10.21979/N9/SQCDCB
 and doi:10.21979/N9/1RDZ7Z. Hope these are of of some help to troubleshoot.
       I see that error log from export_*.log file I attached yesterday also contains the same information, but in different format (xml). 

The command used are:

 One more issue we faced for reExportDataset is: we couldn't run it with DOI . We need to use dataset id only. Following error is thrown:
curl http://$SERVER/api/admin/metadata/:persistentId/reExportDataset?persistentId=doi:10.21979/N9/SQCDCB?unblock-key=$KEY
{ status:"error", message:"Endpoint available using API key only. Please contact the dataverse administrator"}
  If we add Admin API token also the error is thrown.

Thanks and regards,
Jeya
server_log_1RDZ7Z.txt
server_log_SQCDCB_2.txt
server_log_SQCDCB_1.txt

Philip Durbin

unread,
6:53 AM (14 hours ago) 6:53 AM
to dataverse...@googlegroups.com
Hi Jeya,

It's great that you're already getting a stacktrace. Thanks for sending them. I'm looking for a "caused by" line but I don't see one. If there is any more output, please send it along.

As for the other error, if you replace the second ? with a & it should work, like this:

curl "http://$SERVER/api/admin/metadata/:persistentId/reExportDataset?persistentId=doi:10.21979/N9/SQCDCB&unblock-key=$KEY"

That is, & is used to add each additional query parameter beyond the first one. Also, the whole URL needs to be in the quotes or else your shell (e.g. bash) will interpret the & as a command.

I hope this helps,

Phil


To view this discussion visit https://groups.google.com/d/msgid/dataverse-community/b6ce613e-1d83-44ff-b2f6-874282024108n%40googlegroups.com.

James Myers

unread,
8:09 AM (13 hours ago) 8:09 AM
to dataverse...@googlegroups.com

FWIW: PR #11782 adds a print stacktrace at the place that you’d want. Doesn’t help you with 5.14 directly, but people seeing related issues after v6.9 won’t need to add it.

 

One thing you might try – if there are no cached export files for a dataset and you request one specific one from the UI, I think the code will just try to generate that one export format – which may show you more details in the log that when the problem is being caught in the reExportAll method.

 

-- Jim

jeyala...@yahoo.com

unread,
8:25 AM (13 hours ago) 8:25 AM
to Dataverse Users Community
Thanks Phil and Jim for all the suggestions!!
@Phil : I am attaching the files with more logs around the "Error exporting" stack trace. But those are usual IOExceptions logged everyday and I couldn't find the relevance.
Regarding the curl command, I already did try with & in place of ?, but the missing part was " " as you correctly pointed out. After adding it, it is working now. Thanks.
@Jim: Can you please tell me where to look for the cached export files? I had a doubt regarding this from the release notes:  "When replacing an existing format, one must also remember to delete the cached metadata export files or run the reExport command for the metadata exports of existing datasets to be updated." Since we are reExport-ing I I thought it should be ok.
Regards,
server_log_SQCDCB_more_logs.txt
server_log_SQCDCB_1RDZ7Z_more_logs.txt

James Myers

unread,
8:34 AM (13 hours ago) 8:34 AM
to dataverse...@googlegroups.com

Jeye,

The cached export files are in the store used by that dataset. Each dataset’s files are stored in subdirectories related to their PID, i.e. doi:10.5072/ FK2LDVRQ6 has files in the store’s /10.5072/ FK2LDVRQ6/ . In that directory, you should see files with names like export_Datacite.cached. You can delete those.

 

That said, if you’re failing to create the DDI file, presumably it’s not getting cached and therefore you can just try to re-export from the UI.

Reply all
Reply to author
Forward
0 new messages