[Archivesspace_Users_Group] Curious! : Import Job says it failed but resources imported

[Majewski, Steven Dennis (sdm7g)]

Just noting that I’m still seeing this phenomenon with 1.4.2 on a clean install starting with a db:nuke database
and a clean archivesspace/data directory.

It only seems to occur when importing a large number of EAD files in one batch.
If the import is broken up into smaller batches it works as expected. ( No problems with a batch of 10 )
With a batch of 50 or more, it imports all of the guides, then attempts to import the first guide a second time,
and the batch fails with an error message about duplicate ID.  However, clicking on ‘browse resources’, all
of the EAD files appear to have been imported.

In this last case, I selected 64 files, the import job lists 64 files, and I get 64 resources.
Viewing the Failed background job shows 64 files.
The data/shared/job_files/import_job_1/ directory contains 64 hex-renamed xml files plus output.log.
/repositories/4/jobs/1 lists 64 filenames.
But the log shows that after processing 64 files, it attempts to import the first one on the list again
and fails.


— Steve.


On Jan 12, 2016, at 5:56 PM, Majewski, Steven (sdm7g) <sdm7g at eservices.virginia.edu<mailto:sdm7g at eservices.virginia.edu>> wrote:


Very curious error observed:

Trying to move 50 resources from test server to production.

Same version: ArchivesSpace v1.4.2 running on all servers.

Exported all 50 from test server as EAD.
Imported all 50 onto another test server as initial test.
Had to redo when I found that I needed to add param to export unpublished notes.
But both imports completed without complaint.

Attempted to repeat the same import on production server.

Imports appeared to be running successfully for an uncounted number of files, but gave this error message at the end:

================================================== 54-Mss_77-1.xml ================================================== 1. STARTED: Reading JSON records 1. DONE: Reading JSON records 2. STARTED: Validating records and checking links 2. DONE: Validating records and checking links 3. STARTED: Evaluating record relationships 3. DONE: Evaluating record relationships 4. STARTED: Saving records: cycle 1 Created: /revision_statement/import_42935b5d-613c-42fa-a297-d87bd61a77f3 Created: /revision_statement/import_d0d44810-bbf2-448a-803e-59e4abe479b2 Created: /revision_statement/import_b900aae8-c51e-45d6-afd3-31d8cd45b35e Created: /revision_statement/import_42b9bd89-2058-462a-9425-8b317bacb633 5. STARTED: Cleaning up 5. DONE: Cleaning up !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! IMPORT ERROR !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! The following errors were found: id_0 : That ID is already in use ead_id : Must be unique %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Full Error Message: Problem creating 'Papers of Frederick D. G. Ribble': id_0 That ID is already in use, ead_id Must be unique !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!


However, on clicking on the Browse Resources link for that repo, it appears that all of the resources are there.
Double checking on the status of that background import job shows Job Status as “Failed” .

I stopped server, deleted solr index and indexer_state files, and restarted just to be sure of consistent state.
No change. Job has failed but it appears I have all 50 imported resources.

The file in the error message: 54-Mss_77-1.xml, was the first file on the list of Import files in the Batch Job page.
It is not listed twice in that file list, but it appears both at the start and the end of the import log,
with that failure message at the end.

I believe that every other time I’ve had a batch job import failure on a single file, NONE of the files are imported.



— Steve Majewski











-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20160201/f3d79591/attachment.html>