Changing filename causes error

[taojing2002]

Robyn reported the issue on dev.nceas, which is Metacat 3.1.0:

1. make a new package with a data file, save it
2. edit data package and rename the file, save and see error:

HashStore refs files already exist for pid resource_map_urn:uuid:58d3047d-40eb-44cc-88e2-56bff3f3283d and cid: 
8ed23f53022ab5a04687d63b2402e150e9312db3211029e559c269b4ad9a5c26

[taojing2002]

I looked at the tomcat log file and found:

StreamingMultipartRequestResolver.resoloveMulitpart - Metacat is handling the object stream AFTER handling the 
system metadata stream. StreamResolver will store the object with identifier
resource_map_urn:uuid:cdf928df-f5a2-49ac-bcfa-e79309c68930 , declared size 2957 and calculating checksum using
algorithm MD5 [edu.ucsb.nceas.metacat.restservice.multipart.StreamingMultipartRequestResolver:resolveMultipart:197]

The identifier of resource_map_urn:uuid:cdf928df-f5a2-49ac-bcfa-e79309c68930 is the original resource map. I am just wondering why MetacatUI sends the identifier again.

[doulikecookiedough]

@taojing2002 The hashstore error described above occurs when HashStore stores a data object with a given pid with a data file that already exists (and is tagged appropriately).

It seems like when Robyn renames the data file, it attempts to re-store the same data object with the existing pid (and ultimately sees that everything exists where it should be).

Do you think this is an issue with how we handle re-naming of a data file? Should renaming a data file mean it has to be stored again...?

[taojing2002]

@doulikecookiedough I looks like not related to hashstore. I tested the scenario on Metacat 2.19.1, which doesn't use hashstore and got the error:

The requested identifier resource_map_urn:uuid:a3caa3ce-5704-497d-aa6e-0c45cab5cacb is already used by another object andtherefore can not be used for this object. Clients should choosea new identifier that is unique and retry the operation or use CN.reserveIdentifier() to reserve one.

It seems MetacatUI reuses (uploads) the original resource map again to Metacat

[iannesbitt]

@robyngit Can we transfer this issue to NCEAS/MetacatUI?

[robyngit]

Yes, already created a MetacatUI ticket here: NCEAS/metacatui#2590. Closing.