Hashlists used for the EUS synchronization script were being incorrectly generated after the move to UUID-based relationships in the mapping tables. Previously, the two columns being mapped served as a composite primary key for the table, and those were being used to generate the unique hash for that record. This hash could then be compared to the hash from the corresponding EUS table to judge the status of that record (deleted, changed, new). With the new UUID identifiers, those keys went away and the hashing function was just hashing the UUID value. Everything looked like it was missing from the metadata tables.
Steps to Reproduce:
Run the EUS sync script and let it get to the mapping relationships.
Expected Result:
If you're in normal (not forced update) mode, the script should pull and compare hashes from metadata and EUS to see which items need to be moved. Only missing and/or changed items should be affected.
Actual Result:
Everything in the mapping tables looks like it needs to move across. No real harm is being done, but it slows down the process when it tries to poke in every value into the table and fails due to key collision.
Pacifica Metadata version
v0.10.3
Pacifica Core Software versions
v0.99.0
Platform Details
Production MyEMSL
Scenario:
Hashlists used for the EUS synchronization script were being incorrectly generated after the move to UUID-based relationships in the mapping tables. Previously, the two columns being mapped served as a composite primary key for the table, and those were being used to generate the unique hash for that record. This hash could then be compared to the hash from the corresponding EUS table to judge the status of that record (deleted, changed, new). With the new UUID identifiers, those keys went away and the hashing function was just hashing the UUID value. Everything looked like it was missing from the metadata tables.
Steps to Reproduce:
Run the EUS sync script and let it get to the mapping relationships.
Expected Result:
If you're in normal (not forced update) mode, the script should pull and compare hashes from metadata and EUS to see which items need to be moved. Only missing and/or changed items should be affected.
Actual Result:
Everything in the mapping tables looks like it needs to move across. No real harm is being done, but it slows down the process when it tries to poke in every value into the table and fails due to key collision.