8.2 GenAI Issues

This section describes some commonly encountered issues and errors for GenAI and their workarounds.

Issue: When you try to verify whether the vector embeddings were correctly loaded, if you see a message which indicates that the vector embeddings or table did not load in MySQL AI, then it could be due one of the following reasons:
- The task that loads the vector embeddings into the vector store table might still be running.
  Workaround: Check the task status by using the query that was printed by the VECTOR_STORE_LOAD routine:
```
SELECT * from mysql_task_management.task_status where id = TaskID;
```
  Or, to see the log messages, check the task logs table:
```
SELECT * from mysql_task_management.task_log where task_id = TaskID;
```
  Replace TaskID with the ID for the task which was printed by the VECTOR_STORE_LOAD routine.
- The folder you are trying to load might contain unsupported format files or the file that you are trying to load might be of an unsupported format.
  Workaround: The supported file formats are: PDF, TXT, PPT, HTML, and DOC.
  If you find unsupported format files, then try one of the following:
  - Delete the files with unsupported formats from the folder, and run the VECTOR_STORE_LOAD command again to load the vector embeddings into the vector store table again.
  - Move the files with supported formats to another folder, create a new PAR and run the VECTOR_STORE_LOAD command with the new PAR to load the vector embeddings into the vector store table again.
Issue: the VECTOR_STORE_LOAD command fails unexpectedly.
Workaround: Ensure that you use the --sqlc flag when you connect to your database system:
```
mysqlsh -uAdmin -pPassword -hPrivateIP --sqlc
```
Replace the following:
- Admin: the admin name.
- Password: the database system password.
- PrivateIP: the private IP address of the database system.