[BUG]: Query failures when workspace default catalog is not hive_metastore #2207
Description
Is there an existing issue for this?
- I have searched the existing issues
Current Behavior
If the workspace where UCX is running has the default catalog set to something other than hive_metastore, multiple UCX cli commands will fail.
Expected Behavior
The queries executed by UCX should all be namespaced into hive_metastore without relying on the default catalog setting.
Steps To Reproduce
- Execute the assessment workflows
- Update the workspace default catalog to another UC catalog than
hive_metastore - Run
databricks labs ucx create-catalogs-schemas- error occurs. Error message is not accurate. Using--debugto see the correct error
ERROR [d.l.ucx.create-catalogs-schemas] ValueError: cluster_id is required in the configuration. Config: host=https://<redacted>.cloud.databricks.com, token=***, auth_type=pat. Env: DATABRICKS_HOST, DATABRICKS_TOKEN
18:45:11 DEBUG [d.l.lsql.backends] [api][fetch] SELECT * FROM ucx.mounts
18:45:11 DEBUG [d.l.lsql.core] Executing SQL statement: SELECT * FROM ucx.mounts
18:45:12 DEBUG [databricks.sdk] POST /api/2.0/sql/statements/
> {
> "format": "JSON_ARRAY",
> "statement": "SELECT * FROM ucx.mounts",
> "warehouse_id": "<redacted>"
> }
< 200 OK
< {
< "statement_id": "<redacted>",
< "status": {
< "error": {
< "error_code": "BAD_REQUEST",
< "message": "[TABLE_OR_VIEW_NOT_FOUND] The table or view `ucx`.`mounts` cannot be found. Verify the spelling ... (292 more bytes)"
< },
< "state": "FAILED"
< }
< }
Cloud
AWS
Operating System
macOS
Version
latest via Databricks CLI
Relevant log output
No response
Metadata
Metadata
Assignees
Labels
No labels