You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It should not remove existing permissions everytime we add new groups.
Actual Behavior
Everytime this resource gets applied it to hive_metastore it removes all permissions and then adds all permissions in configuration. This will result in outages due to permissioning of hive metastore being completely removed and and readded during this time period. Also the time period it takes to apply this is directly proportional to the number of schemas. We are up to having to apply terraform twice (because 1st apply fails out), took a period of 90 minutes.
Steps to Reproduce
terraform apply this resource with an initial group.
terraform apply this resource adding permissions for the second group (you will see that it says it is only adding the permissions for the second group which it should be)
In reality, you are removing the permissions for all groups and re-adding the initial group, and second group's permissions.
Terraform and provider versions
Latest (issue has always existed)
Is it a regression?
No it has always worked like this and needs to be fixed. It's causing outages for customers using hive_metastore.
Debug Output
Important Factoids
Nope, we do have a large amount of schemas.
Would you like to implement a fix?
No
The text was updated successfully, but these errors were encountered:
forma-wolf
changed the title
[ISSUE] Issue with databricks_XXX resource
[ISSUE] Issue with hive_metastore_databricks_sql_permissions resource
Feb 5, 2025
forma-wolf
changed the title
[ISSUE] Issue with hive_metastore_databricks_sql_permissions resource
[ISSUE] Issue with databricks_sql_permissions resource
Feb 5, 2025
Configuration
Expected Behavior
It should not remove existing permissions everytime we add new groups.
Actual Behavior
Everytime this resource gets applied it to hive_metastore it removes all permissions and then adds all permissions in configuration. This will result in outages due to permissioning of hive metastore being completely removed and and readded during this time period. Also the time period it takes to apply this is directly proportional to the number of schemas. We are up to having to apply terraform twice (because 1st apply fails out), took a period of 90 minutes.
Steps to Reproduce
Terraform and provider versions
Is it a regression?
Debug Output
Important Factoids
Would you like to implement a fix?
The text was updated successfully, but these errors were encountered: