The issue with SQL-endpoint + external hive metastore

[vpismarev]

Hi

We have faced an issue recently (everything worked well 1 week before)

We are using exnernal hive metastore:

"spark.hadoop.datanucleus.connectionPool.maxPoolSize" : "10"
 
"spark.sql.hive.metastore.jars" : "/dbfs/ivyjars/*"
 
"spark.sql.hive.metastore.version" : "3.1.0"

and now it looks like all access for the metastore from sql-endpoints are stuck (I am observing a lot of queued session from query history like "show databases", "show table" and so on )

the version of sql-warehouse:

Channel Current (v 2022.40)

We are using rds for hive metastore and I'm observing a lot of idle sessions such as:

idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,SET application_name = 'PostgreSQL JDBC Driver'
idle,,,COMMIT

What is the issue?

P.S.

The restarting of the SQL -warehouse helps sometimes, but the problem still exists

[Ajay-Pandey]

Generally, this should not be a case. Please raise a support request to databricks