sql-spark-connector issue where the access token expires post 60 mins

[shyamalesh]

Even after setting the default time out in the Azure active directory for the particular service principal to 9 hours. We are still facing the token expiry issue post 1 hour of execution.

Connection code from databricks to Azure SQL -
df.write
.format(DBConnection.connector_type_spark)
.mode("overwrite")
.option("url", url)
.option("dbtable", destination_tablename)
.option("accessToken", access_token)
.option("encrypt", "true")
.option("databaseName", DBConnection.jdbc_database)
.option("hostNameInCertificate", host_name_in_certificate)
.save()

Error message - org.apache.spark.SparkException: Job aborted due to stage failure: Task 157 in stage 29.0 failed 4 times, most recent failure: Lost task 157.8 in stage 29.0 (TID 19900) (10.139.64.13 executor 0): com.microsoft.sqlserver.jdbc.SQLServerException: Login failed for user '<token-identified principal>'. Token is expired. ClientConnectionId:c79c98cb-8d3e-421f-88f0-971508e53a72

Is there any way to connect to Azure sql server from databricks using ECDOMAIN without password ?

Example -
df.write
.format("com.microsoft.sqlserver.jdbc.spark")
.mode("overwrite")
.option("url", url)
.option("dbtable", dbtable)
.option("user", ECDOMAIN\username)
.save()

[mcbdx]

Hello,

I'm having a similar issue, except that the token works for the first 30 seconds only. Did you ever find a solution to this? How are you acquiring your access token?