mcbdx
commented
3 months ago 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?
Open shyamalesh opened 8 months ago
mcbdx
commented
3 months ago 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?
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()