I recommend using SQL Server as metastore. Each time when you init the metastore, you lose all the Hive metadata like databases, tables, etc.
person R4F43L access_time 2 months ago Re: Apache Hive 3.1.2 Installation on Windows 10
In your tutorial i use the derby embeded so i try to use and external derby following this tutorual "Installing Apache Hive 3.1.2 on Windows 10 - https://towardsdatascience.com/".
And it works fine but when i stop the hive server killing the process sometimes i have the same error so i remove metastore_db and with cygwin i do :
$HIVE_HOME/bin/schematool -dbType derby -initSchema
But all hive query are very slow... maybe it caused by the limitation you mention ". So i will try to change the db of the metastore maybe i will try to do with sql server running in Docker following this tutotorial how-to-run-sql-server-in-a-docker-container.
You think hive query can be faster if we change engin MR to Tez, you have a step-by-step to adapt it to tez ?
I share with you some issues i found during my test, in my case i want to write parquet files in hdfs then create hive external tables to query data and insert it on an internal hive table where i also need to delete data (this table must be transactional)
Error: User: MYUSERNAME is not allowed to impersonate MYUSERNAME
I fix it by following : stackoverflow - 43180305
I fix it by following : stackoverflow.com - 29602670
Here the properties i added to fix my issues :
Forum discussions for column Hadoop.
Want to contribute on Kontext to help others?